Zuben E. Sauna

Silent (synonymous) SNPs: should we care about them?

One of the surprising findings of the Human Genome Project was that single nucleotide polymorphisms (SNPs), which, by definition, have a minor allele frequency greater than 1%, occur at higher rates than previously suspected. When occurring in the gene coding regions, SNPs can be synonymous (i.e., not causing a change in the amino acid) or nonsynonymous (when the amino acid is altered). It has long been assumed that synonymous SNPs are inconsequential, as the primary sequence of the protein is retained. A number of studies have questioned this assumption over the last decade, showing that synonymous mutations are also under evolutionary pressure and they can be implicated in disease. More importantly, several of the mechanisms by which synonymous mutations alter the structure, function, and expression level of proteins are now being elucidated. Studies have demonstrated that synonymous polymorphisms can affect messenger RNA splicing, stability, and structure as well as protein folding. These changes can have a significant effect on the function of proteins, change cellular response to therapeutic targets, and often explain the different responses of individual patients to a certain medication.

Exposing synonymous mutations

Synonymous codon changes, which do not alter protein sequence, were previously thought to have no functional consequence. Although this concept has been overturned in recent years, there is no unique mechanism by which these changes exert biological effects. A large repertoire of both experimental and bioinformatic methods has been developed to understand the effects of synonymous variants. Results from this body of work have provided global insights into how biological systems exploit the degeneracy of the genetic code to control gene expression, protein folding efficiency, and the coordinated expression of functionally related gene families. Although it is now clear that synonymous variants are important in a variety of contexts, from human disease to the safety and efficacy of therapeutic proteins, there is no clear consensus on the approaches to identify and validate these changes. Here, we review the diverse methods to understand the effects of synonymous mutations.

Building better drugs: developing and regulating engineered therapeutic proteins.

Most native proteins do not make optimal drugs and thus a second- and third-generation of therapeutic proteins, which have been engineered to improve product attributes or to enhance process characteristics, are rapidly becoming the norm. There has been unprecedented progress, during the past decade, in the development of platform technologies that further these ends. Although the advantages of engineered therapeutic proteins are considerable, the alterations can affect the safety and efficacy of the drugs. We discuss both the key technological innovations with respect to engineered therapeutic proteins and advancements in the underlying basic science. The latter would permit the design of science-based criteria for the prediction and assessment of potential risks and the development of appropriate risk management plans. This in turn holds promise for more predictable criteria for the licensure of a class of products that are extremely challenging to develop but represent an increasingly important component of modern medical practice.

Fc fusion as a platform technology: potential for modulating immunogenicity

The platform technology of fragment crystallizable (Fc) fusion, in which the Fc region of an antibody is genetically linked to an active protein drug, is among the most successful of a new generation of bioengineering strategies. Immunogenicity is a critical safety concern in the development of any protein therapeutic. While the therapeutic goal of generating Fc-fusion proteins has been to extend half-life, there is a critical mass of literature from immunology indicating that appropriate design of the Fc component has the potential to engage the immune system for product-specific outcomes. In the context of Fc-fusion therapeutics, a review of progress in understanding Fc biology suggests the prospect of engineering products that have an extended half-life and are able to modulate the immune system.

Large scale analysis of the mutational landscape in HT-SELEX improves aptamer discovery

High-Throughput (HT) SELEX combines SELEX (Systematic Evolution of Ligands by EXponential Enrichment), a method for aptamer discovery, with massively parallel sequencing technologies. This emerging technology provides data for a global analysis of the selection process and for simultaneous discovery of a large number of candidates but currently lacks dedicated computational approaches for their analysis. To close this gap, we developed novel in-silico methods to analyze HT-SELEX data and utilized them to study the emergence of polymerase errors during HT-SELEX. Rather than considering these errors as a nuisance, we demonstrated their utility for guiding aptamer discovery. Our approach builds on two main advancements in aptamer analysis: AptaMut—a novel technique allowing for the identification of polymerase errors conferring an improved binding affinity relative to the ‘parent’ sequence and AptaCluster—an aptamer clustering algorithm which is to our best knowledge, the only currently available tool capable of efficiently clustering entire aptamer pools. We applied these methods to an HT-SELEX experiment developing aptamers against Interleukin 10 receptor alpha chain (IL-10RA) and experimentally confirmed our predictions thus validating our computational methods.

Characterization of Coding Synonymous and Non-Synonymous Variants in ADAMTS13 Using Ex Vivo and In Silico Approaches

Synonymous variations, which are defined as codon substitutions that do not change the encoded amino acid, were previously thought to have no effect on the properties of the synthesized protein(s). However, mounting evidence shows that these “silent” variations can have a significant impact on protein expression and function and should no longer be considered “silent”. Here, the effects of six synonymous and six non-synonymous variations, previously found in the gene of ADAMTS13, the von Willebrand Factor (VWF) cleaving hemostatic protease, have been investigated using a variety of approaches. The ADAMTS13 mRNA and protein expression levels, as well as the conformation and activity of the variants have been compared to that of wild-type ADAMTS13. Interestingly, not only the non-synonymous variants but also the synonymous variants have been found to change the protein expression levels, conformation and function. Bioinformatic analysis of ADAMTS13 mRNA structure, amino acid conservation and codon usage allowed us to establish correlations between mRNA stability, RSCU, and intracellular protein expression. This study demonstrates that variants and more specifically, synonymous variants can have a substantial and definite effect on ADAMTS13 function and that bioinformatic analysis may allow development of predictive tools to identify variants that will have significant effects on the encoded protein.

AptaCluster --- A Method to Cluster HT-SELEX Aptamer Pools and Lessons from Its Application

Systematic Evolution of Ligands by EXponential Enrichment (SELEX) is a well established experimental procedure to identify aptamers - synthetic single-stranded (ribo)nucleic molecules that bind to a given molecular target. Recently, new sequencing technologies have revolutionized the SELEX protocol by allowing for deep sequencing of the selection pools after each cycle. The emergence of High Throughput SELEX (HT-SELEX) has opened the field to new computational opportunities and challenges that are yet to be addressed. To aid the analysis of the results of HT-SELEX and to advance the understanding of the selection process itself, we developed AptaCluster. This algorithm allows for an efficient clustering of whole HT-SELEX aptamer pools; a task that could not be accomplished with traditional clustering algorithms due to the enormous size of such datasets. We performed HT-SELEX with Interleukin 10 receptor alpha chain (IL-10RA) as the target molecule and used AptaCluster to analyze the resulting sequences. AptaCluster allowed for the first survey of the relationships between sequences in different selection rounds and revealed previously not appreciated properties of the SELEX protocol. As the first tool of this kind, AptaCluster enables novel ways to analyze and to optimize the HT-SELEX procedure. Our AptaCluster algorithm is available as a very fast multiprocessor implementation upon request.

Recent advances in (therapeutic protein) drug development

Therapeutic protein drugs are an important class of medicines serving patients most in need of novel therapies. Recently approved recombinant protein therapeutics have been developed to treat a wide variety of clinical indications, including cancers, autoimmunity/inflammation, exposure to infectious agents, and genetic disorders. The latest advances in protein-engineering technologies have allowed drug developers and manufacturers to fine-tune and exploit desirable functional characteristics of proteins of interest while maintaining (and in some cases enhancing) product safety or efficacy or both. In this review, we highlight the emerging trends and approaches in protein drug development by using examples of therapeutic proteins approved by the U.S. Food and Drug Administration over the previous five years (2011–2016, namely January 1, 2011, through August 31, 2016).

Aptamers as a Sensitive Tool to Detect Subtle Modifications in Therapeutic Proteins

Therapeutic proteins are derived from complex expression/production systems, which can result in minor conformational changes due to preferential codon usage in different organisms, post-translational modifications, etc. Subtle conformational differences are often undetectable by bioanalytical methods but can sometimes profoundly impact the safety, efficacy and stability of products. Numerous bioanalytical methods exist to characterize the primary structure of proteins, post translational modifications; protein-substrate/protein/protein interactions and functional bioassays are available for most proteins that are developed as products. There are however few analytical techniques to detect changes in the tertiary structure of proteins suitable for use during drug development and quality control. For example, x-ray crystallography and NMR are impractical for routine use and do not capture the heterogeneity of the product. Conformation-sensitive antibodies can be used to map proteins. However the development of antibodies to represent sufficient epitopes can be challenging. Other limitations of antibodies include limited supply, high costs, heterogeneity and batch to batch variations in titer. Here we provide proof-of-principle that DNA aptamers to thrombin can be used as surrogate antibodies to characterize conformational changes. We show that aptamers can be used in assays using either an ELISA or a label-free platform to characterize different thrombin products. In addition we replicated a heat-treatment procedure that has previously been shown to not affect protein activity but can result in conformational changes that have serious adverse consequences. We demonstrate that a panel of aptamers (but not an antibody) can detect changes in the proteins even when specific activity is unaffected. Our results indicate a novel approach to monitor even small changes in the conformation of proteins which can be used in a routine drug-development and quality control setting. The technique can provide an early warning of structural changes during the manufacturing process that could have consequential outcomes downstream.

Pharmacogenetics and the immunogenicity of protein therapeutics

volume 29 number 10 october 2011 nature biotechnology To the Editor: The development of anti-drug antibodies (ADAs) to therapeutic proteins can lead to adverse events and also make a biologic less effective for its intended use. Immunogenicity assessments are now critical for the development, regulatory licensure and use of biologics and it has been argued that it is highly unlikely that regulatory approval would be granted for a biologic without an assessment of its immunogenicity1. The development of ADAs does not necessarily affect the safety and/or efficacy of a protein therapeutic and thus risk-based approaches are generally advocated to evaluate the clinical consequences2. For example, although therapies that involve replacement of proteins that are lacking or nonfunctional in patients have had spectacular results in the clinical management of many chronic diseases, such therapies are particularly prone to have adverse consequences as a result of so-called neutralizing ADAs3. Several reviews have previously catalogued the product and patient-related risk factors for immunogenicity (refs. 4,5 and references therein and Supplementary Fig. 1). The past decade has seen a steady shift in the use of recombinant human proteins, substantial improvements in quality and less heterogeneity in therapeutic protein products owing to the adoption of strategies such as quality by design6. Even so, improvements in product quality cannot control for the genetic variability of the patient population due to which some individuals, racial and/or ethnic groups or other subpopulations develop inhibitory antibodies at a higher frequency than others. In this article, using the example of factor VIII in the treatment of hemophilia A, we propose that a pharmacogenetic approach may be an important factor in the accurate prediction of immunogenicity of some categories of therapeutic proteins. Recombinant human protein drugs are mostly recognized by the body as ‘self ’; tolerance of recombinant proteins can be affected, however, by two key differences from the corresponding endogenous protein: (i) mutations in the endogenous protein that render it defective and (ii) the occurrence of nonsynonymous (ns) single-nucleotide polymorphisms (SNPs). Genotyping the endogenous (albeit nonfunctional) protein in a patient can identify regions of the infused protein that are ‘foreign’ to that individual (Fig. 1 and Supplementary Fig. 2a). Thus, in the case of factor VIII, consistent with the precept that synthesis of the factor VIII polypeptide chain is necessary for inducing central tolerance, the nature of the mutation in the patient’s factor VIII gene (F8) should be a good predictor of the frequency of factor VIII inhibitor formation. Hemophilia A patients with missense mutations in F8 develop inhibitors with a frequency of ~5%, whereas the rate of inhibitor development for patients with large gene deletions has been reported to be as high as 88%7. Nonetheless, even within each class of patients, there is variability in the immune response, that is, some patients with a missense mutation develop inhibitory antibodies, whereas a fraction of patients with large deletions do not develop inhibitors. SNPs are the most common source of genetic variation in the human population8 and a recent report has investigated the effects of ns-SNPs on factor VIII immunogenicity9. This study demonstrates the presence of several ns-SNPs in F8 that result in primary amino acid sequence mismatches between the infused factor VIII and the endogenous factor VIII protein of some patients with hemophilia A. Large differences in the frequency of inhibitor development between patients of white-European and black-African descent may be traced to distinct populationspecific distributions of these ns-SNPs10. Concomitantly, there is both indirect and direct evidence that the CD4+ T-cell response is essential for the development of inhibitory antibodies, and recent studies have been successful in identifying T-cell epitopes on the factor VIII protein11,12. On the basis of the above findings, we suggest that a sequence mismatch between the endogenous (tolerizing) peptides and those derived from the infused protein-drug may be exploited as a basis for understanding the pharmacogenetics of immunogenicity. Additionally, a critical determinant for T cell–dependent alloimmunization in an infused protein is the strength at which any foreign (non-self) peptide(s) derived from it (potential T-cell epitopes) binds to one or more of the distinct major histocompatibility complex (MHC)-II molecules on the surface of an individual patient’s antigen-presenting cells13. MHC-II proteins are extremely polymorphic and their distributions also exhibit clear racial and ethnic differences14 (Supplementary Fig. 1c). Thus, even identical non-self peptides will interact differently with the MHC-II repertoire of different patients, for example, binding with high affinity to an MHC-II protein in one individual while not binding at all in another. Several studies have endeavored to associate the nature and location of hemophilia A– causing mutations to immunogenicity, and whether particular human leukocyte antigen (HLA) alleles occur more frequently in individuals who develop inhibitory ADAs. In the past, however, these parameters have been considered independently. Here, we propose that all three parameters, mutations (as well as SNPs) in factor VIII, HLA type and sequence of factor VIII infusions be determined in individual patients. The decision making is hierarchical (Fig. 1) based, first, on determining the regions of sequence mismatch between the endogenous and infused proteins; next, on whether the peptides that represent the sequence variation bind to that patient’s MHC class II proteins (http://tools.immuneepitope.org/analyze/ html/mhc_II_binding.html); and, finally, whether the peptide-MHC interaction elicits T-cell responses. Such a strategy unfortunately cannot be fully validated using clinical data that are currently available. For example, the Hemophilia A Mutation Database (Hemophilia ADB15; http://hadb. org.uk), which is the most comprehensive database of F8 mutations and repository Pharmacogenetics and the immunogenicity of protein therapeutics CHMp/EWp/433478/2010. <http://www.ema.europa. eu/ema/index.jsp?curl=search.jsp&q=22+July+2010. +EMA%2Fchmp%2Fewp%2F433478%2F2010&mu rl=menus%2Fregulations%2Fregulations.jsp&mid=> 12. Finke, L.H. et al. Vaccine 25, B97–B109 (2007). 13. Hoos, A. et al. Semin. Oncol. 37, 533–546 (2010). 14. Marshall, M., Ribas, A. & Huang, B. Evaluation of baseline serum C-reactive protein and benefit from tremelimumab compared to chemotherapy in firstline melanoma. Abstract no. 2609, presented at the 2010 annual meeting of the American Society Clinical Oncology, Chicago, IL, June 4–8, 2010. 15. Robert, C. et al. N. Engl. J. Med. 364, 2517–2526 (2011). 16. Taylor, C.F. et al. Nat. Biotechnol. 26, 889–896 (2008). 17. Finn, O.J. N. Engl. J. Med. 358, 2704–2715 (2008). 18. Schreiber, R., Old, L.J. & Smyth, M.J. Science 331, 1565–1570 (2011). 19. Zitvogel, L., Kepp, O. & Kroemer, G. Nat. Rev. Clin. Oncol. 8, 151–160 (2011). CORRESpONdENCE