© GUNILLA ELAM/SCIENCE SOURCE
There is a growing reproducibility problem across the life sciences. The retraction rate of published papers has increased tenfold over the past decade, and researchers have reported only being able to replicate published results in 11 percent1 or 25 percent2 of attempts. It’s become known as the “reproducibility crisis,” and science is in a race to fix it.
One major factor contributing to this problem is the use of poorly validated research antibodies. Lot to lot, antibodies can vary wildly. Some may not bind specifically to their target, or they may bind a different cellular protein altogether. According to one estimate, researchers around the world spend $800 million each year on poorly performing antibodies.3 (See “Exercises for Your Abs” here.)
While many researchers debate the best way to weed out the good antibodies from the bad, others are developing alternatives. Nucleic acid aptamers and protein scaffolds are increasingly being used to detect proteins of interest. Although they currently constitute only a fraction of affinity reagents, with the lion’s share of the market still going to traditional antibodies, these newer options offer an opportunity to rectify the problems stemming from using poorly validated antibodies in research. Researchers can engineer RNA or DNA aptamers and protein scaffolds to a specific target and function, the molecules are consistent from batch to batch, and they can be produced at a fraction of the cost of antibodies. These new reagents can target proteins that remain inaccessible to antibodies. And researchers have designed them to be functional in a wider range of conditions, including intracellular environments that degrade the antibody structure, opening up applications such as super-resolution microscopy and intracellular live-cell imaging to investigate the molecular dynamics of diverse cellular processes.
So, rather than complain about the poor performance of antibodies, perhaps the scientific community should embrace the new antibody alternatives designed to overcome this problem—and, by doing so, begin to resolve the ongoing reproducibility crisis.
The rise—and pitfalls—of antibodies
© STEVE GRAEPELAntibodies are large protein molecules composed of two heavy and two light chains linked by disulfide bonds. They play a crucial part in the immune system’s ongoing battle to keep our bodies from falling prey to deadly diseases. Through the diversification of gene segments in the antibody sequence, the mammalian immune system produces different combinations of heavy and light chains to bind a wide variety of foreign proteins. When an invader is detected, those B cells that produce the most specific antibodies undergo hypersomatic mutation to fine-tune the antibody’s affinity to a particular antigen, then differentiate into plasma cells that generate the targeted antibody molecules by the million to mark the disease-causing target for destruction. It has been estimated that the human body can create enough different antibodies to recognize 1012 distinct pathogens.4
For decades, life-science researchers have taken advantage of this natural process to develop tags and assays for a wide array of proteins. In the early 1900s, researchers began to cultivate protein-specific antibodies by immunizing rabbits, chickens, goats, donkeys, and other animals with a desired target protein. B cells within the animal host generate antibodies to different antigenic areas (epitopes) on the protein of interest. The antibodies targeting the desired protein can then be isolated and purified for use in biochemical and cell-based assays to document protein expression under different conditions or to identify potential disease biomarkers. But the reliance on an animal host system for production meant lot-to-lot heterogeneity for such polyclonal antibodies. (See illustration adjacent.)
In 1975, Argentine biochemist César Milstein and German biologist Georges Köhler discovered how to generate batches of individual antibodies, produced by a single B cell to target a specific antigen. Once an animal host produces antibodies to a target, the antibody-producing B cells are isolated from the spleen or lymph nodes and fused with tumor cells to generate immortal hybridoma lines. These lines are then screened to identify clones producing antibodies that bind with a high affinity to a specific epitope on the target protein. These cells are then cultured in large-scale bioreactors.
While heterogeneity can arise from drift in the cell line’s antibody expression and downstream production processes, monoclonal antibodies exhibit far less lot-to-lot variation than polyclonal antibodies, and have become the affinity tool of choice in modern research laboratories. Monoclonal antibodies are now routinely employed to localize proteins within tissues, determine protein network interactions, and analyze protein function. They are now being pushed to the limits of their performance in applications such as nanoimmunoassays and in vivo cell imaging. In medicine, antibody therapeutics represents the fastest growing sector of pharmaceutical sales, with 47 monoclonal antibodies currently on the market and a further 300 in clinical trials.5
© STEVE GRAEPEL
But there are many examples where the use of antibodies has actually hindered scientific progress, by providing misleading or inaccurate results. Antibodies have evolved to execute their biological function perfectly, but this does not make them foolproof as investigative tools or therapeutic agents. In fact, many of the very characteristics that aid in antibodies’ function as part of the immune system limit their use in research and medicine.
In the context of an immune reaction, for example, not all B cells produce antibodies that are exquisitely specific. So long as the antibodies exceed a certain threshold of binding affinity for the target, they remain part of the immune system’s defense. In the body, this is a good thing: these less-specific antibodies cross-react with a variety of related antigens, making the antibody defense force more versatile.6 If an invading pathogen mutates or a similar pathogen invades, potentially effective antibodies may already be in circulation. As part of an assay to specifically identify a particular protein, however, such cross-reactivity can be the downfall of the experiment or therapy.
Examined in this light, it is easy to see why taking a molecule that is derived for one purpose and applying it to another may not yield the best results. A clear example of the shortcomings of antibody use in life-science research comes from the Human Protein Atlas project. Mathias Uhlén of the Royal Institute of Technology in Stockholm, Sweden, and colleagues set out to catalog protein expression and localization data across 44 normal human tissue types, 20 different cancers, and 46 cell lines. The team sourced antibodies from 51 different commercial vendors for validation. Of the 5,436 antibodies received, about half failed to detect their target in a Western blot or standard immunohistochemistry assay.7
Nucleic acid aptamers and protein scaffolds can target proteins that remain inaccessible to antibodies, and researchers have designed them to be functional in a wider range of conditions.
The development of novel antibodies that bind new protein targets continues to face several challenges. For example, using the conventional route of immunizing lab animals to produce an antibody against a toxic target molecule will often kill the host animal prior to the generation of sufficiently specific antibodies. Conversely, if a protein target is highly homologous to a host protein, the immune system may not recognize the target as foreign in order to generate antibodies against it.
Given the rapid pace at which molecular biology proceeds in the modern era, new protein-binding reagents are desperately needed. A review of 20 million published research articles from 1950 to 2009 showed that three-quarters of the research focused on just 10 percent of the proteins that were known prior to the mapping of the human genome.8 Rational design of antibody alternatives will allow us to target a broader swath of proteins and function across more platforms to better investigate the scientific questions at hand.
Engineering a solution
© STEVE GRAEPELAlternative affinity reagents developed over the past few decades include both nucleic acid– and protein-based molecules. Aptamers are short molecules of single-stranded DNA or RNA, typically less than 100 nucleotides in length, that form 3-D structures capable of binding specific target proteins. Protein scaffolds, formed from polypeptide fragments or whole proteins, have similarly precise interactions with target molecules. Both types of affinity reagents are produced entirely in vitro, so in principle they are not subject to the limitations of antibody production by animal immune systems, allowing researchers to study proteins for which it is impossible to generate antibodies. And even when antibodies do exist, aptamers and protein scaffolds offer more-precise targeting, because they have been engineered for a specific purpose.
These novel affinity reagents also offer other benefits over antibodies. Both nucleic acid aptamers and protein scaffolds are much smaller than natural antibodies, which typically weigh about 150 kDa. Aptamers and scaffolds are as little as one-tenth that size. This means that their distribution is not restricted in the same manner as that of antibodies, opening up new targets that were previously inaccessible, such as epitopes hidden inside molecular grooves and pockets where antibodies simply can’t fit. Labeling target proteins with these smaller tags in cytochemistry experiments reduces the chance of the target protein being dragged around the cell according to the tag’s biochemistry, and increases the chances of identifying the correct protein localization. Additionally, their smaller size increases these affinity reagents’ tissue penetration, enhancing access to epitopes within tissue sections and decreasing false negative immunohistochemistry results. Smaller molecules are also cleared more rapidly from the body, especially when their size is below the renal cut-off of 45 kDa, making these molecules ideal as imaging agents in the clinic.
Researchers first developed nucleic acid aptamers in 1990 as RNA-based molecules, though DNA variants quickly followed to deal with the low stability of the RNA backbone. Aptamers offer simple chemistry that can be easily functionalized, but they lack stability across a range of temperatures and pH, and in the presence of common buffer components or DNases and RNases found in many media and cell environments. Researchers have used various chemical modifications to increase aptamers’ resistance to nuclease activity, to improve aptamer binding, and to increase their structural diversity, but others have turned to yet another option: protein scaffolds.
© STEVE GRAEPELDeveloped around the same time as aptamers, protein scaffold affinity reagents were originally designed to identify potential therapeutic targets. Researchers soon began to apply this technology to screening for binders to completely novel proteins, by presenting a random sequence as the binding surface. Because protein scaffolds lack the disulfide bonds of antibodies, they retain their structure in a greater variety of cell culture and assay environments, without being attacked by other proteins that break these bonds and cause antibodies to fall apart. Scaffolds maintain function and target affinity at temperatures up to 80 °C and in solutions with a pH as low as 2 and as high as 13.
Because protein scaffolds can be delivered to the inside of the cell, researchers can use them in live-cell imaging, ultimately allowing use of the same reagent in both biochemical and cell biology assays. Additionally, protein scaffolds could help to deliver drugs directly into cells, improving targeting of pharmaceutical payloads and reducing side effects. And as new protein scaffolds are often engineered to lack cysteine residues, aberrant folding during their production within the cell factory is unlikely, increasing reproducibility within the reagents.
Importantly, both nucleic acid aptamers and protein scaffolds are far easier to consistently produce than antibodies. In addition to requiring animal hosts to provide an antibody-producing B cell, functional antibodies can only be expressed in higher eukaryotic cell systems. Antibodies are extensively glycosylated with a complex range of sugars that are critical to their function. Lower eukaryotic organisms, such as insects and yeast, and prokaryotic cells are not capable of the full range of complex glycosylation. As a result of these complexities, production times for monoclonal antibodies are six months on average, often making the generation of new antibodies the rate-limiting step in the advance of new research. This production process is also extremely expensive, and the use of such intractable biological systems breeds batch-to-batch inconsistencies.
Aptamers and protein scaffolds can be made without a host immune system. Now that robust and scalable methodologies for creating custom DNA and RNA molecules exist, effective aptamer binders can be chemically synthesized at a fraction of the cost of producing protein-based affinity molecules. And because protein scaffolds do not contain any posttranslational modifications, they can be expressed in bacterial cells, which are cheaper and easier to control than the eukaryotic systems used for antibody production. Both aptamers and scaffolds can often be available to researchers in a matter of weeks. (See illustration above)
Alternatives at work
© AVACTA LIFE SCIENCES. May not be reproduced without express written permission from the copyright holder. So far, the majority of the industry attention for antibody alternatives has largely focused on the therapeutic development of antibody alternatives. Many companies now have initiated Phase 2 and 3 clinical trials of candidate molecules to treat conditions from vision problems to cancer, and two such molecules have already been approved for therapeutic use. In 2004, the RNA aptamer–based therapeutic pegaptanib (Macugen), originally developed by NeXstar Pharmaceuticals, became the first antibody alternative to gain US Food and Drug Administration (FDA) approval for the treatment of neovascular age-related macular degeneration. Pegaptanib is a 28-base-long RNA oligonucleotide with modifications to protect the aptamer from endogenous nucleases and extend its half-life in vivo to 10 days.9 Administered directly into the eye, this aptamer selectively binds the most common isoform of vascular endothelial growth factor (VEGF), preventing angiogenesis and the increased permeability of the blood vessels within the eye associated with neovascular age-related macular degeneration.10 Five years later, in 2009, the FDA approved a protein scaffold called ecallantide (Kalbitor) for the treatment of sudden hereditary angioedema attacks.
In a therapeutic context, the most important characteristic of aptamers and scaffolds is that they lack immunogenicity, thus avoiding harmful immune responses in patients. One way to ensure that protein scaffolds do not trigger host immunity is to model one’s scaffolds after proteins found in the human body. For example, the FDA-approved Kalbitor is based on the common Kunitz domain of protease inhibitors, and clinical trial participants have not suffered immunogenic responses. When brought to market, Kalbitor was one of only two approved therapies to treat cardiovascular attacks of this sort, which can cause rapid and serious swelling of the face or other parts of the body that may result in permanent disfigurement, disability, or death; the other is a protein therapeutic derived from human blood.
A major factor holding back the field of antibody alternatives as therapeutics is their small size. While this improves their intracellular function and use in research applications, their low molecular weight means that they are rapidly cleared from the body via the kidneys, reducing their potential therapeutic impact. Various strategies have been employed by the industry to overcome this rapid renal clearance, such as adding an antibody domain or an albumin-binding domain to the scaffold, or increasing the molecular weight of the protein scaffolds (though they still remain significantly smaller than a corresponding antibody). The fusion of an antibody domain to a protein scaffold can also help engage the immune system for improved therapeutic benefit.
While their small size can be a hurdle in developing antibody alternatives in a clinical setting, it is a big advantage in their use as laboratory tools, allowing them to penetrate bodily tissues that are inaccessible to antibodies and offering more-precise molecular labeling. In 2012, for example, Silvio Rizzoli of the European Neuroscience Institute and Center for Molecular Physiology of the Brain in Göttingen, Germany, and colleagues used 15 kDa aptamers to capture the dynamics of endosomal trafficking in live cells using super-resolution imaging.11 Doubling the molecular weight of the aptamer resulted in a substantial reduction in image quality, showing the importance of the small size of intracellular labels in accurate imaging of the intracellular space.
The increased intracellular stability of protein scaffolds as compared with antibodies is also critical to their function as research tools and offers potential therapeutic benefit for intracellular targets. For instance, protein scaffolds have been used to investigate the function of the small intracellular domain of a matrix metalloproteinase, which was shown to determine protein turnover to help regulate protein function in cell movement.12 Using antibodies in the reducing environment of the cell interior in such a study would be impossible.
When developing antibody alternatives for research, scientists are purposefully mimicking natural proteins to avoid interference by the immune system. For example, Janssen produces protein scaffolds called Centyrins that are based on the fibronectin glycoprotein of the extracellular matrix. Our own company, Avacta Life Sciences, recently introduced Affimer scaffolds, which are based on the cystatin protein family of common protease inhibitors. The use of consensus sequences from a number of species may allow these reagents to be used across a variety of different model systems.
Nucleic acid aptamers and protein scaffolds may also help fight emerging outbreaks of acute infectious disease. Examples of recent outbreaks that have caused considerable social, economic, and political stress are not hard to come by—SARS in 2003, the H1N1 flu pandemic in 2009, the Ebola crisis of recent years, and the continued emerging threat of MERS. It is impossible to predict such episodes, and alternative affinity reagents could be crucial tools in quickly stemming the spread of pandemic diseases. Screening libraries of 10 billion sequences can take as little as 7 weeks. And while the processes required for optimization, scale-up, and subsequent culture and validation of substantial quantities of the required affinity reagent remain to be explored, this all may take only a matter of months.
Over the past few decades, the rate of advancement of genomic technologies has outpaced proteomics. Yet it is the expressed protein within a cell, not the underlying genetic blueprint, that executes correct or aberrant function. In order to unify and make sense of the numerous data sets being produced, scientists need tools that enable the unraveling of proteomics. This requires affinity reagents that can specifically target individual protein isoforms and glycoforms, and that can tag all the proteins within a cell or organism. While the concerns over antibody irreproducibility are increasing, the solution may already be available.
Jane McLeod is a science writer at Avacta Life Sciences, where Paul Ko Ferrigno is the chief scientific officer. Avacta Life Sciences sells peptide aptamers, one of the main forms of antibody alternative.
- C.G. Begley, L.M. Ellis, “Drug development: Raise standards for preclinical cancer research,” Nature, 483:531-33, 2012.
- F. Prinz et al., “Believe it or not: How much can we rely on published data on potential drug targets?” Nat Rev Drug Discov, 10:712, 2011.
- A. Bradbury, A. Plückthun, “Reproducibility: Standardize antibodies used in research,” Nature, 518:27-29, 2015.
- B. Alberts et al., Molecular Biology of the Cell, 4th edition (New York: Garland Science, 2002).
- D.M. Ecker et al., “The therapeutic monoclonal antibody market,” mAbs, 7:9-14, 2015.
- N. Baumgarth, “How specific is too specific? B-cell responses to viral infection reveal the importance of breadth over depth,” Immunol Rev, 255:82-94, 2013.
- L. Berglund et al., “A genecentric Human Protein Atlas for expression profiles based on antibodies,” Mol Cell Proteomics, 7:2019-27, 2008.
- A.M. Edwards et al., “Too many roads not taken,” Nature, 470:163-65, 2011.
- J. Ruckman et al., “2’-Fluoropyrimidine RNA-based aptamers to the 165-amino acid form of vascular endothelial growth factor (VEGF165): Inhibition of receptor binding and VEGF-induced vascular permeability through interactions requiring the exon 7-encoded domain,” J Biol Chem, 273:20556-67, 1998.
- P. Sundaram et al., “Therapeutic RNA aptamers in clinical trials,” Eur J Pharm Sci, 48:259-71, 2013.
- F. Opazo et al., “Aptamers as potential tools for super-resolution microscopy,” Nat Methods, 9:938-39, 2012.
- R.D. Wickramasinghe et al., “Peptide aptamers as new tools to modulate clathrin-mediated internalisation—inhibition of MT1-MMP internalisation,” BMC Cell Biol, 11:58, 2010.