July 1, 2016 (Vol. 36, No. 13)
Dominic Zichi Director of Bioinformatics SomaLogic
Sheri Wilcox Senior Director, Discovery Sciences SomaLogic
Jeffrey Walker Director, Strategic Alliances and Technology SomaLogic
Nebojsa Janjic CSO SomaLogic
Elucidating Biology with a Highly Multiplexed Proteomics Platform
Proteomics is the comprehensive study of protein expression. The goal of proteomics today is to enable a large number of protein measurements on a large number of samples, analogous to now routine large-scale genomic measurements. The SOMAscan® assay is helping realize this goal.
Mass spectrometry (MS) and antibody-based techniques are dominant proteomic technologies, but both suffer from challenges for large-scale protein profiling. Because most proteomic studies are aimed at complex mixtures (blood, for example) the relatively high abundance of a few proteins can lead to biased results. We refer to this as the “albumin/IgG problem,” since albumin and globulins account for over 93% of proteins in blood.
Discovery MS can identify large numbers of proteins and provide relative changes in abundance for small numbers of samples, but limited sensitivity results in biases toward high-abundance proteins. MS also suffers from low throughput, relatively poor reproducibility, and limited dynamic range.
Antibody-based technologies are ubiquitous in the field. These techniques are able to detect low abundance proteins (sub-nanomolar). Antibodies are also adaptable to measurements in complex matrices, but individual antibodies can be subject to nonspecific binding. Even high affinity antibodies will bind nontarget proteins in a matrix where the target is dramatically less abundant.
ELISAs overcome this issue by using two antibodies in a sandwich format; an initial antibody captures the target and a second antibody detects. The specificity of each antibody helps isolate the target among abundant nontarget proteins. While the sandwich method proved to be a helpful advance, it no longer meets the needs of proteomic researchers wishing to perform highly multiplexed, large-scale profiling studies. The main reason is that ELISAs do not scale well in a highly multiplexed assay due to nonspecific binding of both capture and detection antibodies.
A Better Way to Profile Proteins
The SOMAscan assay allows for simultaneous detection of thousands of proteins (multiplexing) while assaying large numbers of samples (throughput). In its current version, the assay can measure 1,310 protein targets from as little as 65 µL sample (serum, plasma, cerebrospinal fluid, tissue, cells, and other complex matrices). The low sample volume requirement maximizes the utility of precious clinical samples. In addition to this high-level of multiplexing, the assay covers a large dynamic range detecting proteins from femtomolar to micromolar concentrations.
A Novel Reagent
The SOMAscan assay uses a unique set of protein-binding reagents aimed at conformational epitopes to achieve large-scale multiplexing and throughput. SOMAmer® reagents (Slow Off-rate Modified Aptamers) are single-stranded DNA constructs containing modified functional groups mimicking amino acid side chains involved in protein binding. Like high-affinity monoclonal antibodies, these reagents allow for affinity capture with outstanding specificity to the given protein epitope. Since recognition is directed to conformational epitopes, binding to shared epitopes across highly conserved proteins can occur.
SOMAmer reagents derive their high affinity and specificity from functional group and shape complementarity to protein epitopes. The novel protein-like side chains introduced into these reagents allow the SOMAmer DNA backbone to adopt unique secondary and tertiary structures and facilitate tight binding to protein targets.1,2 The complementary interactions with protein targets and the unique SOMAmer structures adopted by these molecules are illustrated (Figure 1).
Each SOMAmer reagent is selected through an iterative in vitro evolution method known as SELEX (Systematic Evolution of Ligands by EXponential enrichment). Selection pressure enriches for both high affinity and slow off-rates that are key to SOMAscan assay performance. Once selected, the sequence is chemically synthesized using standard solid-phase techniques, generating a uniform and reproducible customized reagent.
SOMAscan Assay for Large-Scale Proteomics
The SOMAscan assay is a two-catch assay that overcomes the “albumin/IgG problem” by combining the initial binding with “kinetic specificity enrichment” to measure low abundance analytes in complex matrices. This format allows for two elements of specificity, analogous to an antibody sandwich assay, using a single reagent. Because it relies on a single affinity reagent and dissociation kinetics (slow for specific and fast for non-specific interactions), the SOMAscan assay scales easily to thousands of simultaneous measurements that perform like individual ELISAs.1,3
To perform the assay, SOMAmer reagents are immobilized to a streptavidin-bead using a biotin on the reagent. The bead-bound reagents are then mixed with biological samples allowing proteins to bind during catch-1, Figure 2A,B. Unbound proteins are washed away before bound proteins are tagged with biotin, Figure 2C.
A photocleavable linker then allows light-activated release of the complexes into a buffer formulated to reduce nonspecific complexes, Figure 2D. Since all SOMAmer reagents are oligonucleotides, they are inherently negatively charged. A polyanionic competitor facilitates the “kinetic specificity enrichment” that reduces weakly bound, nonspecific complexes, Figure 2E.
The biotin on the protein is then used to immobilize high-affinity complexes that remain after kinetic challenge in catch-2, Figure 2F, and unbound SOMAmer reagents are washed away. At this point beads retain complexes of proteins and SOMAmer reagents. In the final steps of the assay, SOMAmer reagents are dissociated from these complexes, Figure 2G, hybridized to complementary sequences on a microarray, Figure 2H, and quantified by fluorescence.
The fluorescence signal for each SOMAmer reagent is directly proportional to the amount of protein in the initial sample. This transformation from protein to nucleic acid signal is made possible by a special feature of nucleic acid ligands: the same unique sequence determines both folding into a 3D structure with precise shape complementarity to its target, as well as hybridization to complementary probes.
Large-Scale Protein Profiling
The advent of large-scale proteomic profiling opens a world of applications in biomedical science just as large-scale genomic profiling did. Utilities range from elucidating fundamental biology questions like disease natural history to biomarker discovery for diagnostics, drug development, discovering new drug targets, and elucidating mechanisms of action.
A powerful example of large-scale proteomic profiling is a study of Duchenne Muscular Dystrophy (DMD) comparing blood profiles from DMD and control subjects to find new therapeutic targets in this disease.4 Results comparing ~4,000 proteins between DMD and control subjects (Figure 3) revealed a wealth of differential expression.
In another notable example, large-scale proteomic profiling discovered a circulating protein, GDF11/8, in young mice that reduced cardiac hypertrophy in old mice.5 This discovery sparked a wave of interest and research to understand complex biology with broad implications in musculo-skeletal disease.
A final example is a recent study (in press) of cardiovascular disease that ranks among the largest proteomic profiling studies ever conducted. Blood from 2,700 case and control samples were profiled, collecting more than three million measurements in a matter of weeks. The results were used to develop a promising biomarker test to manage patients with previous cardiovascular events based on risk of suffering another event.
Given that the bulk of biological function is manifested by changes in proteins, proteomics holds great potential to impact numerous applications. This potential has yet to be fully realized largely due to limitations in existing proteomics technologies, particularly the need to choose between a technique that measures many proteins in a few samples and one that measures a few proteins in many samples. The SOMAscan assay addresses these limitations by allowing rapid profiling of thousands of proteins in thousands of samples, providing a powerful tool for better understanding of biological processes with broad utility in biomedical research.
1. Rohloff et al. (2014) Nucleic Acid Ligands with Protein-like Side Chains: Modified Aptamers and their Use as Diagnostic and Therapeutic Agents, Molecular Therapy – Nucleic Acids 3, e201.
2. elinas et al. (2016) Embracing proteins: structural themes in aptamer-protein complexes, Current Opinion in Structural Biology 36, 122–132.
3. Gold et al. (2010) Aptamer-Based Multiplexed Proteomic Technology for Biomarker Discovery, PLoS One 5, e15004.
4. Hathout et al. (2015) Large-scale serum protein biomarker discovery in Duchenne muscular dystrophy, Proceedings of the National Academy of Sciences 112, 7153-7158.
5. Loffredo et al. (2013) Growth Differentiation Factor 11 Is a Circulating Factor that Reverses Age-Related Cardiac Hypertrophy, Cell 153, 828-839.
Dominic Zichi (firstname.lastname@example.org) is director of bioinformatics, Sheri Wilcox is senior director, discovery sciences, Jeffrey Walker is director, strategic alliances and technology, and Nebojsa Janjic is CSO at SomaLogic.