Proteomics 2.0 and Its Impact on Drug Development

Mo Jain is the founder and chief scientific officer of Sapient.

While genomics has proved transformative in decoding the genetic underpinnings of disease and remains a cornerstone for drug development today, population-scale discovery has been notably shifting in recent years “beyond the genome,” focusing on complementary data sources that may be leveraged to accelerate the drug development process. This shift has been catalyzed by the observation that, given the largely static nature of the human genome (which is set from conception), genomic measures capture only a minority fraction of population attributable risk for most common human diseases, from cancer to autoimmune disorders.¹

Genetic Code Versus Zip Code

It has long been appreciated that dynamic non-genetic factors, including external factors such as diet, lifestyle, environment, and toxicants, as well as internal factors including organ physiology, cell-to-cell communication, and tissue repair and remodeling, may have far greater influence on overall health outcomes. As such, there is increasing interest in generating expanded omics datasets, including through transcriptomics, proteomics, and metabolomics, which can read out these effects on health status and disease pathobiology over time.

Proteins as Drug Targets

The vast majority of therapeutic agents work by binding to a protein target and modulating its functionality. Genomics or transcriptomics data have traditionally been used as surrogates to understand proteins that may be altered in a disease state, but across the ~20,000 protein groups that comprise the human cell, there is actually limited correlation between presence of a genetic mutation and its protein level expression—and this protein disassociation is only exacerbated in disease states.² Simply put, genomics and transcriptomics as surrogate measures of protein levels are in many cases insufficient to determine phenotypes.

Direct Measures, Dynamic Insights

Proteomics—the direct measure of proteins expressed by the genome to assess their presence, structures, functions, variants, and interactions—enables us to more directly and deeply understand the dynamic processes contributing to disease and drug response. It provides a comprehensive view of protein activity and the protein modifications that also contribute to and influence the cellular environment. Through proteomics, we can identify functionally relevant drug targets and biomarkers, elucidating the underlying mechanisms of diseases to inform therapeutic strategies.

Initial proteomic surveys of disease have only further confirmed this thesis, from the identification of more than one hundred thousand protein-disease associations to new clinical insights into pharmacotherapeutic mechanisms of action.^3,4 The expansion of proteomics analysis as a cornerstone of drug development, however, has been limited by technological hurdles, including the necessary sensitivity, specificity, and speed of measure needed to perform comprehensive protein measures at scale, across more studies and sample matrices, until quite recently.

Innovations Enabling Proteomics 2.0

Proteomic technologies and methodologies are today evolving at a rapid pace, progressively unlocking our ability to study a broader landscape of proteins and their isoforms. These key innovations have converged to transform the utility of proteomics in drug discovery and development processes in just a short time and have important implications for drug discovery and development processes by enabling the following.

Significant sensitivity gains

A key physical challenge in proteomics is the wide dynamic range of protein concentrations in biological systems, particularly within the plasma proteome.⁵ High abundance proteins, such as albumin or fibrinogen, are present at approximately a billion-fold higher concentration than lower abundant proteins, thereby masking the measure of key proteins present in lower concentrations, including tissue injury and signaling proteins. While traditional targeted protein assays such as the enzyme-linked immunosorbent assay (ELISA) use antibodies that can amplify the signal of low abundance proteins, they are limited to measuring at most dozens of proteins at a time.

Traditional mass spectrometry-based methods, which directly measure proteins via direct peptide sequencing, can alternatively be used in a nontargeted approach to broadly assay more proteins per sample. This includes important proteoforms and post-translational modifications (PTMs), which can provide further insight into functional mechanisms of the protein and can represent potential drug targets. The challenge to date has been that mass spectrometry has lacked the sensitivity to capture the low-abundance portion of the proteome, with signals from abundant proteins dominating the readout, even with extended analytical cycle times.

As a means to enable greater coverage breadth, affinity-based methods for global proteome profiling have emerged. These approaches use a mixture of antibodies (e.g., Olink™) or aptamers (e.g., SomaLogic™) to bind and quantitate hundreds to thousands of proteins per plasma sample. Method advancements over the last decade now enable high-sensitivity measurement of more than 5,000 proteins in plasma. However, affinity-based approaches are largely limited to measuring whole native proteins and cannot capture proteoforms and PTMs without the development of additional antibodies or aptamers for each specific isoform of interest.

This is where recent innovations in mass spectrometry technologies and methodologies have driven a step change in protein measurements, closing the gap or even exceeding the breadth of protein measures relative to affinity-based methods by enhancing sensitivity while maintaining a high degree of accuracy and precision. Key advances have come in upfront protein preparation, most notably in the development of nanoparticle enrichment techniques to isolate proteins from complex biospecimens. These systems “compress” dynamic range by depleting high-abundance proteins while at the same time enriching low-abundance proteins, allowing for a greater number of proteins and their isoforms to be more evenly detected across the proteome. Now, mass spectrometry-based proteomics methods can offer protein coverage exceeding 5,400 protein groups per plasma sample and more than 12,000 protein groups in tissue and cells. To maintain consistency in protein counts across large studies, state-of-the-art nanoparticle systems employ specific surface chemistries that allow the nanoparticle to “capture and hold” the proteins tightly in place. This minimizes protein drop-off during sample preparation and processing steps to mitigate sample-to-sample variance and deliver highly reproducible measures.

Specificity at scale

When it comes to affinity-based proteomics, questions regarding the accuracy and specificity of protein measures continue to exist because they are indirect measurements through binding agents. Measurement specificity has always been a hallmark of mass spectrometry, as the technique involves direct analysis and sequencing of enzymatically-derived protein fragments known as peptides. Such peptide sequencing allows for the precise identification of proteins and their proteoforms, including protein variants and PTMs. Mass analyzer technology innovations such as trapped ion mobility continue to increase the number of peptides that can be captured per protein and thereby provide more points of quantitation.

Pairing this next-generation mass spectrometry instrumentation with optimized nanoparticle enrichment techniques, we can now achieve specificity at scale, delivering robust quantitative measures across more proteins, including exosomal proteins, extracellular vesicles, PTMs, and proteoforms.

Mass spectrometry also offers an orthogonal approach to cross-validate protein measures made via affinity-based proteomics workflows, to confirm if the antibody or aptamer did in fact bind to the correct protein.⁶ For proteins identified via affinity proteomics, mass spectrometry can be deployed to further investigate those proteins’ PTMs and proteoforms for more granular understanding of protein functions and disease mechanisms.

Accelerated speed to insights

While hardware innovations have transformed the speed at which samples can be processed and mass spectrometry data can be collected—allowing for deep proteome coverage with precision and accuracy at high throughput—equally important is the software innovations that ensure that the data can be efficiently handled and analyzed. Cloud computing, paired with AI and machine learning (ML) tools that support workflows from spectral analysis to peptide mapping, are allowing rapid interpretation of population-scale datasets that drive statistically powerful investigations into human disease pathology and therapeutic development.

Taken together, proteomics workflows can now provide the sensitivity, specificity, and speed that drug developers need to effectively deploy protein-based biomarker strategies in their pipelines. Enhanced specificity and sensitivity allow for the identification of novel dynamic biomarkers and therapeutic targets that were previously undetectable. The ability to analyze PTMs and proteoforms provides deeper insights into disease mechanisms, enabling the development of targeted therapies that modulate specific protein functions. Improved analytical speed facilitates large-scale studies and the integration of proteomic data into clinical workflows, supporting personalized medicine approaches.

As Proteomics 2.0 continues to evolve and its applications continue to expand, it is clear that these orthogonal data streams will be integral to accelerate drug discovery, development, and deployment and central to the future of precision medicine.