Cardiovascular Proteomics Grows Up: Will It Move Out Of The Basement?

The late teens are often when a young adult chooses between two paths. In one scenario, energized by new abilities and burgeoning independence, he runs amok, making bad decisions and creating havoc and embarrassment for friends and family. In the other, she wisely harnesses the same adolescent liveliness and stamina to advance confidently in new directions, perhaps even doing things previously not possible. In their AHA statement (1), Lindsey and colleagues identify this critical juncture for the technologies and concepts of proteomics as applied to cardiovascular disease, making a sound case that the latter path is being taken by this adolescent.

The mysteriousness of the protein—in all its biochemical uniqueness, affection for interaction with other molecules, recalcitrant reproduction (there is no “PCR for proteins”) and proclivity for modification—is reviewed in detail, serving to remind the reader why there will never be a next-generation sequencing (or even Sanger sequencing) equivalent, one-size-fits-all technique, for proteomics. As a result, innovations in protein chemistry and most importantly, mass spectrometry, continue to drive biological inquiry. The time has passed when a mass spectrometry result in the main text of a paper on basic biological mechanisms or clinical research is strange; in fact the opposite is now true, wherein mass spectrometry is regularly employed towards a multitude of protein-related questions across biological systems, in what may be the most important innovation to have come from the last two decades of intense research in proteomics.

One of the first people to the microphone at early proteomics meetings almost always asked: “does the proteomics data confirm the microarray data.” While the questions about the relatedness between transcripts and proteins have become more nuanced, it turns out this question is still actively debated today, with the resounding answer being: it depends. Hopefully what we have learned in the intervening years is that this question need no longer be about the abilities of the respective techniques: notwithstanding differences in dynamics range, protein measurements, like those of their nucleotide counterparts, can now be robustly quantitative, and journals reviewers must continue to demand that proteomic experiments are done (and reported) in such a manner. Many of the present unknowns in proteome dynamics will be resolved through a combination of quantitation, including at the single molecule level, and biological perturbation. Virtuous adherence to quantitative measurement, including controls, in proteomic mass spectrometry can also help erode the western blot’s primacy in protein analysis…taking with it many of the false positive/negative results, and monochromatic view of the protein, inherent with westerns.

The purpose of science is to serve mankind: how is proteomics performing in this charge? There are two conceptually distinct ways to directly impact human health through translational research: improve prognosis/diagnosis (read: new biomarkers) or improve treatment (read: new mechanisms). It turns out that proteomic investigations in cardiovascular disease are making sound progress towards both these goals, revealing in the process that a decision between the aforementioned modes of research is perhaps a false dichotomy. Discovery research can identify both basic mechanisms of disease and targets for diagnosis in the same model system. It is also increasingly appreciated, as noted by Lindsey et al., that single molecules are unlikely to have the discriminatory power for improved diagnoses. An ongoing challenge for the application of proteomics to the clinic—indeed one that can only be surmounted by clinicians and discovery scientists working closely together—is to clearly frame the clinical question, such that the formidable technological capabilities of the proteomic machine can be wisely marshaled. A new frontier of personalized diagnosis awaits, if the techniques and clinical questions can finally be appropriately paired.

Regardless of the fairness therein, scientists, funding agencies and (ultimately) the taxpayer will continue to compare proteomics with genomics, including in the context of whether big or small science is likely to provide the next biomedical transformation. Your correspondent will now endeavor to do likewise in reflecting on what the proteomics community needs to do for the ongoing investments of public/private funds and research effort to be deemed worthwhile and thus to endure.

Data Driven Discovery

Many discovery-based proteomics papers end with a figure showing gene ontology analyses. As pointed out in a prescient paper (2) by Professor Michael Dunn in the early days of cardiovascular proteomics, “A major challenge will now be to investigate the contribution of these changes [detected by proteomics] to altered cellular functions resulting in cardiac dysfunction.” Professor Dunn was talking about the ~100 protein changes detected by the state-of-the-art 15 years ago—the challenge of determining what the proteins are doing in disease is even more daunting with today’s improved techniques. Novel computational strategies must be devised to interrogate the proteomes of the cell as stand alone entities; to let the data teach us what our previous studies of other tiers of biological information have missed.

Democratize Proteome Knowledge

The proteomics community needs to go further with allowing all analytical platforms to be open source, to enable simple data visualization tools to be freely available and to mandate mature data comparison/analysis tools be applied prior to, and continuing after, publication. For a long time the idea of an expiration date on the data, an anathema outside of proteomics, was an excepted (if tacitly) feature in the field resulting from changing sequencing databases and highly variable metrics for protein identification. We are now getting closer to an infrastructure that enables data memory, cross-laboratory comparisons and solidification of proteome knowledge.

Enable Other Approaches

The human genome project was sold to the public with the tagline it would cure diseases. Many scientists believed it would. It didn’t. However, the sequencing of the genome has revolutionized research, enabling proteomics and other technologies and changing our basic understanding of many diseases. This result was a terrific return on investment for the public. Like we learned with the genome, the proteome is not a fixed endpoint, and ongoing proteomic projects, grandiose or modest, must reveal new principles of normal biology or disease pathogenesis that likewise spur innovation in the technical realm.

Undertake Large-Scale Human Studies

The time is now for the proteomics community to devote more effort to large clinical cohort analysis. Given the well-documented challenges with translating discoveries in mice, the onus is on other approaches to demonstrate a better success rate. Given the privileged integrator function of the proteome—recording and processing the effects of common genetic variation and lifetime environmental risk—proteomic investigations in human populations have the potential to change the treatment disease. But these insights will only come if the studies are designed to reveal features that are clinically actionable, improving efficiency and/or reducing cost of care.

The perennial question of techniques or ideas shaping scientific progress continues to play out in proteomics. The scientific statement by Lindsey and colleagues does an excellent job of reviewing how the field has evolved, identifying the ongoing challenges and setting a path for the future. If cardiovascular proteomics is to move out of the basement, allowing it to continue to drive biological discovery and to become functioning part of clinical decision-making, technology development must maintain its intimate bond with biological and clinical theory.