In the mid-19th century, Rudolph Virchow first proposed the cellular basis of human disease . Since then, we have gained a detailed knowledge of how cells operate in general and how molecules such as DNA, RNA and proteins are key regulators of cell function. Using this knowledge to determine what goes wrong in each and every disease or, particularly, to understand the variation in cellular dysfunction between different individuals with the same condition is an area in which we are still just ‘scratching the surface’. Yet, as the underlying biology is a major determinant of disease phenotype and clinical outcome, it is of prime importance to unravel disease mechanisms and translate this knowledge into new diagnostic and therapeutic tools. Turning biology into meaningful clinical applications is the basis of translational research.
So, where are we at present? On the one hand, we are overwhelmed by reports in the literature describing much success in the laboratory using molecular tools; on the other hand, decades-old textbook wisdom prevails in everyday clinical practice. How is this possible?
Translational research is a time-consuming, complex, difficult and not always exciting task. The level of collaboration, persistence and coordination needed make that many researchers continue with the next proof-of-concept study in the laboratory rather than try to translate their findings into clinical applications. The causes of this problem are multifactorial, but one is the lack of availability of the right tools. In the current issue of the Journal of Internal Medicine, Pontén et al. provide an overview of an enormous enterprise to build such a tool, of which the relevance for translational research cannot easily be overestimated. The Human Protein Atlas (HPA) project, which started in 2003, aims to generate specific antibodies for the major transcripts coded by all the genes (about 20 000) in the human genome .
With the human genome project and the wide availability of techniques such as PCR, microarrays and next-generation sequencing, we have witnessed a revolution in genome research over the last two decades. Thus, at the DNA and RNA levels, we can now understand the biology of diseased cells with a degree of detail and magnitude that is hard to perceive and also to scientifically digest. At the protein level, however, we are still lagging behind with the availability of tools, and initiatives such as the HPA play a major role in approaching the situation with genomics. Methods such as mass spectrometry increasingly allow the spectrum of proteins present in diseased cells to be determined but are still in the early days of having the capability for high-throughput validation of many samples for specific proteins of interest; however, this type of validation is of key importance in translational research. This is where antibodies, and thus the HPA, come in. The HPA initiative is unprecedented in this sense, because of the high level of quality control and quality assurance, the level of annotation and documentation and last but not least the portal by which all of this information is accessible to the whole scientific and clinical community.
Another reason to value the HPA project is that proteins as molecules currently provide many more opportunities for developing therapeutic and diagnostic tools than molecules such as DNA and RNA. Almost all drugs interfere with proteins, and hardly any with specific genes at the DNA or RNA levels. Assays that measure specific proteins can be turned into low-cost, point-of-care, diagnostic tests that can even be available over the counter. Pregnancy tests and faecal immunochemical tests for colorectal cancer are well-known examples of such protein assays.
The goal of the HPA is ultimately to generate and validate antibodies against all human proteins, i.e. against at least one major isoform of each protein-encoding gene. In addition, the aim is to make available to the scientific community all information generated on the expression of these proteins in healthy and diseased tissues, in a way similar to genome browsers such as Ensembl and NCBI. The report by Pontén et al. illustrates that the project is well on its way in achieving these goals, and already many researchers have gratefully used the results of the HPA project, often in combination with other proteomic and genomic analyses . An additional goal of the HPA is to identify and explore the potential of the antibodies generated for use as biomarkers of clinical relevance, for example to personalize therapy. This latter goal may be the most challenging. A sensitive and specific antibody is not automatically a biomarker. An established and clinically applicable biomarker is essentially an extensively validated (both technically and clinically) diagnostic test that discriminates between two different phenotypes that should receive different therapies. For a single biomarker, this process can easily take up to a decade or more. To do this for the thousands of antibodies generated by the HPA consortium is obviously more than a single project can accomplish, even when it is as well organized as the HPA. This is clearly a task in which the translational research community should participate, and this is facilitated by the open-access policy of the HPA.
Human biology is highly complex, and any project that aims to unravel this complexity must acknowledge this fact; this also applies to the HPA project. The aim of the HPA is to generate antibodies against at least one major transcript of every human gene. Here, the problems of complexity already kick in because it is not as simple as one gene, one protein. On the contrary, a single gene may code for multiple variants of proteins, often with different functions, and in this way, the 20 000 genes may code for many hundreds of thousands of proteins. In cancer, this is further complicated by genomic rearrangements (e.g. translocations) that result in new, cancer-specific, proteins. The BCR/Abl fusion protein is a prominent example of this. Recent advances in next-generation sequencing that allow sequencing of complete cancer genomes have shown that these types of rearrangements are much more common, also in epithelial cancers, than originally thought. This field is basically not covered by the HPA and therefore may qualify for a new cancer protein project.
Having said that, the HPA as it is now already is an enormous asset to biomedical research, and the challenge to the scientific community is to use this to the benefit of the patient. Leveraging the full potential of our increased understanding of disease biology and translating this into meaningful clinical applications will only work when physicians and biomedical researchers of different disciplines join forces and actually conduct studies such as biomarker-driven clinical trials in a much more coordinated way and at a higher pace than we are doing now. The HPA is ready for this.