Wednesday, March 26, 2008

Protein Biomarkers part I: Fact versus Fiction

Preface to Biomarkers

Its difficult to know where to begin when discussing the nascent biotechnological sector known as Biomarkers. I think therefore it may be helpful for the uninitiated to give the term some context and definition. In its simplest definition, a biomarker is any substance which is used as an indicator of a biologic, or phenotypic state. Under this definition, biomarkers are used every day when physicians listen to a patient lungs, take blood pressure, perform standard blood titers to measure hematocrit levels or white blood cell counts, during urine tests for drug metabolites or diabetes and so on. Biomarkers can be DNA, such as the gene which signifies Huntington's disease or any other genetic disorder, or RNA in expression profiles. Also of note, the concept of biomarkers is a cornerstone of the larger topics of personalized medicine and preventive medicine.

This being said, the term biomarkers has been expicitly defined by the National Institutes of Health
A characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes or pharmalogical responses to a therapeutic intervention
Therefore the term biomarker can be applied accurately in reference to individual genotypes. The most notable success in this area has been the selection of HER2 positive breast cancer patients during the acceleration of Roche's Herceptin through the clinical trial process. The power of DNA gene profiling, such as is done with microarray chips and/or genomic sequencing projects is enormous and ongoing. However, correlating DNA "biomarkers" (a.k.a. genes) to diseases is nothing particularily new and in this author's opinion is simply a semantic repackaging of efforts that began in the genomics era. In contrast, protein specific biomarker concepts are just now beginning to evolve, are grasped much less readily by the general public and are more easily leveraged with scientific knowledge. Therefore, keeping the broader definition in mind, the rest of the article will focus specifically on protein biomarkers and technologies.

The Promise of Biomarkers
Any scientist worth their salt will quickly find themselves inundated with conditional modifiers when writing about biomarkers. This is because the hype surrounding the field is historically paralleled only by the hype surrounding the sequencing of the human genome. Astute observers will also note that the excitement around biomarkers is justified in many ways by the same arguments used previously for the genomics bubble, which has of yet not exactly lived up to investors' expectations. Therefore it is convenient here to include a section describing what I will euphemize as the promise of biomarkers, in this way I hope to make plain that while scientifically exciting, the biomarker field is very much hypothetical from an investor standpoint.

Organic life, for the most part is comprised entirely of proteins. It is important to remember that in the human organism, roughly 21,000 genes encode at least one million different proteins. Theoretically speaking, every disease state imaginable will be able to be detected by either foreign or aberrant proteins, or abnormal protein levels. (The same can not be said for genes). Identification and characterization of any protein entity which can designate disease activity will result in a "bio"marker. Clinical detection of these protein signals can then in turn be used to monitor and gauge specific aspects of a given disease. Furthermore, any number of effective biomarkers may be used together in a panel to create a diagnostic "fingerprint" of the pathology in a single sample.

Biomarkers are an Intermediary between treatment and disease

These protein indicators have an incredible potential value in the understanding of the mechanism of disease progression. In addition, as more individual patients are screened using biomarker panels, different pathology "fingerprints" will likely divide many of the known disease categories into subtypes, with the result of much more medically descriptive diagnoses. This knowledge, and the ability to select between subtypes, can also allow the healthcare industry to expedite successful therapeutic development by selecting clinical subjects with those subtypes more responsive to a given treatment. At the same time, a decrease in negative incidents in the marketplace will be possible, if prescribing physicians have an analogous yardstick for pathology subtype treatment options. If these reasons aren't enough to get excited about the the prospect of biomarkers, the detection and identification of such proteins can also lead to preventative treatments, potential drug targets, diagnostic tools, and surrogate endpoints for clinical trials.

Protein Biomakers can gauge disease subtype and progression. This will allow therapeutic developers to:
  • Diagnose patients earlier and more accurately
  • Perform more directed clinical trials
  • Assess disease state primary endpoints with confidence
  • Decrease unresponsive or adversely affected patients in the marketplace
  • Potentially discover new therapeutic targets

In essence, the nascent field of biomarker discovery can be described as disease pathway research.
A validated biomarker is simply an identified protein constituent which is affected in some way by a pathological state. Some might point out that in fact, this has been the modus operandi of great many academic biomedical researchers for decades, and therefore the notion of "biomarkers" is more of a new buzzword than a research concept. This may be true with regards to academia. However insofar as the drug development and pharmaceutical industry is concerned, seeing real value in disease pathway illumination has only recently gained traction.

The Business of Serendipity
Everybody has a pet theory as to the causes of Big Pharma's lack of R&D productivity, and mine goes something like this. Until relatively recently, drug developers' business strategy relied on a blockbuster model as the goal, and academically validated targets as the starting line. They each made huge libraries of small molecule pharmacophores and screened the same targets. Eventually corporate risk aversion manifested itself in me-too therapeutics and huge marketing sales forces. The point here is that it wasn't (and perhaps still isn't) seen as scientifically or financially feasible for these companies to screen for novel effected proteins along distant and undiscovered disease pathway branches.

However, summarizing the recent prevailing expert opinions; so long as the pharmaceutical industry leaves the deepening of disease understanding to academia, the inevitable result will continue to be the same big pharma corporations screening the same published targets with most likely the same pharmacophore libraries (or more recently recombinant protein, or RNA libraries). This lack of scientific competitive advantage is one reason behind the recent trend of niche market focus in the pharmaceutical industry. Moreover, the revelation that disease understanding has intrinsic value via potential biomarker discovery, in addition to proprietary target identification, has cemented this strategy of intellectual capital at least amongst the pundits.

The fact that success in biological discovery most often proceeds by serendipity is probably only surprising to non-scientists. Nevertheless, investors need to weigh risks, and business plans necessarily must project return on investment. Therefore, short of cutting as broad a swath as NIH funding, the pharmaceutical industry needs to increase exposure to proprietary understanding of disease in a way amenable to statistical evaluation. Nothing fits this bill more perfectly than proteomics.

I assume that a number of my readers will be nonplussed to discover that the major thesis laid out at this point is that protein biomarker discovery is tantamount to proteomics. However the major objective in on my part has been to simply to disentangle the fact from the fiction and the jargon from the investment opportunities. To this end, I believe this quote sums up the recent proteomics history fairly well.
The age of proteomics has been announced in the mid 1990’s and since then a tremendous rush of proteome gold miners started grounding the claims. There were people claiming to (i) analyze the whole proteome, (ii) determine the interactome, (iii) determine the global phosphoproteome , (iv) determine the secretome, (v) and many more. Such buzzwords give the impression that with some mass spectrometers, computers, and algorithms we have the Holy Grail in our hand and will soon be able to elucidate the biology.
Now 10 years later, quite a number of studies are published every week dealing with high-throughput proteomics experiments. Very often, in such studies, the technical limitations are addressed or long lists of qualitatively identified proteins are reported, only.
So -- in fact, proteomics as a science and as a sector remains only in its infancy in terms of technological development necessary to elucidate the dazzlingly complex interactions of proteins necessary for biomarker discovery. However due to the somewhat dire situation of the drug discovery industry described above, coupled with the potential of protein biomarkers, there is little doubt that these nascent technologies have recently benefitted, and will continue to receive substantial investments in the coming years.

I'd like to break part I off here and let this article stand as a review to bring the unindoctrinated protein biomarker investor up to speed. In part II I plan to cover the vision of interplay between biomarkers and therapeutics and finally get down to which protein biomarker technologies I see as the most promising areas for investment.