Phorum

September 14, 2008

Bruce Tomczuk, PhD 

In the past entries, I have described my analysis of the woes of the pharmaceutical industry and I have detailed solutions, stressing the necessity to switch to a systems biology approach toward disease treatment.  In this entry, I ] extract examples from an article written by Lisa M. Jarvis in the September 1, 2008 issue of Chemical & Engineering News, on systems biology in action. 

In this article, Jarvis delineates systems biology approaches to a treatment of Cystic Fibrosis.  The first example describes the early work of Aurora Biosciences which was later acquired by Vertex, who continued the project.  The project was based on the discovery that CF is caused by a mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) gene.  This mutation causes an excess absorption and a decreased secretion of chloride ion which prevents the formation of a critical thin film of hydration.  Consequently, the lungs suffer a defect whereby dry pockets develop that get clogged with thick mucus and provide a breeding ground for bacterial infections. 

The Aurora project

Thus, the original Aurora project set out to discover “correctors”, that is, compounds that could increase the number of chloride ion channels as well as “potentiators”, compounds that could increase the flow of chloride ions through existing chloride ion channels.  Note, the phenotype descriptions in quotations!  The research strategy didn’t rely on the molecular mode of action of the compounds, just the resultant phenotypic action.

Cut to the chase, after several years of research, Vertex is currently conducting clinical trial with VX-809, a “corrector” in Phase I and VX-770, a “potentiator” in Phase II.  These compounds have the potential to not only provide symptomatic relief but also modify the disease.  The project had taken on a life of its own as Joshua Borger, CEO of Vertex, has stated that “[he] couldn’t stop this project if [he] tried”.  Preliminary clinical results of VX-770 have shown that after just two weeks of treatment, not only was lung function improved, but the amount of salt in their sweat was also lowered which demonstrates the systemic disease modifying effects of the drug.  Again, the exact molecular mechanism of action of VX-770 is not completely understood, but the importance is that the biological activity of the drug corrects the defect underpinning the disease. 

It is interesting to note that on a parallel track, the Vertex compounds have provided the tools to further dissect the nature of the disease: how the CF protein folds, which partners it folds with, and where it self-interacts.   

The PTC Therapeutics project 

A second project that highlights the wonders of systems biology is PTC Therapeutics.  Scientists with this project uncovered that in a smaller subset of CF patients, there is a “nonsense” mutation, a premature stop signal on the mRNA that causes the ribosome to fall off before the full CTFR protein is produced.  PTC Therapeutics used a systems biology approach to look for compounds that would allow the normal CFTR protein to be produced.  This approach was based on the fact that gentamicin and other aminoglycosides could already elicit such an effect, albeit weakly.  PTC utilized a HTS screen based on a cellular assay producing firefly luciferase from LUC nonsense-containing mRNAs expressed in HEK293 cells.  The screening hits were optimized to yield PTC124 via traditional medicinal chemistry techniques except that the SAR followed “nonsense suppression activity”.  Again, although the exact molecular mechanism of action of PTC124 is unclear, the compound appears to bind to the 28S ribosomal submit to induce a conformational change that enables the full translation of the CFTR protein. 

Encore 

These are just two examples that utilize a systems biology approach to successfully discover new drugs.  Some of these drug candidates have progressed to proof-of-concept in patients with potential for disease modifying treatments.  Similar approaches could be used for any disease, and I have expanded on some approaches in previous entries.  In my opinion, this is the strategy for successful drug discovery and one which must be implemented ASAP. 

Note: The author (BT) declares no financial ties to PCT Therapeutics or Vertex Pharmaceuticals, Inc.

Bruce Tomczuk
June 25, 2008

In the past entries, I have described my analysis of the woes of the pharmaceutical industry and I have detailed some solutions.  In this entry, I want to use all of the foregoing discussions in defining an ideal pharmaceutical company called ASAP Pharmaceuticals, Inc.

The basis premise is that diseases are complex and the pharma industry has oversimplified biology since the 1980’s when molecular biology had blossomed.  It is this reductionist strategy that has produced the dearth of drug approvals that are apparent today.  The pharma industry became enamored by the idea of one molecular target for one disease—specificity was, and still is, the cornerstone in most of the large pharma.  In reality, disease states are rarely caused by a single molecular target.

One order of Systems Biology, ASAP

The answer is to switch to a systems biology approach to disease modification.  Systems biology, in its simplest form, treats organisms as complex systems by mapping out, rather than reducing, biological complexity.  Thus, the cornerstone of ASAP Pharmaceuticals, Inc. is a group of biological mathematicians/ programmers called the “Disease Simulations Group”.  This group is charged with the description of diseases by mathematical models with all of the attendant complexity.  Sensitivity analysis of the models can aid in the identification of pressure points and limiting pathways. The subsystems involved in these pathways can be transformed into HCS (high content screens) using human primary cell- or mixed-cell culture systems by the “Cellular Biology Group”.

The Cellular Biology Group would be divided into 4 possible departments, which would focus on symptoms modification (SM), disease modification (DM), regenerative medicine (RM) and preventive medicine (PM).  The cellular biologists in these groups would be charged with working with the Disease Simulations Group to transform and to validate the perfected disease models into HCS.  The SM group would focus on general symptomatic treatments for example pain relief (analgesics).  The DM group would focus on diseases where the biological pathways have become unbalanced.  These include a broad array of diseases with large unmet medical need such as atherosclerosis, type II diabetes, and asthma.  The RM group would focus on disease states where this imbalance has caused “irreversible” structural damage to cells, organelles, or organs.¹  These diseases include myocardial infarction (MI), COPD, stroke, CHF, etc. The fourth biology department, preventative medicine (PM), is futuristic and would evolve from the improved biological understanding of diseases.  Thus, in the advent of consistent and/or progressive biomarker changes indicative of a disease phenotype, it would be feasible to correct the biological imbalance.

And all the rest….

The remainder of the pharmaceutical research enterprise—pharmacology, chemistry, drug metabolism and pharmacokinetics —would align themselves with the various disease groups.  Aside from the different screening technology, migrating from molecular target to cellular phenotypic readouts, which would necessitate a difference in chemistry structure-activity relationship readouts, the remaining sciences would basically remain the same.  Disease teams would be matrixed across these departments borrowing cellular biologists, pharmacologists, chemists, and drug metabolism scientists.

So what are we waiting for?

A new pharmaceutical model awaits us, but will require the ingenuity to interpret HCS and to decipher complex disease pathways.  If we make this paradigm shift in biology, the remainder of the drug discovery process will still need to be navigated.  It is a process that has never been easy, but will continue to be improved. 

Drug discovery depends very much on starting with good ingredients in order to have a good final product.  One of those ingredients is the biology (systems biology vs. molecular target) and the other is the chemistry (drug candidate and its properties).  We have possessed the capability of synthesizing a wide variety of small molecules for a long time.  In fact, a standing Medicinal Chemistry joke is that the chemists always make drugs, if only the biologists could figure out their use.  If we get on the right track with the first ingredient, we are bound to make excellent therapeutic discoveries.   However, the pharma industry needs to change its ways ASAP.

1. In regards to regenerative medicine, it was satisfying to note that Pfizer has recently instituted a global center for regenerative medicine (Nature Biotechnology, 26(5), May 2008).

April 2, 2008
Bruce Tomczuk, PhD

In my entry from March 3rd, I proposed a regenerative strategy for the treatment of a chronic degenerative disease, COPD. There are many diseases in which the end result is a degeneration of a cell type, an organelle, or an organ. A partial list would include Parkinson’s, Stroke, Cognitive impairment, Depression, Spinal cord injury, Type 2 Diabetes, Myocardial infarction, Heart Failure, Wound repair, Macular degeneration, Blindness, Hearing loss, and Alopecia.

Zebra fish
Some species, such as newts and zebra fish, can regenerate injured or severed body parts. Although humans don’t display such dramatic abilities, some tissues are capable of regeneration. Examples include new skin cells to replace damaged cells from sunburn or lacerations. Some nerve and muscle cells can regrow after injury and the hematopoietic system can repopulate following blood donation. These observations give rise to the hypothesis that the genetic basis for regeneration is inherent in humans, but that it can no longer be evoked. Indeed, the genes responsible for zebra fish fin regeneration are also found in humans (Schebesta, M., Lien, C-L., Engel, F. B., Keating, M. T., TSW Development; Embryology 1 (S1), 38-54 (2006)).

Regeneration Basics
Regeneration may depend on a cellular dedifferentiation process (i.e., cells must reprogram from a lineage-specific committed state into a puripotent progenitor state that can subsequently differentiate into functional cells of a different lineage). The dedifferentiation process produces unspecialized progenitor cells which share similar traits to human stem cells. Under appropriate signaling, the dedifferentiated cells can re-differentiate into cardiomyocytes, insulin-secreting cells, neurons, etc. Thus, these cells could provide replacement of damaged tissues or organelles. It has been determined that “terminally” differentiated mammalian cells can, indeed, dedifferentiate when stimulated with appropriate signals. One such signal is msx1 that can trigger the dedifferentiation process. Furthermore, there is evidence that small-molecules can both promote the dedifferentiation process as well as direct the redifferentiation process resulting in cellular regeneration.

SMART
Dedifferentiation
The laboratories of Sheng Ding at Scripps have reported a small molecule, reversine, a 2,6-disubstituted purine can cause dedifferentiation into a stem cell state (Chen, S., Takanashi, S., Zhang, Q., Xiong, W., Peters, E. C., Ding, S., Schultz, P. G., PNAS, 104, 10482-10487, 2007; http://www.scripps.edu/chem/ding). He demonstrated that reservine could reprogram myoblasts into osteoblasts or adipocytes.

Stem Cell Differentiation
Using HCS screening of puripotent embryonic stem cells and chemical proteomics (as discussed in the Phorum on January 16 and March 1), Ding has discovered a 4, 6-disubstituted pyrrolopyrimidine, TWS119, that can induce neuronal differentiation. A chemical proteomic strategy using TWS119-linked agarose affinity chromatography has identified the target as glycogen synthase kinase (GSK-3). However, GSK-3 is involved in a variety of cellular processes and may not be an attractive target for regenerative therapy.

Osteoporesis:
Ding has also used a HCS screen with mesenchymal progenitor cells and a bone biomarker, ALP to discover a 2, 6, 9-trisubstututed purine compound, purmorphamine. The mechanism of action of purmorphamine has been determined to be the activation of the hedgehog signaling pathway.

CNS:
An HCS screen utilizing imaging of neuronal and astroglial markers in primary adult hippocampal progenitor cells resulted in the discovery of a 4-aminothiazole, neuropathiazol, which specifically induced neuronal differentiation (Warashina, M., Min, K. H., Kuwabara, T., Huynh, A., Gage, F. H., Schultz, P. G., Ding, S., Angew. Chem. Int. Ed., 45, 591-593 (2006)). Another small molecule that activates the hedgehog signaling pathway is being pursued for stroke by Curis (http://www.curis.com). As a possible treatment for spinal cord injury, EGFR inhibitors, such as Tarceva, have been shown to promote axonal regeneration (Koprivica, V., Cho, K-S., Park, J. B., Yiu, G., Atwal, J., Gore, B., Kim, J. A., Lin, E., Tessier-Lavigne, M., Chen, D. F., He, Z., Science, 310, 106-110, (2005)). Finally, while not generally considered a degenerative disease, it has been established that all known antidepressants induce neurogenesis in animals. Braincells, Inc. has 2 small molecules, BCI-540 (Phase II) and BCI-632, as antidepressants that act solely by inducing neurogenesis and not by neurotransmitter enhancement (http://www.braincellsinc.com).

Cardiovascular:
A regenerative therapeutic strategy toward myocardial infarction and heart failure has been explored by several groups. Keating has demonstrated that both GSK-3 inhibitors (Tseng, A. S., Engel, F. B., Keating, M. T., Chem. Biol., 13 (9), 957-963, (2006)) or a combination of FGF1 with p38MAP kinase inhibitors provide cardiomyocyte proliferation (Engel, F. B., Hsieh, P. C. H., Lee, R. T., Keating, M. T., PNAS, 103, 15546-15551, (2006)).

Metabolic:
It is well established that there is a degeneration of the islet beta cells in the pancreas of type II diabetics. Thus, a therapeutic approach would be beta cell replication and survival. It has been reported that GSK-3 inhibitors not only promote beta cell survival (protection), but also promote beta cell regeneration (Mussmann, R., Geese, M., Harder, F., Kegel, S., Andag, U., Lomow, A., Burk, U., Onichtchouk, D., Dohrmann, C., Austen, M., J. Biol. Chem., 282 (16), 12030-12037, (2007)).

The Bottom Line:
This is not a comprehensive review of small molecules as regenerative therapies. Thus, if there are other entries, perhaps readers can go to the archive and “click” on “Comments” at the bottom of the blog entry to add to this list.

From the foregoing discussion, not many small molecules as regenerative therapies have progressed far in clinical trials. Perhaps the farthest developed candidate is BrainCells’ antidepressant BCI-540 that is currently in Phase II. In addition, there appear to be some common regenerative targets, GSK-3 and Hedgehog signaling pathway, which may not lend itself to specificity for tissues. It does mean that there is still much to be learned about the control of the regenerative process. The Pharma industry would be wise to focus more on this promising avenue of novel disease-modifying therapeutics.

The author (BT) has no financial ties to any commercial entity mentioned in this article.

March 3, 2008

Bruce Tomczuk, PhD 

In the entry from January 16, I proposed a new drug discovery paradigm that would represent the target disease state by the best mathematical model known in order to identify pressure points which define validated therapeutic systems.  The physiologies used to describe the disease should be deconstructed to include all of the known pathways.  The regulatory control and kinetics of the pathways should be explicitly represented if possible.  Once validated, the pathophysiological subsystems of the model usually provide the human primary cellular or mixed cellular assays one would need for an HTS.  Lead generation and optimization would be performed using these cell or mixed cell assays.  

A Hypothetical Case Study: COPD

Let us examine how one would institute this new paradigm in reality.  Let us assume that you would like to find a therapy for COPD (Chronic Obstructive Pulmonary Disease), a disease with a large unmet medical need.  This is a disease where the end result is the irreversible degeneration of an organelle, the alveoli of the lung.  Unfortunately, there are absolutely no disease-modifying agents to treat COPD.  The best treatment against the progressive nature of this disease is the avoidance of risk factors, such as smoking.  However, the cessation of smoking is a very difficult behavioral modification probably related to the addictive properties of smoking.  In fact, it is not uncommon to witness COPD patients removing their oxygenators in order to smoke.  The standard of care for COPD is just palliative treatment to improve breathing.  Thus, mild-moderate COPDers are generally put on short-acting bronchodilators when needed.  This usually progresses to more chronic treatment with long-acting bronchodilators.  As part of this stage, rehabilitation for breathing and exercise is generally added to the care.  As the disease progresses to the moderate stage, inhaled glucocorticosteroids are added to treat repeated exacerbations.  In the severe stage, patients are placed on long-term oxygen administration from portable and fixed home oxygenators.  Thus, there are suboptimal drugs to treat COPD.  Let us assume that our lofty hypothetical research goal is to discover a disease-modifying therapy that actually reverses the progressive loss of alveolar structure in the lower airways.

Up and running 

In this hypothetical example, the top-down analysis of COPD would describe the loss of alveolar structures in the lower airways as a balance between pneumocyte apoptosis and alveolar regeneration.  A novel approach for disease-modification would be to enhance the native ability for alveolar regeneration.  The alveoli endothelial cells are comprised of type I and type II pneumocytes.  Type I pneumocytes are responsible for the gas exchange that occurs in the alveoli and type II pneumocytes are responsible for the regeneration of type I pneumocytes.    Recent progress has produced a cell culture of type II pneumocytes or a heterocellular culture containing both type I and type II pneumocytes (www. novathera.com; Isakson, B. E., Seedorf, G. J., Lubman, R. L., Boitano, S., In Vitro Cellular & Developmental Biology-Animal, 38 (8), 443 (2002)). Thus, the HCS would be a phenotypic cellular screen assessing the regeneration of type I pneumocytes in elastase-treated (emphysema phenotype) cell cultures.  Compound libraries could be tested probably in 96 well-format.  Positive controls could be agents that are known to induce alveolar regeneration—all-trans-retinoic acid (ATRA), granulocyte colony-stimulating factor (G-CSF), adrenomedullin, and hepatocyte growth factor (HGF) (Massaro, D., Massaro, G. D., Proc. Am. Thorac. Soc., 3, 709 (2006)). 

Target fishing, anyone? 

Hits from phenotypic screens would be reverse analyzed in a chemical proteomic sense in order to determine the molecular mechanism of action (http://www.scripps.edu/chem/ding/index.htm).  In this scenario, we would use the small molecule as bait for target fishing.  Thus the small molecule is incubated with appropriate cellular or sub cellular fractions.  The small molecule would form complexes with molecular targets.  Now, a fishing line is needed is order to pull out these compound-target complexes.  One technology that has been utilized is the well-known biotin molecule which would form a strong complex with avidin on a solid support.  Thus, a biotinylated compound-target complex would be retained on an avidin affinity column.  The retained complex would be eluted for identification (McMillan, M., Kahn, M., DDT, 10 (21), 1467 (2005)).  A more recent technology is the use of magnetic beads attached to the small molecule as the bait.  After incubation with the cellular or subcellular fraction, a magnet is used to fish out the complexes (http://www.invitrogen.com/content.cfm?pageid=11049).  Of course, one needs to be aware that the extra functionalization of the small molecule could interfere with the target binding.  When operating in the dark, several differentiated derivatives may be necessary to corroborate results.   Who are YOU?

The eluted proteins can be identified by mass spectral analysis by a fairly robust process.  The sequence of the targets can be compared to the known proteome for identity.  Let’s assume that the identity of our target is P311, a protein that has been localized in alveolar epithelial cells and has been associated with alveolar development (Zhao, L., Leung, J. K., Yamamoto, H., Goswami, S., Kheradmand, F., Vu, T. H., Am. J. Respir. Cell Mol. Biol., 35, 48 (2006)).  P311 is an 8 kd protein with unknown function.  The protein does not belong to any known family proteins.  Furthermore, the structural motifs of P311 do not provide clues to its function.  Thus, in order to set up a HT screening assay, we would need to set up another HCS, such as the well-known ability of P311 to cause neurite genesis in PC12 cells (Fujitani, M., Yamagishi, S., Che, Y. H., Hata, K., Kubo, T., Ino, H., Tohyama, M., Yamashita, T., J. Neurochem., 10.1111/j.1471-4159.2004.02738.x).  A ready-made kit using an optimized PC-12 subclone, Neuroscreen ™ -1 cells could be utilized as the HCS using an ArrayScan ® Reader (http://www.cellomics.com/component/frontpage/).  One would use the cellular assay for lead generation/lead optimization.  Obviously, if the target were identified as an enzyme or GPRC, then conventional enzymatic or receptor binding assays would be instituted for a molecular HTS in place of the cellular HCS. 

The Bottom Line:  COPD was chosen as a somewhat clean-cut example where there would be little disagreement on the top-down analysis, i.e., an imbalance of alveolar regeneration versus apoptosis.  Other diseases would be more complex and would require causality analysis (e.g., the virtual disease models of Entelos, Inc.).  The outlined paradigm should be capable of producing the most efficacious compound that produces the desired phenotypic change, in this case, neurite outgrowth.  This HCS data should directly translate into an increase in type I pneumocyte regeneration in elastase-treated heterocellular alveolar epithelial cells.  In turn, this should correlate to an observed efficacy of in vivo models of COPD, such as cigarette-smoking or elastase-induced emphysema, assuming target fidelity across species and appropriate pharmacokinetics of the small molecule.  The possibility of a true disease-modifier has to be a better outcome than a me-too bronchodilator. 

The author (BT) has no financial ties to any commercial entity mentioned in this article.

January 16, 2008
by Bruce Tomczuk, PhD 

In the January 7, 2008 edition of Investor’s Business Daily, Peter Benesh’s article “Pharma Forecast Calls For More Of The Same” has focused on the woes of the Pharma Industry. (1)  My favorite line of this article was a quote by Michael Shulman, the editor of ChangeWave Biotech Investor, who stated “These guys don’t have a clue how to develop new drugs.  There is nothing interesting coming out of Big Pharma for as far as the eye can see.”  The gauntlet has been thrown down.   To review my entry from December 19^th entitled “The Grim Prognosis”, I proposed that the lack of new drug productivity was the deliberate reliance of the Pharma Industry on a drug discovery paradigm that focused on one molecular target for one disease strategy.  Indeed, to this date, the vast majority of large pharma still pursue this strategy even in the face of advances in systems biology that suggest that complex diseases may not be effectively treated by interventions at single nodes.(2)   

A New Drug Discovery Paradigm 

Systems Biology
A systems biology approach could be utilized that would implicitly incorporate the complexity found in the disease state.  One of the most interesting approaches to a systems approach of the disease state has been the mathematical modeling approach by Entelos. (3)  In this approach, a disease state is described in virtual patients.  A number of virtual patients could be used in order to encompass the observed clinical subgroups or the number of different patient phenotypes.  As an example, virtual patients in rheumatoid arthritis (RA) might be represented by methotrexate responders, non-responders, and reduced-responders as well as anti-TNF biologic responders and non-responders.   

It’s the complexity, _ _ _ _ _ _
Next, the general sets of pathophysiologies indicative of these virtual patients are defined.  To continue with the RA example, these might include a definition of the common clinical endpoint of ACR20 into changes in synovial cell density, the rate of cartilage degradation, and the levels of synovial IL-6.  Each subsystem is then deconstructed in greater detail from the whole-body dynamics to the molecular level, if possible.  Obviously, at some point, the limits of knowledge are reached.  In this situation, the system description implicitly incorporates the unknown detail.  The subsystems that can be described in molecular detail are explicitly represented by mathematical equations including the regulatory mechanisms, biological redundancies, and the multiple physiological timescales.  The solution of these underlying equations provides a predictive biostimulation of human pathophysiology.  Obviously, the disease representation is validated for the known therapeutic treatments.  This mathematical disease model allows one to interrogate the system to set clinical endpoints based on the mechanism and physiology associated with the proposed treatment.  Entelos has developed detailed representation (termed PhysioLabs®) of several diseases of the metabolic (e.g., diabetes and obesity), cardiovascular, immunology and inflammation, and respiratory systems. 

HTS assay
Sensitivity analysis of the PhysioLab® representation can identify the pressure points in the “system” which might represent new therapeutic avenues.  The subsystems involved in this pressure point can be transformed into HTS assays.  These HTS assays are most likely human primary cell- or mixed cell-culture systems.  One could use the BioMAP® systems that have been implemented at BIOSEEK, Inc (4, 5).  The BioMAP® systems have been chosen to represent the various disease states, much like the subsystems that might be represented in the PhysioLabs®.  As an example, asthma has been represented in the BioMAP® systems that comprise a mixed endothelial and mast cell, bronchial epithelial cell, and endothelial and Th2 blast cell assays.  In an HTS mode, test compounds are screened across the various BioMAP® screens at various concentrations.  The readout from these systems includes multiple proteins, such as cytokines, chemokines, adhesion receptors, and other mediators.  These data are quantifiable and reproducible.  A database of over 1300 unique compounds and over 8000 compound profiles has been generated for comparison.  These reference compounds and profiles represent approved therapeutics and clinical-stage compounds.  Hits from this type of screening could be further characterized in terms of molecular targets, but is not necessary, since it is the integrated signal found in the cellular environment that would be important for disease modification. 

OMICS plus
Omics will provide the parts list of the subsystems that are found in the disease representation, but detailed mechanistic studies will need to be performed in order to map out the regulatory controls, such as cross-talk, negative/positive feedback loops, and redundant or alternative pathways.  Determination of concentrations and localization provide further data for the model.  In addition, studies on the kinetics of each component of the system can provide the necessary physiological timescales of the system that can be plugged into the mathematical model of the disease. 

The Cure
In summary, a proposal for a new drug discovery paradigm would be to represent the target disease state by the best mathematical model known in order to identify pressure points which define validated therapeutic systems.  The physiologies used to describe the disease should be deconstructed to include all of the known pathways down to the molecular target and omics levels, if known.  The regulatory control and kinetics of the pathways should be explicitly represented if possible.  Once validated, the pathophysiological subsystems of the model usually provide the human primary cellular or mixed cellular assays one would need for an HTS.  Lead generation and optimization would be performed using these cell or mixed cell assays.  In this manner, Pharma could escape from its insistence on one target for one disease since the pathophysiological cell systems used would be molecular target agnostic.  

(1)     Benesh, P., Investor’s Business Daily, p. A8, January 7, 2008.
(2)     Hopkins, A.L., Nat. Biotechnol., 25, 1110 (2007).
(3)     www.entelos.com
(4)     Butcher, E.C., Berg, E.L., Kunkel, E.J., Nat. Biotechnol., 22, 1253 (2004).
(5)     www.bioseekinc.com  Disclaimer: The author (BT) has no financial ties to either Entelos or Bioseek, Inc.