The “individualization” of medical treatments, revisited with evolution

The longer I maintain this blog, the more I find unexpected (to me, at least) intersections and relationships between various topics that I write about. Of course, a lot of it simply has to do with the fact that one of the overarching themes of this blog is skepticism and critical thinking, which leads one to seek patterns in various pseudoscience, but sometimes it’s a little more interesting than that. For example, a couple of weeks ago, I wrote a post about the “individualization” of treatments in “alternative” medicine and how it’s largely a sham that alt-med practitioners claim that their therapies are more “individualized” than those of “conventional medicine.” More recently, I’ve been slapping down some truly ignorant statements about evolution by a fellow surgeon, Professor of Neurosurgery Dr. Michael Egnor, who’s clearly being groomed by the Discovery Institute to be one of their more prominent spokesmen because he is a NEUROSURGEON. (Sorry, I’m just using caps to simulate how the DI likes to shout out Dr. Egnor’s profession.)

One of Dr. Egnor’s frequent statements is that understanding “Darwinism” (an epithet he spits out between clenched teeth when he probably really means “evolution”) Is not just irrelevant to medicine but actually harmful to the practice of good medicine. I was actually thinking about that dogmatic rejection of reality at the time I was listening to a talk at the Society of Surgical Oncology Meeting on Friday on systems biology by Dr. Leroy Hood of the Institute for Systems Biology in Seattle. As I sat there listening to the talk, I marveled at the complex biological systems that were being studied and how understanding how these systems evolved was being applied to the molecular diagnosis of disease. Later that night, because it was cold, with falling sleet and snow, I remained ensconced in my hotel room, surfing the Internet, and, almost on cue, I found this little gem from the Access Research Network:

Who can ever cease to be amazed at the complexity of cells and genetic systems? A Perspective essay in today’s Science, concerned with “gene regulatory networks for development”, confirms yet again the reality of complexity. The authors bring out the role of subcircuits in the gene regulatory network. These are “often of elegant and sometimes counterintuitive design, even more so, the ways they are combined in the overall network.” The authors add: “Among the most fascinating aspects of gene regulatory networks are their design principles, for these are often interestingly different from what would seem the “simplest” solution.” The modular construction of these subcircuits is fascinating because it is at a much deeper level than the feedback circuitry we are more familiar with. “Thus, although subcircuits are indeed the modular functional components of developmental gene regulatory networks, they are to be distinguished from simpler “building blocks” or “motifs” that are used for many diverse developmental functions (e.g., feedforward or feedback elements, per se).” The complexity goes even deeper: “As we have come to understand developmental gene regulatory networks, there arises an impression of “overlayered” circuit design–or more precisely, deployment of multiple subcircuits–that in different ways support the same end result.” What are we to make of the inter-dependent modular nature of regulatory networks? Is this the ‘tinkering’ model of Darwinism, or do we have the ‘exquisite design’ model of ID? “We may interpret this as we like–as overengineering; or as design deluxe, replete with bells and whistles; or as the expected result of an evolutionary process in which individual regulatory modules have been added in and overlain at different times.” The authors plump for the ‘tinkering’ model. But even they appear to concede that what we have before us are irreducibly complex systems: “once integrated into the regulatory system, they are there to stay, barring evolutionary redirection”. All I can say is that the hallmarks of ‘tinkering’ are noticeable by their absence! I’ll go for the ‘design deluxe’ option!

Talk about synergy between topics that I had been thinking of writing about! Yep, it’s Michael Behe’s irreducible complexity (IC) argument applied to systems biology of the type pioneered by Leroy Hood. I’m sure no one’s surprised that “intelligent design” creationists would jump all over systems biology as yet more “evidence” of irreducible complexity, when it is no more “slam dunk” evidence against evolutionary theory than Michael Behe’s similar claim for the alleged IC of the bacterial flagellum, for example, or the oldest argument from incredulity of all, the one that Charles Darwin himself answered, the claim that the eye could not have evolved. In fact, given how functionally redundant many of these networks are (with overlayered “circuit” design and interlocking redundant networks of regulatory genes), they would seem to be against the very definition of IC, given that components can be removed from many of these systems and, thanks to redundant pathways that take up the slack, the overall network will still function. Now we see creationists moving towards arguing that these interlocking networks of regulated genes couldn’t possibly have evolved, even though, as complex as they are, it is not difficult to imagine selection pressure acting both at the level of the genes and the modules of gene products that form the “subcircuits” and how small accumulated changes over many generations could lead to complex networks. After all, if gene duplication can lead to new biological information, as the duplicated gene, under selection pressure, gradually diverges from its original structure and acquires new functions, it’s not that hard to imagine a similar process happening to the “modules” of these biological networks. Indeed, the authors of the Science article themselves basically say this:

Gene regulatory networks for development are the direct product of evolution, and the character of their design both illuminates evolution and is illuminated by it.

[…]

There is a plethora of regulatory jobs different from one another–such as the development of embryos, or of stem cells, or of adult body parts–that all require different kinds of subcircuits. The subcircuit components of gene regulatory networks have evolved independently of one another, and at different rates (2), and are assembled in different contexts in related organisms. Both in their functional organization and in the separate evolutionary histories of their subcircuits, gene regulatory networks are modular in construction.

And the conclusion:

We may interpret this as we like–as overengineering; or as design deluxe, replete with bells and whistles; or as the expected result of an evolutionary process in which individual regulatory modules have been added in and overlain at different times, so that some are more ancient and others more new (1). However, once integrated into the regulatory system, they are there to stay, barring evolutionary redirection. But the generality of this quality of developmental gene regulatory networks is emerging as a fact of life–it is what we see in modern animals. The consequences of evolutionary history determine the shape of the control apparatus that determines life processes. Perhaps in current system design we are seeing something of the grim pressures that modern lineages survived in past evolutionary bottlenecks–of the absolute necessity for lineage survival of genomic regulatory systems built to run and not to fail.

It’s a bit unfortunate that the authors used the word “design” so much; it got the ID activists all hot and bothered even though there was nothing in this commentary to support that these networks were consistent with ID. It’s the same lazy figure of speech that seems to have led Dr. Egnor to think that, when doctors say an organ or structure is “designed” to do some function or other, they really mean it was designed by some being.

All this was going through my mind, when it occurred to me that systems biology could be viewed as the latest contribution of evolutionary theory to medicine. I was actually rather that I had taken some notes on Dr. Hood’s talk, which was quite fascinating, if a bit more optimistic than I consider warranted. It started out with an explanation of biology as an informational science, with the key discoveries being:

  1. The theory of evolution, which told us how biological information evolves.
  2. Gregor Mendel’s discoveries, which told us how biological information is passed on to an organism’s progeny.
  3. Watson and Crick’s discovery of the DNA double helix, which told us how biological information is stored and passed on.
  4. The Human Genome Project, which gave us the basic information stored in human DNA upon which the organism is based.

Dr. Hood’s contention is that biological information is mostly digital (the genetic code) and thus ultimately knowable but that environmental factors represent analog information that impinges and modifies the inherent biological information. Those environmental factors represent factors that act both on the individual organism, leading certain networks to be activated or deactivated, and on the population as selection pressure for evolution. Dr. Hood spoke of a hierarchy of biological information that ranged from DNA to RNA to protein to regulatory networks to cells to multicellular organisms to populations to ecologies, information from all of which are used to try to understand gene regulatory networks. (I may have missed a link or two; he was really talking fast at this point.) Systems biology consists of a “parts assessment” (understanding the components of the biological networks) and the assembly of individual circuits and understanding how they are put together to form more complex collective circuits. These circuits consist of elements (called “nodes”) and interactions (called “edges”); these interactions can give rise to emergent properties. And Dr. Hood is by no means alone in studying these questions. For example, there’s the Evolutionary Systems Biology Group at SRI and the work going on at the Institute for Genomics and Systems Biology at the University of Chicago, just to pick a couple out of a search.

To study these systems, Dr. Hood is developing quantitative measurements, global measurements, dynamic measurements, and computational methods of mathematical integration of many different data types. The researchers working at the institute span many disciplines, from engineering, to mathematics, to genomics, to cell biologists, to immunologists, to bioinformatics, to evolution. The goal of studying systems biology is stated as producing medicine that is 4P: “Predictive, preventive, personalized, and participatory.” Yes, the last two P’s are what caught my attention. Alternative medicine practitioners claim that they are providing “personalized” treatments. Indeed, homeopaths even go so far at times as to say that each person requires a different treatment. My retort to such nonsense is that the sort of work that was described last week is what real personalized medicine will be about, as I’ll try to explain. At the same time, it demonstrates yet another example of the power of evolutionary biology, when coupled with multiple other disciplines, to produce medical advances through the understanding of complex biology.

Using these techniques, Dr. Hood described research showing how diseases can be viewed as perturbations in the signaling by gene networks and subnetworks, and how these perturbations can be measured by looking for changes in organ-specific proteins in the blood. Most individual biomarkers are rather useless as single markers, with a few exceptions, and even these exceptions are fraught with difficulties and problems with specificity. These problems could perhaps be obviated by looking at large numbers of organ-specific secreted proteins whose perturbation is associated with perturbations in key points in various biological circuits. As an example, for the case of prion diseases like Creutzfeldt-Jakob disease and mad cow disease, it was shown that a systems biology approach, in which 900 key signaling components whose levels were perturbed through the course of the disease were winnowed down to identify a protein signature that can identify the perturbed networks before neurologic symptoms arise, leading to a potential blood test for cattle for mad cow disease and for humans.

The ultimate goals of this research are two-fold. The first goal is early diagnostics. As recently as 5-10 years ago, the technology was not up to this, but now it is becoming possible to use microfluidics and protein arrays to measure 2,500 proteins simultaneously from a single drop of blood. The sensitivity is reaching the 40 attamole (40 x 10-18 mole) range. The result, it is hoped, will be the ability to identify the perturbations in various molecular circuits that are associated with disease, so that intervention can occur at an earlier point in the process. Indeed, Dr. Hood presented his vision of such a blood test that one would do every six months with a drop of blood no larger than what diabetics use to check their blood sugar. If the results are abnormal, then a trip to the doctor would be in order. The second goal is therapy, with drugs designed to target key nodes in these circuits and push their activity towards normal, thus normalizing the disrupted system. This second goal is a bit further from realization compared to the early diagnostics. The ultimate result would be a true personalization of medical treatments based on the detailed molecular “fingerprint” of each patient. This is far more substantive “individualization” than any homeopath could ever achieve, and that’s what medicine is working towards now. It also depends on systems biology and genomics, the understanding of which is very difficult, if not impossible, if not studied in the context of evolution:

…most of the known mutations causing Mendelian disease in humans occur at evolutionarily conserved positions in proteins: in other words, patterns of negative selection (i.e. selection against changes at these positions) apparent from comparing related genes (in the same or in other species) reveal essential sites in proteins that, if mutated, cause human disease. We have recently shown that detailed evolutionary modeling, including signals of positive as well as negative selection, can significantly improve our predictions for both Mendelian and even complex disease.

Of course, I would be remiss if I didn’t report that Dr. Hood’s talk, as several talks that I’ve heard from visionaries, was more than a bit pie-in-the-sky. While his use of systems biology to study prion diseases and develop tests that detect mad cow disease and human prion diseases before symptoms appear is very impressive, Dr. Hood seems not to realize that developing these early diagnostic tests may not be a quick as he thinks for common human diseases like cancer, diabetes, or heart disease. The main reason is the lifespan of humans and the long “lead time” before subclinical disease becomes detectable. Let’s take the example of an early detection system for breast cancer. Let’s say that Dr. Hood’s blood test can identify the tumor at a very early stage or even identify precancerous conditions that turn into breast cancer over time (or, to confuse things even further, consider the more likely result: a percentage probability that the patient has breast cancer). Most breast cancers, by the time they are clinically detectable, have been around for several years, as long as a decade or more in some cases. Validating any blood test like Dr. Hood’s would require many years to verify that its results actually do predict the ultimate development of the disease and to learn how to recognize false positives, particularly given the number of proteins being measured. Similarly, consider one issue brought up by such a test. Let’s say that it detects “breast” cancer at such an early stage that no imaging modalities (mammogram, ultrasound, MRI) can detect it. But we know it’s there. What’s a woman to do? Well, assuming that the development of Dr. Hood’s molecularly targeted drugs that will reverse the “systems perturbation” responsible lags behind the development of the early diagnostics test (a reasonable prediction), it is quite likely that the woman will have to decide whether to wait until the tumor becomes detectable by conventional imaging or have both of her breasts removed to treat the tumor. Similarly, at least in the realm of diagnostics, it’s hard to imagine that such tests won’t start producing a whole lot of false positives, particularly if administered every six months, as Dr. Hood envisions, at least in the early stages of their use, until enough data arises to allow us to filter out the noise.

Issues like these, however, can usually be resolved with time and research. Despite the “gee-whiz” factor and a bit of old-fashioned excessive hype, systems biology is likely the wave of the future that will indeed likely result in something similar to Dr. Hood’s vision of “4P” medicine. Indeed, it’s already in its embryonic form and may become fairly widespread within my lifetime. It may even put surgeons like me out of work (other than trauma surgeons, who would be dealing with the more macroscopic “systems perturbations” that will always be with us), if the vision of being able to design drugs to eliminate systems perturbations associated with disease is ever realized. Be that as it may, the bottom line remains that, no matter how much ID advocates might want to label the circuits and systems in systems biology as “irreducibly complex,” they almost certainly aren’t, any more than the eye or the bacterial flagellum is. They can say that, because of their intricate complexity, they must have been “designed” all they want. In actuality, these circuits can only be come to be understood well enough to manipulate their activities for therapeutic effect through understanding how they evolved and using lower organisms in which various modules of these systems are conserved as model systems to manipulate the circuits before trying to manipulate them in humans.

Creationists, for example, can deny that understanding evolution is a great help in medical research and developing the next generation of cutting edge diagnostics and treatments all they want, but that won’t make it so. I’m also going to go out on a limb a bit and predict that the very alternative medicine aficionados who are so enamored of the “individualization” that their practitioners supposedly offer them won’t recognize the real individualization of treatment that will be provided by systems-based approaches.