Prediction of side effects


The phenomenon of side effects of treatments is neither specific for personalized medicine nor for medication. Whatever we do in order to achieve anything usually has some unintended side effects. I deliberately distinguish here between intended and unintended effects, as both can be adverse or beneficial.


Aspirin - bright and dark sides 


Let me illustrate the concepts on a very well-known example: Aspirin, one of the oldest drugs that started as a natural compound from the bark of trees and quickly got synthesized chemically as a drug. The original observed (and thus intended) effect was anti-inflammatory and analgesic (reducing pain). It has a myriad of side effects both beneficial as well as adverse. To start with the adverse side effects, Aspirin is known to cause in quite a number of cases stomach bleeding, which can be rather severe. A reason why the drug would never make it to the market today as it would fail in clinical studies due to this adverse effect alone. However, the beneficial side effects would have never been seen as they were detected by serendipity over the decades of using the drug. 


Aspirin is now regularly applied in low dosage (up to about 90 mg or 100 mg/day) to prevent strokes and heart attacks. This is due to its ability to prevent (respectively reduce in low dosage) blood clotting responsible for closing small blood vessels in brain and coronary heart vessels already affected by stenosis (reduced diameter due to plaque formation). Of course as another side effect this may complicate a bit the blood clotting responsible for stopping the bleeding of wounds but the looming danger of a stroke or heart attack is far more dangerous that that. 


More recently, Aspirin was also found to interfere with metastasis of cancer cells, basically due to the same reason it prevents blood clotting: Platelet cells are used a living shields by traveling metastatic cells in the blood escaping the immune surveillance, which would destroy cancer cells. Under the influence of Aspirin platelets no longer shield the metastatic cells, which then may fall prey to the immune system (Scientific American, May 2017). 



Side effects are an inevitable consequence of the interconnections of biological functions


This nice example not only illustrates the ambivalent nature of side effects (life-threatening vs life-saving) but also highlights why side effects are basically inevitable. Platelets are really rewired on a molecular basis by Aspirin and this involves about 60 genes that are changed in their expression by Aspirin. Basically, all of these genes are involved in multiple biological circuits and functions. Therefore, it is impossible to change one circuit (in this case the undesired shielding of metastatic cells) without affecting a variety of other functions as well. 



Figure 11: Origins of side effects 


As illustrated in Fig 11 there are different ways side effects become inevitable. First the drug or biological designed to have the intended effect and target usually affects more targets that intended causing side effects. The other way is that most drug targets are involved in more than one function. As a consequence necessarily more is changed than intended - causing side effects. 


Personalized medicine and side effects


How can Personalized Medicine help with unintended side-effects despite their inevitability? Side effects cannot be prevented by personalized medicine but it can help in two respects: Moderating the severity by personalized adaption of the dosage and changing to less adverse side effects by changing the drug and consequently its target(s) based on genetic or epigenetic information. Different people have quite different kinetics of drug metabolization and duration within the body. Knowledge of genetic or epigenetic parameters influencing metabolization or secretion is a basis to adjust dosage and frequency of drug intake in order to reach the optimal ratio between intended and unintended effects. As already outlined in the Aspirin story, biological systems are networks often with multiple ways to achieve the same thing (functional redundancy). This is necessary to ensure that a single event destroying one way does not lead to death but can be compensated much the same way as airplanes have all critical steering transmission lines in triplicate. Knowledge about the personal peculiarities of such (partial) networks ca be used to select another drug affecting the same disregulated process but at another point of the network. 


The bottom line to all that is that we need to know how side effects arise before we can either harness them into new treatments (see  Aspirin and metastasis, not yet a treatment) or avoid adverse effects (such as the stomach bleeding, which is now countered by other drugs given in parallel). Personalized medicine is seeking the molecular basis of diseases and drug actions and thus is best suited to answer many of the crucial questions. 


What’s coming up next?


Next week I will expand into a related field, the selection of the optional therapy which also requires more knowledge about what is going on in the organism.