Virtual Pharmacology
Molecular modeling for medicines
Peter Fedichev, Post-ScienceDrug development is an expensive and risky process.
Consulting companies estimate that it takes up to a billion dollars to develop a successful drug, and it usually takes more than 10 years. The most terrible thing is that due to increased safety requirements, more and more medicines, for which hundreds of millions of dollars and years of work have already been spent, cannot pass clinical trials. The number of refusals to register medicines is increasing. This means that if this goes on, then very soon there may not be people who will agree to finance this process because of the high risks.
In addition, the diseases that medicine faces are becoming more complicated. Simple diseases are cured, it is increasingly difficult to find any one cause of the disease (such as a bacterium or virus). This way it is impossible, for example, it is still impossible to cure Alzheimer's disease or diabetes mellitus. Often extremely expensive therapies are not able to defeat fatal diseases and only slightly prolong the patient's life (more than half of the treatment costs may fall on the last year of the patient's life). Worse, all people are genetically different, and each person has their own cancer or diabetes. For one patient, inside one cancerous liver tumor, a puncture can give four or more types of cancer cells. This means that while the number of diseases that we must learn to treat is constantly growing, the price tag of hundreds of millions of dollars and huge risks limit people's desire to do this.
Unless some breakthrough technologies appear that will significantly speed up and reduce the cost of developing more and more specific and safe medicines, perhaps in 100 years we will live as long as we live now. In principle, this is not such a big problem, because over the past hundreds of years, the maximum life expectancy has practically not changed. We are lucky that we live at a time when life expectancy has increased rapidly in a short time, mainly due to improved sanitation standards and the advent of effective antibiotics. But breakthrough technologies are emerging, and the key here is the use of modern computing tools, artificial intelligence.
Biology is a very capacious science of facts, which, unfortunately, are mostly random. To the question "Why is this so?" The answer is one: "It just happened. Evolutionarily. Because there was such a very complex trajectory of the development of genes in human ancestors." To successfully treat diseases, you need to link a lot of facts, but there are a lot of random connections in these facts that are extremely difficult to analyze. There is no logic, a single creator or a single scheme by which all this would be organized. Therefore, to analyze genetic networks and signaling pathways that lead to the development of certain diseases, it is simply necessary to use powerful computing tools.
Computers allow you to analyze signaling pathways, to study how some genes affect others. Since nature is built on the principle of maximum survival, i.e. maximum stability, many genes delegate connections between themselves to so-called regulatory genes, which are simultaneously associated with a large number of genes. Identifying those master regulators that are activated in diseases and are not active in a normal state is the task of identifying targets for medical interventions. A variety of methods are used for this. This is mainly genomics; now transcriptomics is emerging, i.e. genome sequencing followed by the study of gene expression levels and metabolites in cells.
Thus, there is a way to study tens, and possibly in the near future, hundreds of thousands of parameters of each cell. Subtle differences in gene expression are revealed, and most importantly, it turns out which of the genes, being expressed at the wrong time, cause a change in a healthy cell to a sick one. They have a chance to become the targets for which the pharmaceutical industry will develop drugs that will really help from a particular disease.
So far, most of the drugs have been found, oddly enough, by trial and error, even if it was about the industry. By the method of so-called high-performance screening, dozens, hundreds, sometimes under a million substances were tested on certain cellular models of certain diseases, of which dozens and hundreds were promoted further as potential drugs. Unfortunately, in the past years, large companies have spent millions of dollars screening hundreds of thousands of substances against certain pathologies, where none of these substances gave a worthy candidate.
The computer method of molecular modeling allows you to simultaneously study tens of millions of molecules for binding them to a specific target. Since these molecules are created against a specific target, which should not be in a healthy person, there is less chance that such a molecule will become toxic and will not pass further stages of preclinical and clinical studies.
Thus, the creation of such technologies not only speeds up or reduces the cost of research, but often allows, in principle, to find at least something from the most important diseases. In this way, for example, molecules from viral hepatitis have recently been obtained. This summer they were approved in the United States, and therefore in the world. For the first time in the history of mankind, viral hepatitis C has become, if not curable, then a controlled disease. For a certain type of breast cancer, very effective, so-called targeted drugs have appeared, which significantly increased the prognosis of patients' survival.
We are at the beginning of this process. Computers for drug development began to be used about 15 years ago, but about the last 5 years we finally see examples when it really began to work.
Is it possible to apply molecular modeling in any other field? Of course you can. In addition to the search for new drugs, people are faced with the task of creating new materials: polymers, materials for energy or photoelectronics, etc. Wherever it is necessary to come up with a molecule that has not yet been in nature, good molecular modeling methods are needed. Biology is probably one of the polygons where these methods are being worked out to the stage where we can first find an artificial molecule for any biological molecule that regulates it, and in the future we will be able to create new crystals, semiconductors, materials with unique properties that have never been in nature before. For example, for quantum computers, which, as many believe, will require the creation of fundamentally new materials, the properties of which can only now be imagined.
Pharmaceuticals is one of the most high-tech areas of business. This business, obviously, will use any opportunities that appear, including the possibilities of molecular modeling. International pharmaceutical and biotech companies are very receptive to any new technologies, they are constantly trying them. I am sure that only thanks to the advent of such technologies will we see new medicines that will probably cost less money than they do now, which will really help patients, even if it is necessary to make a medicine that only a hundred patients on the planet will need. Now such a question is not even worth it: because of the costs and risks, no company will make a medicine that ten thousand people will need. But in the future, personalized medicine will be able to diagnose, develop and apply therapies that will make sense for small groups of everything, maybe even just one patient!
The author is PhD, Scientific Director of Quantum Pharmaceuticals, Head. MIPT Laboratories of Systems Biology
Portal "Eternal youth" http://vechnayamolodost.ru14.05.2013