Customized medicines with pharmacogenetics

More than 95 percent of the population carries at least one genetic variant that is responsible for abnormal drug metabolism. Maaike van der Lee conducts research into these variations and explains how we can apply the knowledge about them in clinical practice.

Maaike van der Lee works at the Clinical Pharmacy and Toxicology department at the LUMC. There she works as a pharmacist and conducts research in the field of pharmacogenetics: a science that investigates the influence of genetic variations on the individual response to medication. “Everyone responds differently to medicines,” says Maaike. “This is partly caused by variations in our genes. If you know how someone reacts to a medicine, you can adjust the dose accordingly or prescribe a different medicine.”

Liver enzymes

There are several factors that influence drug metabolism. Variations in the piece of DNA that codes for liver enzymes are a common cause. “These enzymes,” Maaike explains, “are responsible for breaking down medicines. If you have a genetic variation there, you break down a medicine more slowly or faster. If the breakdown is slower, you get more of the drug in your blood and you have a greater chance of side effects. On the other hand, if you break down a medicine faster, it is removed from your blood too quickly and therefore less effective.”

A drug where pharmacogenetics is well applied is Clopidogrel. This blood thinner is often prescribed after a heart attack. “Liver enzymes convert Clopidogrel in the body into an active substance,” says Maaike. “Due to a variation in the DNA, the drug is not converted in some patients. There is an alternative medicine, but it is basically a second choice. By conducting a genetic test, the doctor or pharmacist can determine very precisely who responds to clopidogrel and who should be prescribed the alternative medicine.”

At the LUMC, Clopidogrel determines in advance whether the drug is effective, but generally the genetics are looked at afterwards to explain side effects. For example, in a patient who has tried several antidepressants and suffers from many side effects. The genetic information then helps to determine an alternative treatment. “Ultimately, we want to first do a test on everyone who starts taking a medicine for which we know that genetics plays a role. This way you prevent someone from trying a drug and suffering from side effects,” Maaike adds.

AI model

When determining drug metabolism, enzyme activity is currently classified into four categories: none, reduced, normal and increased activity. However, the genetic variants are much broader than these categories, Maaike explains. “We have done research into the CYP2D6 enzyme. This liver enzyme is responsible for processing a quarter of commonly used medications. We saw more than 200 different genetic variants in a group of more than 650 patients. With the current method you group more than 200 variants into four categories.”

Together with her colleagues, Maaike has investigated an alternative approach. “We have mapped all genetic variations in the CYP2D6 enzyme in our patient group using long-read sequencing. In addition, we have information about the activity of the CYP2D6 enzyme, from the blood levels of the patients.” The researchers used this information to train an AI model and predict the functioning of the enzyme. “With the current grouping into four categories we could explain 54 percent of the difference in enzyme activity, but with a continuous scale it was 79 percent. The difference in percentage is so high because the continuous scale includes all variants and does not group the patients.”

The prediction of an individual's drug metabolism is much more accurate with this new approach. But Maaike explains that a broad application of this method is not yet possible. “We have now demonstrated this for one medicine. And there are also other factors you want to investigate, such as dosage and patient side effects. Our biggest limitation right now is the data. We need good data and large cohorts. After all, an AI model is only as good as the data you put into it.”

DNA medicine card

The Netherlands is a leader in the field of pharmacogenetics. The pharmacogenetics working group of the Royal Dutch Society for the Promotion of Pharmacy (KNMP) has developed pharmacogenetics guidelines, which are integrated into prescribing systems. In practice, this means that the pharmacist or doctor receives a notification when an adjusted dose or an alternative medicine is recommended, based on the genetic information. “This way of working is already highly integrated and we see a shift, with more and more general practitioners working with this in addition to hospitals.” Patients are increasingly requesting a test themselves, and with the arrival of DNA medication cards it is becoming more common. “I think pharmacogenetics will become a standard part of healthcare. You only have to do the test once and it provides demonstrable health benefits.”

Do you want to know more? During the one-day conference “Innovations in DNA & RNA Technologies” on April 16, Maaike van der Lee will give a lecture about pharmacogenetics and how we can apply this knowledge in clinical practice. Visit the website for more information and Sign In for a free visit.