For many people around the world who suffer from several chronic diseases, swallowing a large number of medications daily is literally a bitter pill to swallow. It actually deters some of them from taking their drugs regularly. 

Now, researchers at the Hebrew University of Jerusalem (HUJI) have developed a unique algorithm that predicts already in its design stages – based on its chemical structure – if the chemicals in the pill have a bitter and repulsive taste. The algorithm, created using Machine Learning, will be able to speed up the process of drug development and make unnecessary the performance of animal experiments to assess the bitterness of the materials under development. 

Drug-development research effort focuses on the development of effective and safe medications to treat disorders with a minimum of side effects. The use of sophisticated and automated processes has led to the approval of thousands of drugs by the US Food and Drug Administration (FDA). In 2020, only 53 new drugs were actually approved. But there is a problem that does not get an optimal solution – many drugs have a very bitter taste that makes it difficult to take them by mouth. 

Although introducing the drug into capsules usually solves the taste problem, this is a solution that is not always possible as this makes the medication larger and may make it more difficult for the patient to swallow it. While many of us manage to swallow pills despite therir terrible taste or problematic size, there are populations, including children and the elderly, who have great difficulty doing ao, 

Studies have shown that when drugs had a very bitter taste, children refuse to take them, posing a health hazard from incomplete drug treatment. For example, a study published recently found that over 90% of pediatricians report many children who are unwilling to take medication because of their taste, which exposes them to health damage and lack of proper medical care. Moreover, because of the potential risks posed by the deterrent taste of many drugs, the FDA recently sought to add a warning on flavor to the list of side effects in prescriptions for children’s drugs, to warn parents of those problematic tastes.

Pharmaceutical companies recognize that there is a problem of very bitter taste only in the advanced stages of development or in the clinical trials themselves when the drug is given to thousands of people who serve as experimenters to test its effectiveness. In case a particularly unusual taste problem is identified, companies will have to go back and change the formulation of the drugs in order to try and mask the repulsive taste (which is not always possible). This leads to further delay in releasing the drug to market, millions of dollars in losses, and additional animal experiments. Most drug companies prefer to launch a drug even if they have already realized that it is very bitter, in the hope that the public will manage to swallow bitter pills. 

The research group of Prof. Masha Niv, a taste researcher at the Institute of Biochemistry, Food and Nutrition at HUJI’s Faculty of Agriculture, Food and Environment understood the severity of the problem and were able to develop an algorithm that predicts — based on its theoretical chemical structure – if the molecule has a bitter taste. The study was recently published in the journal Computational and Structural Biotechnology.

Eitan Margolis, a doctoral student in Niv’s laboratory, explained: “We collected molecules from the BitterDB database established in our lab that contains information about more than 1,000 bitter compounds. We integrated it with information on the taste of molecules that we received from our colleagues at the Glaxo Smithkline pharmaceutical company and the natural-materials company AnalytiCon Discovery. We taught the computer to understand which combinations of chemical properties are most important for a substance to be perceived as having a very bitter taste and thus ‘trained’ it to quickly detect whether a particular molecule may be very bitter or not. So, in fact, we created the BitterIntense algorithm. After testing the model’s performance, we saw that it was correct in more than 80% of cases; this strengthens its potential for developing taste-friendly drugs without the need for early synthesis and testing in humans or animals. Since this is a computational model, we will continue to improve its accuracy as more information about bitter molecules accumulates.” 

Contrary to previous scientific beliefs, it has been found from the results of predictions that very bitter drugs do not tend to be more toxic to the liver than less-bitter drugs. However, the fact very-bitter substances seemed to be more toxic to the heart due to their ability to block potassium channels is particularly interesting because receptors for bitter taste are also expressed in the heart. Current studies focus on the physiological functions of taste receptors that are expressed outside the mouth, added Niv.

And what about medicines for COVID-19? Margolis added that “we have seen that the prevalence of highly converted drugs among those currently in development or approval for Coronavirus treatment is higher than medications in general. This result is interesting because loss of taste is a very significant and common symptom among those infected with the virus. However, the fact that very-bitter drugs can have a beneficial physiological effect emphasizes the need not to disqualify outright those very-bitter drugs based solely on their taste, but only to prepare accordingly.” 

In conclusion, the researchers believe that BitterIntense is a method that can reduce costs, minimize animal experimentation and shorten the time the drug arrives at the clinic. The ability to detect high bitterness in an integrative manner in the process of discovery will help in the development of drugs appropriate for children and geriatric patients. In addition, BitterIntense is also relevant to biotechnology companies and that are working on the development of new sweeteners (sometimes also bitter) or natural ingredients that aim to be integrated into food products. 

Margolis shared a personal experience related to the research: “After the publication of the scientific article, many young people and adults approached me through social media and told me that they very much identify with the problem of drug bitterness. Some said that until now, they had to chew pills or take them with foods with a very strong taste – and even then, it doesn’t always work. Others shared the frustration of giving bitter pills to children who in many cases spit them out, and then they didn’t know whether the child had to get another pill or not. We hope that our research will provide a better solution to the problem of the bitter taste of drugs and to develop innovative food ingredients.”