Based on many years of research, the Serbian and Croatian startup Kidneya Therapeutics is developing a platform for the early diagnosis of kidney diseases, taking specific environmental factors into account. Prof. Nikola M. Pavlović, a medical doctor specialized in nephrology, and the software engineers Dragan Milovanović and Deni Ćosić are collaborating to combine medicine and Artificial Intelligence. In this interview with 5-HT, the team of Kidneya Therapeutics explains how AI methods can facilitate the early diagnosis of kidney diseases and accelerate the development of new drugs.
Why is it important to discover kidney diseases at a very early stage?
Nikola M. Pavlović: Chronic nephropathies are an enormous problem for healthcare systems worldwide. Once a patient has reached an advanced stage, kidney diseases require costly treatment like dialysis or organ transplantation.
Dragan Milovanović: The problem is that some kidney diseases progress very slowly. You need 20 years to reach a point at which they can be recognized, but at that point, the patient is already in an advanced stage requiring transplantation or the like. By using digital biomarkers, we enable medical doctors to discover renal diseases at an early stage or to predict if a patient has a high risk of developing these diseases.
What kind of biomarkers can predict or indicate kidney diseases?
Nikola M. Pavlović: In 30 years of research, I have investigated the role of environmental factors in nephrology. Aristolochic Acids, toxic chemicals produced by Aristolochia plants like Aristolochia clematitis, have proven to contribute to a great extent of chronic renal diseases. They can cause Aristolochic acid nephropathy (AAN) and Balkan endemic nephropathy (BEN), as well as upper urothelial cancer (UUC). Aristolochia plants grow in abundance all over the world, they can contaminate soil and groundwater, and they can also be absorbed by crops, thus entering the human food chain. These findings, which I published in the Journal of Agricultural and Food Chemistry (JAFC) together with my colleague Wan Chan, were honored as JAFC Research Articles of the Years 2016 and 2017. Aristolochia plants are widely present around the world, especially in European countries like Germany or France, but they are not yet recognized as potential hazard.
How do you make practical use of this knowledge?
Dragan Milovanović: By applying Artificial Intelligence and Machine Learning, our system combines clinical data, biochemical findings, and data about geolocations in order to identify patients with a high risk of developing chronic kidney disease.
Nikola M. Pavlović: Through the analysis of blood and urine samples we determine the level of Aristolochic acid in the human body and the amount of Aristolochic acid which has been transformed and bound to the DNA. Furthermore, we add data about the patient’s geolocation. For South Serbia, our first test area, we have already analyzed the concentration of Aristolochic acid in groundwater, soil, and vegetables from the fields.
Deni Ćosić: Our system will be usable as a medical application: Doctors can input the patient’s medical records, and the program will output the patient’s individual risk of developing chronic kidney disease.
How do you plan to make your solution accessible to medical professionals outside of South Serbia?
Dragan Milovanović: In general, our system can be applied anywhere in the world. As soon as the development is completed, medical doctors worldwide will be able to use our technology, probably based on a subscription model. For the use in new regions, however, it will be necessary to take samples from soil, water, and crops first.
Kidneya Therapeutics focuses not only on the early diagnosis but also on the development of new drugs against renal diseases. How does your research on Aristolochic acids help you in this field?
Nikola M. Pavlović: It has been shown that Aristolochic acids can be very dangerous, but Aristolochia plants have also been used in old Chinese medicine for millennials. Alongside the very adverse effects, they also have some beneficial effects. It is a very tricky job, but with AI methods we found a way to test the potential beneficial effects of some of the compounds of Aristolochic acid.
How can AI improve the development of a new drug?
Dragan Milovanović: Our basic idea is to find derivatives of Aristolochic acid which have only positive but no negative effects. AI models can be used to generate molecules with different functionalities which are similar to the starting molecule. This way, we have already identified 100 new molecules which are slightly different from the original Aristolochic acid.
Deni Ćosić: In the next step, we apply a machine learning model to test binding probability and affinity between these molecules and the receptors of different proteins and enzymes in the human body. Finally, we will test the metabolic effects of these derivatives – what happens when they are inside the system of the human body?
Dragan Milovanović: The use of modern technologies enables us to considerably speed up the process of drug development.
What is the story behind the foundation of Kidneya Therapeutics?
Dragan Milovanović: Professor Pavlović and I met more than a year ago. When one of my family members was suffering under a kidney disease, I asked for his advice. We started talking, and what amazed me was his interest and his willingness to know more about AI and machine learning. This is how the idea for Kidneya Therapeutics was born. At the moment, the professor is learning a lot about AI and machine learning, and Deni and I are learning a lot about medicine.
What do you hope to achieve by being part of our Digital Hub?
Nikola M. Pavlović: We are looking for collaborations with various kinds of institutions in different countries, especially in Germany. By establishing connections to renal doctors, toxicologists, geneticists and many more we aim to learn more about the epidemiology of chronic kidney diseases and to compare our data with findings from other European countries.
Dragan Milovanović: We are also looking for cooperation with pharmaceutical companies to initiate clinical trials. Furthermore, as a startup, we are also searching for investors to help us develop our team and our system. Being part of your network gives us a great opportunity to establish new connections like these.