Understanding the Liver

The role of liver cells in drug research 

For five years, more than 200 scientists worked to gain a better understanding of the physiologicalprocesses that take place in the human liver. Their aim: to develop new technologies andprocedures for pharmaceutical research. The team included systems biology experts from BayerTechnology Services

Researchers are attempting to model processes in the liver (red)

Doctor Lars Küpfer has a goal. “Through my work, I’d like to improve the development of therapeutic drugs and their application for patients,” says the process engineer, a Senior Scientist at Bayer Technology Services. The computer is his most important tool. He uses it to work on models and computer simulations primarily aimed at making one thing possible: to better understand – and perhaps even predict – the behavior of certain molecules in the human body with the aid of virtual patients. These simulations are a valuable supplement to experimental results and clinical studies, without which regulating authorities will not approve any new medication. In these studies, a particular preparation has to demonstrate, among other things, that it really works, and also has to show what constitutes a sensible dosage to achieve its effect. On top of that, a new drug has to prove that it represents a real advantage for patients over existing medications.

Computer models are already capable of reproducing human physiology. And when they are input with the right data, then they allow predictions to be made not only for known cases but also for situations that have not been investigated previously.

“Being able to predict the behavior of substances better helps us to optimize our clinical development program.”

Dr. Jörg Lippert

Global Head of Clinical Pharmacometrics, Pharmaceuticals

At Bayer Technology Services, Lars Küpfer works as part of a systems pharmacology group, a special offshoot of systems biology that focuses primarily on the interplay between processes in biological systems – and tries to understand these in as quantifiable a way as possible. Systems pharmacologists investigate the ways in which pharmacological substances are distributed through the human body, and how, as a result of this distribution, an effect occurs that benefits the patient.

This raises the question: why bring a process engineer’s expertise into something like this? Küpfer laughs and explains that it ultimately comes down to using extremely complex mathematical formulae that are interdependent in multiple ways to describe what happens. And this, in principle, is not much different from the classical work of a process engineer, such as modeling the chemical processes in a reactor, for example.

Having said that, Küpfer is quick to stress that systems pharmacology is a young discipline that brings together many different skills: “Our group includes biologists, chemists and physicists working side by side with mathematicians, doctors and, as it happens, process engineers.” Only by working together can the complex processes at work in the human body be modeled and simulated. For example, what happens to a substance in the body, and how it is metabolized.

In the past five years, the experts have made great strides. A major contribution came from the “Virtual Liver Network” (VLN), a German program funded to the tune of 50 million euros by the country’s Federal Ministry of Education and Research (BMBF). The program focused on the largest and most complex organ in the human body: the liver.

In and out again, very fast
How long does a substance, once administered, remain in the liver? This computer simulation, created as part of a VLN project, demonstrates the progress of the concentration of a substance in the blood vessels in the liver at three points in time. The quantity of the substance present increases from blue (none) to green to red (maximum concentration).

More than 200 scientists from 36 institutions were involved in a total of 44 subprojects. Lars Küpfer and his colleagues from Bayer Technology Services were among them. “Until today, the VLN has been the only program worldwide to focus so comprehensively on one single organ in one country, and with such a high level of financial support,” says British doctor Adriano Henney, who coordinated the VLN as programme director.

Although the program focused on a single organ, it covered many different subsidiary aspects. Some projects concentrated on what was taking place in individual liver cells or with the ways cells communicate with each other. Others looked at large sections of tissue or the organ as a whole, while still others looked at the interactions between the liver and the rest of the body. Some projects aimed to bring together the different physiological levels, and in particular to make them relevant to clinical practice. This “Vertical Integration” work package ran under the leadership of Bayer Technology Services.

A few clicks on his keyboard are enough for Lars Küpfer to demonstrate the results of their work. In a video presentation we see how a substance is carried through the bloodstream into the liver. Within fractions of a second, the substance spreads into even the finest blood-vessels, leaving the organ again a short time later via the venous system. It is a computer simulation, developed by Küpfer and his colleagues together with researchers in Bremen and Aachen. “In the model, experimental data move through the vascular structures into the liver,” Küpfer explains. “But at the same time, the mass balances also go in, meaning information about how and at what speed a substance is metabolized in the liver. These inputs are specified on the pharmacological side.”

Dr. Jörg Lippert and Dr. Lars Küpfer at work
Capturing events in the body in formulae and curves: systems pharmacologists
Dr. Jörg Lippert and Dr. Lars Küpfer at work.

But this is not all that the model can do. It can also be used to predict, for example, how behavior changes in a damaged liver, or even a dead one. “Ultimately, we were able to show that these simulations correspond with experimental data to a very high degree,” Küpfer is pleased to say. VLN director Adriano Henney believes the various contributions made by Bayer Technology Services to have been “extraordinarily valuable.” In every form of science, it is vital to make a connection between pure theory and practical application. “Because of Bayer Technology Services’ close proximity to the pharmaceutical business, they are well aware of both business demands and actual patient needs. This aligns their developments very much toward real-life application, and adds greatly to the value those developments represent.”

It was no coincidence, therefore, that the experts at Bayer Technology Services also led the “Clinical Translation” work package. Their goal: to put the knowledge garnered from research (some of which was purely theoretical) into a useful clinical form – i.e. into a form that could be used directly with patients.

In one project, Bayer Technology Services researchers, working with partners from the Institute of Clinical Pharmacology in Stuttgart and the University Hospital in Dresden, investigated how quickly six different substances could be broken down in the liver when all of them were introduced into the body at the same time. This substance cocktail was chosen with deliberate care so that, in breaking down each substance, a different liver enzyme would come into play. For the researchers, it was particularly interesting to find out whether the individual substances would be broken down at different speeds by different patients. “Any dissimilarities might then be traceable to variations in the genes of the enzymes involved, which would make them dependent on the so-called genotype,” Küpfer explains. “We then used the computer model to investigate whether relationships of this kind between the genotype and the breakdown rate exist.” 

The process engineer has a clear goal in mind: “It is conceivable that, one day, we will be able to customize the dosages of a particular substance to match each specific genotype.” In this way, the very best therapy could be determined for each individual patient.

The work also has practical relevance in terms of the interactivity of different drugs. “If two drugs are broken down by the same liver enzyme, it can lead to interactions,” Küpfer says. The extent of such interactions can then be evaluated using computer modeling to determine how great they could be in an extreme case, taking into account the different genetic and physiological co-factors of the particular patient.

For their differentiated approaches, the systems biologists at Bayer Technology Services make use of a software platform that includes the PK-Sim and MoBi programs; the platform was developed in-house and has been expanded continuously over the last ten years. The abbreviations stand for the simulation of the pharmacokinetics of substances, and the modeling of biological systems, respectively. And in fact, based on very little experimental data and a number of fundamental individual physiological values, it is now possible to predict how a drug could circulate through the body in, for example, young children or older people. “By using this program, we are able to take into account such circumstances as the fact that, in younger children, the liver makes up a larger proportion of their body weight,” Küpfer explains. Among older people, by contrast, both body fat as a percentage of weight and the speed of blood flow are reduced. Both factors influence the metabolization of substances. “We can represent these kinds of things with our software platform extremely well,” says Küpfer.

“The developments made by Bayer Technology Services are extremely practical and valuable.”

Dr. Adriano Henney

Programme Director, Virtual Liver Network

90 liters per hour

Without the liver, nothing works. Weighing in at around two kilograms, the organ produces important messenger substances and proteins – used in the blood clotting process, for example – and plays an important role in metabolizing carbohydrates and fat. Some substances are also stored in the liver, while others are eliminated by it, especially foreign substances such as medications. If this happens too fast, the drug does not have enough time to be completely effective. If it happens too slowly, then it can lead to an increased risk of side effects. For this reason, the liver plays a key role in the development of therapeutic drugs. With blood flow of around 90 liters per hour, the organ is well supplied, with blood coming from both the venous and arterial systems. In modeling the liver, these complex and often interrelated functions represent a very special challenge.

Küpfer and his colleagues have continued to push ahead with the development of PK-Sim and MoBi in the course of their VLN projects. “With the current version, we’re in a position to be able to make even better predictions about individual patients,” the process engineer is pleased to say. For instance, the experts were able to gain a good understanding of individual data from clinical studies by taking into account the specific physiology of each of the patients taking part.

For pharmaceutical companies involved in research, such a tool is invaluable. “Being able to better predict the behavior of substances for specific patient populations or even for individual patients helps us to plan clinical studies more effectively, and thus to optimize our clinical development program,” explains Dr. Jörg Lippert, who is responsible for the Clinical Pharmacometrics department at Pharmaceuticals. 

Adriano Henney would agree: “During my career, I’ve seen a lot of candidate substances fail during phase II and III clinical studies. If theoretical models can help us avoid those kinds of misguided development efforts at an early stage, we will be able to save valuable resources – and put them to use elsewhere, in the interests of patients.”

The VLN coordinator is happy with the results of the immense five-year program in other ways, too. “We have achieved a lot in that time, and learned a great deal,” Henney says. “Many processes can now be modeled.”

Ultimately, this also opens up new possibilities in pharmaceutical research, but there is still much to be done. Regardless, Lars Küpfer takes a long-term view of his work. Why shouldn’t we one day be able to model the human body so well that we can accurately foretell the effects of particular substances using complete virtual patients? Is that going too far? “Why?” Küpfer asks, then points to the development of the airplane. “There was a time when hardly anyone could even conceive of the possibility that airplanes might one day be designed largely on computers. And today, that’s exactly how it’s done.” 

(*) Required
Your opinion? Comment now