Tuesday, 28 March 2017 06:28

Working group – a Portrait: Leibniz Research Centre for Working Environment and Human Factors (IfADo) Featured

InVitro+Jobs presents scientists and their innovative research in a regular feature called “Working Group – a Portrait”. We focus on newly developed methods, their evaluation and their potential for reducing and where possible replacing animal experimentation according to the 3R principles of Russel & Burch (reduce, refine, replace).


Working group – a Portrait:
IfADo – Leibniz Research Centre for Working Environment and Human Factors,
Department of Toxicology, Research Group Systems Toxicology

The IfADo – Leibniz Research Centre for Working Environment and Human Factors in Dortmund is a research institute, where scientists investigate the growing challenges in the working environment in the face of an increasingly ageing population (ergonomics), regarding the part played by the immune system with regard to infections and cancer (immunology), the part played by the cerebral cortex and neuroplasticity in experiencing, behaviour, learning and memory formation (psychology and neurosciences), and with toxicology, investigating the toxic processes caused by chemicals in the liver or the nervous system (1).

Image 1: photo of the department of toxicology.
Image source: IfADo

Head of the Department of Toxicology is the physician and toxicologist Prof. Dr. med. Jan Hengstler. The department investigates the exact mechanisms and health effects of chemicals to which people can be exposed at the workplace. The department comprises four working groups: the Research Group Systems Toxicology, the two Junior Research Groups Liver Toxicology and Cellular Toxicology, and the Networking Group Neurotoxicology and Chemosensation (2). The groups Liver Toxicology, Neurotoxicology und Cellular Toxicology provide Systems Toxicology with their data for the development of computer models. A main focus is on the development of more conclusive and faster test methods.

Research into new methods for replacing animal experiments
In order to avoid or at least reduce animal experiments, the IfADo scientists research new methods suitable for replacing animal procedures. To this end, together with scientists from Switzerland, Belgium and Egypt, the Junior Research Group Cellular Toxicology headed by Dr. Rosemarie Marchan has developed a chip technology the size of a mobile phone. The chip platform combines liver and intestine organoids, minute spherical organ units. Pharmaceuticals are first metabolised into their active molecular form by liver organoids and then their effects on intestine organoids are investigated (3). A few months ago, the research team received the Global 3R Award for its development.

Image 2: Liver organoids. A. The organoid comprising hepatocytes possesses a layer of endothelial cells and forms bile canaliculi (green). Such systems can be used to research whether substances are transported in bile or inhibit the transport of bile salts. B. Specific immune system cells are integrated into the organoid (Kupffer cells, green). This is part of current research initiatives to integrate immune system functions into organoids.
Image source: Hengstler, IfADo

In this method, the organoids are loaded into the tilted platform und circulate to the microchambers on the platform. In the microchambers, which are connected by microchannels, the two parallel rows of tissue types are cultured under continuous perfusion with a nutrient fluid. After being cultured for eight days, the organoids are exposed to the test fluid, a chemotherapeutic agent. The therapeutic agent has a cytotoxic effect, in that it is first transformed in the liver by the xenobiotic metabolising enzyme Cyp2B6 into 4-hydroxycyclophosphamide and then transported to the target organ (e.g. a liver tumour). There it is transformed into acrolein (4) and binds to the (so-called adduction). As a result, a cell death mechanism (apoptosis) is triggered and the (cancer) cells die (5). The scientists were able to confirm that this is also the case in the minute organ-like systems. The flow rate can be regulated by tilting the chip platform. After perfusion of the test substance, all organoids are immunohistochemically stained and investigated, for instance using a confocal laser microscope.
Despite this success, the scientists point out that there are diverse differences between liver cells in the natural organ and liver cells cultured from induced pluripotent stem cells.

Liver cells at present only similar
The all-rounder iPSCs (induced pluripotent stem cells) can be obtained relatively easily nowadays, for instance from skin cells from probands or patients. One challenge that has not yet been satisfactorily solved is reprogramming to the desired cell type. This is not only necessary for questions regarding the replacement of animal testing, but scientists also expect a perspective for treating liver diseases, for instance using a stem cell therapy or in the more distant future the transplantation of artificial replacement organs.
However, the liver cells obtained from iPSCs are only similar to the natural cells, not identical. Therefore it is important for the IfADo researchers to develop methods that can detect and analyse these differences in the smallest details.

In a research project with cooperation partners from Saarland University, Charité - Universitätsmedizin Berlin, Dortmund University and experts from the United Kingdom and Sweden, IfADo scientists conducted so-called gene network analyses with hepatocyte-like cells developed from human embryonic stem cells and induced pluripotent stem cells, comparing the results with those from fresh isolated human liver cells (6). First of all, microarray experiments were conducted for all the gene expression analyses, to allow the scientists to see which of the approximately 22,000 human genes are active in the different cell types. The results were analysed using a CellNet computer algorithm.

Image 3: Comparison of liver cells in a real liver (top) and cultivated hepatocytes (bottom). The illustrations on the left show microscope raw data; the illustrations in the middle show reconstructions, with the cytoplasm in green, the cell nuclei in blue (hepatocytes often have two nuclei) and the bile canaliculi in yellow. in the illustrations on the right, the cytoplasm was removed so as to better depict the cytoplasm. Both in vivo and in vitro, the hepatocytes form bile canaliculi and display similar morphology.
Image source: Hengstler, IfADo

The researchers found out that hepatocyte-like cells display the characteristics of a combination of liver cells, small intestinal cells, connective tissue cells and stem cells, thus they are a kind of multi-organ cell type. The hepatocyte-like cells corresponded with real hepatocytes to only 68 per cent. Toxicological tests, however, require mature hepatocytes with all their metabolic characteristics. It was also observed that the hepatocyte-like cells developed from stem cells also formed transcription factors1 that caused an undesired development towards the intestine. The goal is to improve the differentiation of stem cells into hepatocytes to develop improved in vitro test systems that can replace animal experiments in toxicity testing (7).

Research with practical applications
The research results find practical application for instance in commissions for the evaluation of maximum workplace concentrations or in EU projects such as EU-ToxRisk, where feasibility studies for long-term toxicology (lung, liver, kidneys and nervous system) and developmental and reproductive toxicology are to be conducted with new, human-specific methods using human cells, tissues or organ-like systems, Omics technologies, systems biological approaches and computer models.
The intention is to clarify the harmful influence of chemicals on humans by means of mechanistic investigation of so-called adverse outcome pathways (AOPs) in cells, tissue and organs. Adverse outcome pathways describe a process from a triggering molecular event that in the course of time leads to a damaging effect through a chemical stress factor and a cascade of biological reactions through different biological levels (organelle, cell, tissue, organ, entire organism).
The aim of the EU project is to replace animal experiments by allowing toxic substances to be identified in advance with the help of animal-free methods so that they then do not need to be evaluated in animal tests (8).

Initiative Virtual Liver
Prof. Hengstler is a member of the Scientific Leadership of the project cluster Virtual Liver Network, a national initiative funded by the German Federal Ministry of Education and Research (BMBF). This initiative’s goal is to be able to achieve true multi-scale modelling of liver physiology within a single organ and thus come to a dynamic understanding that can drive forward the application of modelling and simulation to medical practice (9). Within this framework, Prof. Hengstler has concentrated on the mechanisms of liver regeneration and developed a computer model for understanding these mechanisms. The initiative Research Network Systems Medicine of the Liver - LiSyM is a continuation of the Virtual Liver project with the goal of being able to predict the course of liver diseases with the aid of a mathematical model.
Prof. Hengstler’s research team is currently investigating molecular mechanisms that can lead to a liver disease, including a project on the role played by the protein Rab 18 and its effects on fatty liver disease, and a project to adapt a computer model for treating chronic liver damage.

“...the major breakthrough is a very complex long-term goal...”

We interviewed Prof. Hengstler on the status of current developments in the field of human liver models as replacement methods for animal experiments in long-term investigations and how he sees the prospects of ending animal experiments in the area of toxicity testing by 2025.

scientists need cells, tissue or organoids in order to be able to replace animal experiments in regulatory toxicology. What is the current situation: Which animal experiments can easily be replaced and where are there still difficulties?

Prof. Hengstler:
In vitro methods are good for recognising whether substances have a mutagenic effect, trigger irritant effects, or either inhibit or induce enzymes; in the past we needed animal experiment to investigate these effects. However, it is difficult replacing animal experiments, in which substances are tested for more than 28 or 90 days. That is currently the subject of intensive research.

How long can scientists culture cells, tissue or organ-like systems and use them in tests nowadays?

Prof. Hengstler:
It is relatively easy to maintain cell systems for one or two weeks. If nutrients, growth factors, pH value and other critical factors are kept constant and contaminations are avoided, some systems can be cultured for 3 months or longer.

Do these longer periods of use have advantages? Can the systems be used to replace long-term experiments, in which substances are administered to animals over a period of 28, 90 or even more days?

Prof. Hengstler:
That is a decisive question, because simply being able to keep a cell system in a culture for a longer period of time does not guarantee that it becomes more similar to the in vivo situation, i.e. a real organ in a human body. Another as yet unanswered question is whether a culture system, which for instance is treated for a year with a liver-damaging substance, develops similar diseases to the liver of a patient. Just one example: Some substances cause healthy hepatocytes in the liver to be replaced by connective tissue cells. This leads to fibrotic bands of connective tissue, which can constrict blood vessels. This disturbed blood flow due to the fibrotic bands can lead to death. Although altered cell functions and cell death after long-term exposure to hepatotoxic substances have been observed in currently available cell or tissue systems, but the connective tissue strands typical for hepatic fibrosis2 cannot yet be reconstructed. This shows that these tissue systems do not react exactly as a real organ does, even if they can be kept in culture for a very long time. However, something else is already possible: When we expose a culture with hepatocytes to a liver-damaging substance, after just one day we can see an increase of certain biomarkers that are normally associated with an elevated risk of liver fibrosis. This demonstrates why two strategies are currently being pursued: one is to “reproduce” organs (or similar structures) in culture, so as to cause there the damage and diseases that occur in a real organ. The second strategy is based on simpler cell systems and biomarkers; the results are then so interpreted that a rise of a particular biomarker in a short-term experiment means that a more prolonged exposure means an elevated risk of disease. In our laboratory we pursue both strategies and cannot say yet which one will prevail, even if we have recently had more success with the second approach, simpler systems combined with well-researched biomarkers.

How do scientists know that a miniature organ is enough for simulating the human organ? When can a miniature organ or cell system even be seen to be suitable?

Prof. Hengstler:
We have set up a database to answer this question, in which we have compiled all available information on substances und their effects on human organs. For instance, it is known that the drug paracetamol almost always leads to liver damage when its, concentration in the blood exceeds a certain value, in this case approximately 1 millimolar (mM)3. Thus if an in vitro system with hepatocytes or a miniature liver does not develop any damage at a concentration of 1 mM paracetamol, then the system is no use. Fortunately, the toxicity of paracetamol can be proven correctly in our in vitro systems, i.e. at the right concentration. If one then conducts this experiment with several hundred substances, one gets an overview of the in vitro system’s accuracy and usability. The in vivo situation is quite accurate for the vast majority of the substances, but for certain substances one would underestimate the toxicity for the liver if one used the current in vitro system, as it unfortunately does still fail in some situations.

What is the problem in such a case?

Prof. Hengstler:
The problem is that none of the currently available systems with artificial mini-organs has a proper capillary network that supplies blood, therefore an interaction with immune cells such as takes place in the real organism is not possible. For this reason the currently available artificial miniature livers fail when it comes to recognising those toxic substances that also affect immune cells. One of the next big challenges will therefore be to develop prediction systems that are also suitable for these mechanisms.

One major area of in vitro systems research is the attempt to generate adult somatic cells from stem cells or precursor cells. What consequences does it have regarding suitability for replacing animal experiments when hepatocyte-like cells developed from stem cells only correspond to hepatocytes by 68 per cent?

Prof. Hengstler:
The similarity score, in this case 68 per cent, is a major methodological advance. It is ascertained by determining the messenger RNA (mRNA)4 of all the genes, of which humans have about 22,000. This is quick and cheap using today’s analysis methods. These mRNAs in hepatocyte-like cells derived from stem cells are then compared with all the mRNAs from real hepatocytes and a mathematic algorithm then calculates a similarity score. The advantage is that no genes are omitted from the evaluation and the score therefore provides an objective parameter. In the past, some scientists only picked out those genes that provided great similarity to real hepatocytes and therefore got a much too optimistic picture. In practice, 68 per cent means we are still too far away from real hepatocytes. We therefore still use cultivated primary hepatocytes in our in vitro systems, i.e. hepatocytes originally isolated from human livers. It would be a great step forward if we could finally use the more readily available hepatocytes from stem cells, but that will only make sense when we have achieved a similarity score of at least 90 per cent.

What factors must pre-exist for the cells to feel “good” enough that the similarity scores will rise?

Prof. Hengstler:
We are pursuing two strategies. For one thing we know which controlling factors in the available hepatocyte-like cells are either hypoactive or hyperactive, especially gene transcription factors. This is why in one of our projects we are trying to get these hypoactive or hyperactive transcription factors into the right state. A second strategy is to recreate the real liver’s micro-environment with its complex pattern of cytokines, chemokines and their extracellular matrix in vitro. We will then combine the two strategies, depending on how promising the results are.

Scientists are trying to cultivate the most important human organs in miniature format and connect them to be able to simulate a systemic approach. American researchers* have reported that they are already able to connect 12 organs with each other. What is the current state of development of artificial tissues and miniature organs for replacing animal experiments in Europe?

Prof. Hengstler:
After the progress made in the field of microfluidics we have used this technology to connect several organ systems in vitro, for instance in our papers Frimat et al., 2011 (10) or Frey et al., 2014 (11). One advantage of these fluidic systems is that the flow can be regulated; thus the medium can first flow through a miniature liver, where certain chemicals are metabolically activated, i.e. only then are they transformed into their toxic state, and are then transported to another organ, in which they develop their toxic effects. Substantially more complex fluidic systems are also possible. Now that this has worked well, American groups are also excited by the technology. In principle this is a technical prerequisite for investigating communication between organs. However, in my opinion the critical restriction is not that we are connecting many in vitro systems; if one wanted, it would already be possible to interconnect hundreds of fluidic cells. The problem is rather that the individual artificial organs still have critical restrictions. If for instance there are no micro-vessels, the artificial organ cannot interact correctly with immune cells; if there is no innervation, infused neurotransmitters would have no benefit, etc. For this reason we are currently working on perfecting the in vitro systems before we reconnect them on microfluidic chips.

How should one imagine a prediction model? What data are fed into it?

Prof. Hengstler:
Prediction models or system simulations are very important in the context of in vitro systems. The advances made in computer sciences allow us to simulate tissue reactions using model calculations. This can be very helpful when very many different factors affect an endpoint, for instance on a metabolic pathway. That is how we identified a possible treatment for a certain disease, hyperammonaemia (Ghallab et al., 2016 [12]; Schliess et al., 2014, [13]). It would have been very difficult to achieve this simply by the power of thought or conventional calculation.

Further prediction models are the physiologically based pharmacokinetic models, PBPK for short. These models calculate for instance which concentrations of a substance in the blood and organs results, when a certain dose is ingested orally. This is important because in vitro systems have neither oral ingestion nor kidneys, liver or intestines that would eliminate the substance again. One can therefore use PBPK models to calculate which substance concentrations should best be used in an in vitro system. Unfortunately, PBPK models are still relatively inaccurate when we know little about a substance.
A third model type deduces the possible toxicity of a substance based on the similarity of the chemical structure, for instance if there is a higher probability of a mutagenic effect. Die accuracy of these prediction models especially depends on how good and comprehensive the data is that they work on.
Despite all the limits these prediction models still have, I still think they are very important. An intelligent combination of in vitro systems with suitable model simulations will probably deliver greater advances in the future.

If the data from prediction models is based on animal experiments, doesn’t this yet again cause problems regarding comparability due to species differences?

Prof. Hengstler:
I think so, too: Animal tests (or also strategies with in vitro systems based on animal cells), have to take into account the problem of interspecies differences. Some good, although not yet perfect, techniques for interspecies extrapolation already exist. In vitro systems in general, including those based on human cells, have to deal with the difficulty of extrapolating from the simpler in vitro system to the more complex in vivo situation. At present the greatest measure of safety for humans is achieved by combining the advantages of both strategies.
What is your assessment of the perspective from now until the year 2025? Do you think it will be possible to replace animal experiments in the area of regulatory toxicology?

Prof. Hengstler:
I assume that considerable progress will be possible by 2025. We will be able use in vitro tests to obtain some information regarding toxicity that at present is usually gained from animal experiments. For instance, we can expect to have reliable pretests in a few years. This will allow us to identify from many substances those that are so dangerous that they are best sorted out immediately and not even tested on animals. However, the major breakthrough, a complete simulation of humans and test animal in vitro and in silico, which would mean that we do not need any organisms at all for experiments, is a very complex long-term goal that won’t be solved in the foreseeable future. I can imagine that this will pose a challenge for generations of scientists.

* Prof. Anthony Atala from the Institute for Regenerative Medicine at the Wake Forest School of Medicine in North Carolina

(1) www.ifado.de
(2) www.ifado.de/toxikologie
(3) Jin-Young Kim, David A. Fluri, Rosemarie Marchan, Kurt Boonen, Soumyaranjan Mohanty, Prateek Singh, Seddik Hammad, Bart Landuyt, Jan G. Hengstler, Jens M. Kelm, Andreas Hierlemann & Olivier Frey (2015): 3D spherical microtissues and microfluidic technology for multi-tissue experiments and analysis. Journal of Biotechnology 205: 24–35.
(4) Aktories, K., Förstermann, U., Hofmann, F.B. & Starke, K. (2009): Allgemeine und spezielle Pharmakologie und Toxikologie. Repetitorium. Verlag Urban & Fischer.
(5) Marquardt, H., Schäfer, S. & Barth, H. (2013): Toxikologie. Wissenschaftliche Verlagsgesellschaft, Stuttgart.
(6) Patricio Godoy, Wolfgang Schmidt-Heck, Karthick Natarajan, Baltasar Lucendo-Villarin, Dagmara Szkolnicka, Annika Asplund, Petter Björquist, Agata Widera, Regina Stöber, Gisela Campos, Seddik Hammad, Agapios Sachinidis, Umesh Chaudhari, Georg Damm, Thomas S. Weiss, Andreas Nüssler, Jane Synnergren, Karolina Edlund, Barbara Küppers-Munther, David C. Hay & Jan G. Hengstler (2015): Gene networks and transcription factor motifs defining the differentiation of stem cells into hepatocyte-like cells. Journal of Hepatology 63: 934–942.
(7) http://www.ifado.de/blog/2017/01/04/forschungsprojekt-von-der-stammzelle-zur-leberzelle
(8) http://www.ifado.de/toxikologie/2016/10/10/eu-projekt-gefahrenstoffe-effizienter-und-tierversuchsfrei-bestimmen/
(9) http://network.virtual-liver.de/de/
(10) Frimat, J.P., Becker, M., Chiang, Y.Y., Marggraf, U., Janasek, D., Hengstler, J. G., Franzke, J. & West, J. (2011): A microfluidic array with cellular valving for single cell co-culture. Lab Chip. 11(2): 231-237.
(11) Frey, O., Misun, P.M., Fluri, D.A., Hengstler, J.G. & Hierlemann, A. (2014): Reconfigurable microfluidic hanging drop network for multi-tissue interaction and analysis. Nat Commun. 5: 4250.
(12) Ghallab, A., Cellière, G., Henkel, S.G., Driesch, D., Hoehme, S., Hofmann, U., Zellmer, S., Godoy, P., Sachinidis, A., Blaszkewicz, M., Reif, R., Marchan, R., Kuepfer, L., Häussinger, D., Drasdo, D., Gebhardt, R. & Hengstler, J. G. (2016): Model-guided identification of a therapeutic strategy to reduce hyperammonemia in liver diseases. J Hepatol. 64/4: 860-871.
(13) Schliess, F., Hoehme, S., Henkel, S.G., Ghallab, A., Driesch, D., Böttger, J., Guthke, R., Pfaff, M., Hengstler, J.G., Gebhardt, R., Häussinger, D., Drasdo, D. % Zellmer, S. (2014): Integrated metabolic spatial-temporal model for the prediction of ammonia detoxification during liver damage and regeneration. Hepatology 60/6: 2040-2051.

1 Transcription factors: a protein, that acts as a kind of „bookmark“ and binds to DNA or a protein complex, thus activating the RNA polymerase and producing a matrix (messenger RNA) at a specific site of the DNA.
2 Hepatic fibrosis: reconstruction process in the liver, in which liver tissue is replaced by connective tissue.
3 Millimolar: unit of mass concentration, 1 Molar corresponds to 1 mol/litre, i.e. approx. 6.022 times 10 to the power of 23 particles, 1 millimol = 1/1000 Mol.
4 Messenger RNA: the single strand transcription unit of a DNA segment. It acts as a matrix for the production of proteins.