“Different culture systems for different purposes”: an interview with Dik van Gent

Written by RegMedNet

In this interview, Dik van Gent, Erasmus University, discusses the challenges and future potential of ex vivo modelling systems such as organoids and organs-on-chips.

Dik van Gent

Dik van Gent studied biology in Utrecht and did his PhD research at the Netherlands Cancer Institute (Amsterdam), investigating the HIV DNA integration reaction. After receiving his PhD in 1993, he conducted post-doctoral research at the National Institutes of Health in Bethesda (MD, USA). In 1996 he moved to the Erasmus University Rotterdam (now Erasmus MC), Department of Molecular Genetics, and has since received funding from the Netherlands Scientific Organization, the Netherlands Cancer Foundation, the Association for International Cancer Research and the European Union. He has authored approximately 70 peer-reviewed papers and is currently a board member of the Netherlands Society for Radiobiology and director of the Research Master program Molecular Medicine.

Please introduce yourself and your institution

I have been in this department for 20 years and during that time I have been working on DNA repair and DNA damage responses. The whole department of molecular genetics at Erasmus Medical Center is centered on these areas, how they work mechanistically and how we can exploit our knowledge to increase the efficacy of treatments for patients. In the past years we have moved towards translational research, particularly on tumors, so that we can apply our knowledge to a setting that is clinically most relevant.

Last year, you published a paper on ex vivo tumor culture systems for functional drug testing in Future Science OA. Can you explain how this research came about?

Our group started this type of research approximately 6 years ago. We moved from cells and mouse models towards using material derived from clinical specimens because at that point a new treatment option for hereditary breast and ovarian cancer had appeared: PARP inhibitors. Many people thought that there could be a broader use for these inhibitors than only BRCA-gene mutated cancers, so we wanted to set up a functional assay to look for the function that the BRCA1 and BRCA2 proteins carry out, rather than the specific mutation in each protein. For this, we needed to use fresh, live material from the clinic.

We set up functional assays with these specimens and, in doing so, had to optimize growth conditions for these specimens. They are quite different from cell lines that you normally have in the lab because they have already been adapted to grow in environments other than, for example, plastic. We can now grow thin slices of tumor for about a week, allowing us to test various treatments on them.

What are the advantages of ex vivo disease modelling systems?

I think the major advantage is that it’s the closest you can get to what’s actually happening in the tumor; the material comes from the tumor and there are no growth steps in between. That means that in just a few hours, we can have it in culture and see, over time, how it develops.

A secondary advantage is that they help reduce the need for animals. Some models, especially patient-derived xenograft models, require quite a lot of animals for studies and we think we can prevent the use of these animals. In some aspects, these systems may even be better than patient-derived xenograft models.

How can ex vivo modelling systems help identify new therapies and modes of action?

I think it’s two-fold: one way is that we can use some of these models for medium-throughput drug screening. In organoid-type models, stem cells from the tumor are cultured into several tumor spheroids from the primary tumor. You could have a whole panel of breast tumor organoids on which you try out various treatments and combinations of treatments.

The other is better diagnosis and personalized medicine approaches, where we can test the individual patient tumor through a biopsy and test which treatment would be best for this particular patient. For this application, tumor organoids are not as well suited as they require several months to culture; I think different culture systems can serve different purposes.

What are the challenges of developing biomimetic ex vivo modelling systems?

I think one of the main challenges is to mimic the natural environment of the tumor as much as possible. I think it’s very important to aim for a culture system that maintains tumor tissue growth and characteristics for a prolonged period of time. That means for several days the tumor cells should keep on proliferating, should not start to die and should maintain tissue architecture.

I think it is very important that the people who establish these growth systems make a good effort to characterize their tissue integrity every time so we know that what is done to culture the material doesn’t itself cause changes. I think that’s another of the main challenges at this moment; there are no real standard procedures so that different researchers may use different methods, making it difficult to compare what researchers are doing. A major challenge will be to develop a more uniform way of performing these kinds of experiments.

How close are we to existing engineered models replacing in vivo models? What more is required to get there?

I think I see three major areas where we still need to improve the systems. One of them is to prolong tissue growth. The challenge is that if we keep the growth of cells constant, as soon as the tissue expands we have a problem. There is no blood supply so the thicker your tissue is, the less well nutrients can penetrate the middle of the tissue and it can go into necrosis. We need something to deliver nutrients into the center of the tissue as this will be important for longer term experiments.

Another one is mimicking other organs. At the moment, a key reason to do an animal experiment is that you can also measure normal tissue toxicity, which is required for all kinds of regulatory reasons. You can’t put your drug into a patient without showing that the toxicity is not too high.  The way to go here might be a battery of various organoids of normal tissue. Such a battery could show that the major organs are not too affected by your treatment, although I think it may be a challenge to completely replace animals.

The third real challenge is to mimic interaction of the tumor with immune cells. In this situation, ex vivo models might be even better than animal models, because immune cells can be added to these tissue slices to look at the interaction. This can’t be investigated very well in vivo because it is difficult to put human immune cells into an animal model.

What are your predictions for the development of engineering models in the future?

A possible improvement would be to go from culturing tissue sections ex vivo to a ‘cancer-on-chip’ approach. You could put the tissue in a microfluidics system where media flows over the tissue in a more controlled fashion. One might imagine the development of something that looks like the blood supply, flowing through the chips. This is being worked on and I think it will enable us to grow much larger pieces of tissue than we can at the moment.

In order to mimic other organs, I think we will be able to make libraries of organoids, where a set of organoids can be subjected to certain treatments. This will enable you to see how your treatment could affect various organs in the body as a first stage of investigating how normal tissue will react.

Adding the complexity of an immune system is something that we have to find out more about. There are some really promising ideas that could help, particularly in the area of immune therapy where personalized medicine is something that still needs more development. In many of trials of immune therapy, only a small percentage of the patients benefit from the treatment. It would be great if you could find out beforehand whether the patient would benefit: yes, or no. 

Read Dik’s Future Science OA review: Ex vivo tumor culture systems for functional drug testing and therapy response prediction