Tumor samples exposed to a new three-dimensional model of human skin developed by the Wistar Institute reflect how tumor cells respond in vivo more accurately than those grown in monolayer tissue culture, according to an institute official.
This finding suggests that organotypic, or tissue-like, models can be used to screen drug candidates ahead of animal testing. Such a model could enable researchers to “characterize all the different cell types, and look at the role of stem cells in transformation, look at characterized cancer stem cells, and look at the influx of stromal fibroblasts or endothelial cells, so you can make artificial blood vessels,” according to one of the scientists who developed the new model.
Meenhard Herlyn, a professor and chair of the molecular and cellular oncogenesis program at the Wistar Institute, talked about the 3D models of human skin developed in his lab during a presentation at the Society for Biomolecular Sciences conference, held in King of Prussia, Pa., last week.
Herlyn said that today, it is possible to create quite complex structures using four or five different cell types, which mimic much of what is seen in human skin.
“I argue that [drug makers] have spent a lot of time and money on drugs that have some effect in animal models, but were not really curative” in human subjects, Herlyn told CBA News following the conference.
These compounds also did not show much effect in 3D models, so “the argument is that when testing compounds for their potential as lead compounds, it is good to add one more layer of scrutiny before one invests a lot of effort and money into animal and patient studies,” said Herlyn.
Such a layer of scrutiny could comprise 3D cultures in which cells are embedded in a collagen matrix and therefore behave differently than if they had been exposed to the drugs and to the growth medium.
“I think the [3D] models are ideally suited for many of the pharmacological studies, even drug uptake or the influx of hypoxia or necrosis,” said Herlyn. Any studies one could do in solid tissues, one could do under much more controlled conditions in these so-called organotypic cultures, he said.
“I am quite convinced that in the next decade, [3D cultures] will become a routine to sort out which drugs work best, and which combination of drugs work best,” Herlyn said.
Like any of the more sophisticated model systems, 3D cultures are not easily amenable to high-throughput screening. However, efforts are underway “in our own laboratory” to change that, said Herlyn.
“I am quite convinced that in the next decade, [3D cultures] will become routine to sort out which drugs work best, and which combination of drugs work best.”
“We call it medium-throughput screening. We think that we can easily test several thousand compounds at a time in 384-well plates.” But he and his team cannot test tens of thousands of compounds, such as a pharmaceutical company might do in a high-throughput campaign, because the readouts will be a little more complicated and are likely to be based on imaging techniques.
In addition, much of the technology still needs to be scaled up. “The big challenge for the next year is to scale up these models for higher throughput. That is not yet completely worked out,” said Herlyn.
“You could, as a compromise, first test for activity in a [monolayer] culture approach, then follow up with a 3D model, and then go into animals,” Herlyn said. To use these structures in the first screen may be too involved, and perhaps the technology has not evolved to that point yet. However, after the primary screen, researchers have too many candidate compounds to go right into animals, and they need to sort through those and weed the less effective and more toxic ones out.
“Our [belief] is that these 3D cultures are suited to further weed out compounds before one goes through more pharmacological modifications or before one goes to animals, almost like a secondary or tertiary screen,” said Herlyn. The technology is still too low-throughput for a primary screen.
In 1976, the Wistar Institute was making monoclonal antibodies against certain cancer cells, said Herlyn. When he and his team characterized these antibodies, they realized that when they compared normal melanocytes and malignant melanoma cells, the melanocytes that they grew in culture were reacting similarly to melanoma cells.
Herlyn said his lab was surprised by this because these antibodies did not bind to melanocytes in skin, but they were binding to melanocytes in culture. “So we thought that we must somehow have transformed them,” he said. Then his lab learned to grow keratinocytes, and “lo and behold, if we mixed the melanocyte and the keratinocyte, it makes a lot of sense because the cells live close together in our epidermis,” he added.
Within three days, all the tumor markers that were previously expressed on the melanocytes disappeared because the keratinocytes shut them down. “That observation led us to realize how important it is that cells continue to be in communication with each other, even if they are in culture,” said Herlyn.
Herlyn’s group realized that this culture method of combining melanocytes and keratinocytes is suited to understand the biology of normal skin, but also to understand how tumors develop, because one can manipulate cells, and up-regulate a gene or down-regulate a gene, and one can do that with two or three genes within a cell.
This can be done under highly controlled conditions, Herlyn said. Keratinocytes and melanocytes can be grown in different dishes and then mixed together, and “they do all the same things they do in real skin,” said Herlyn. If this is done with malignant cells, it is important to know what stage of cancer they came from, because those cells will do the same things as cancer cells in patients, and one can create lesions that are histologically very similar to patients’ lesions, Herlyn said.
This [3D] system has been understudied, because researchers have used genetic models in mice to recreate diseases. “I think research has given us a lot of insight into a lot of genes, but when it comes to the biology of some of the organs, mice are not humans, and the skin of a mouse looks very different from that of a human, so there are limitations to a murine model,” Herlyn said.
The murine genetic model and the synthetic skin model should be used in parallel, based on the inherent strengths of each model, Herlyn said.
“That is what we have been doing over the last few years, and what we will do increasingly in the future, because the model is really very versatile and reproducible,” said Herlyn, who added that the 3D model is much cheaper than making murine genetic models.
“The next step is to, first of all, get these cultures highly reproducible, and get a very clear, black and white readout,” said Herlyn. That is most likely done through imaging, he said. Once the image is acquired, scientists can, with the appropriate software, do the image analysis relatively automatically, Herlyn explained.
“The major issue is that you want to test in the same dish normal and malignant cells,” Herlyn said. Cells can be made visible when transfected with different dyes. “That way you know which cells die, and which cells remain alive,” he said.