Dr. Aziz Cayli knows CHO (Chinese hamster ovary) cells so well that he decided to establish his own company in 2005. Prior to establishing CellCa GmbH in the city of Laupheim, Cayli held posts at Roche in Penzberg and Boehringer Ingelheim in Biberach. CellCa is a specialist in upstream processing and with its current total of 15 staff, the company achieved revenues of around two million euros in 2009. We talked with Dr. Aziz Cayli who did his doctorate on the metabolism of CHO cells.
Yes, it is still the most common cell type used for the production of recombinant proteins. However, many companies today have their own CHO clones and CHO subtypes.
As far as I am aware, it is actually a matter of pure chance that the pharmaceutical industry is working with CHO cells to the extent that it does. However, it is not a matter of chance that industry is working with cell cultures and cell lines. The cell lines, whether they are BHK, CHO, human or NS0 cells, have the advantage of being derived from cancer cells, and hence do not die. Under suitable environmental conditions, these cell lines continue to grow infinitely. The fact that industry often uses CHO cells is entirely down to Professor Lawrence Chasin who established the CHO cell line in the 1960s.
Following the publication of this achievement, many laboratories further investigated the CHO cell line and developed industrial production processes. Industrial companies then found that the production of recombinant proteins worked well with the CHO cell line, and the pharmaceutical industry was able to benefit from these further developments. In the meantime, comprehensive experience and knowledge has been built up and it would not make sense for the pharmaceutical industry to switch to other cell lines, as this would take years, if not decades.
We are developing media and processes that try to mimic the natural environment of CHO cells as closely as possible. We use equipment, media and artificial conditions to create conditions that are similar to the conditions in the animal body/tissue, as this is what the cells need to grow and produce protein.Let me describe this in more detail: In animals, the blood supplies the cells with all the nutrients they need, including sugars, amino acids, antioxidants, vitamins and trace elements. We are attempting to provide the CHO cells with all these nutrients in the artificial media we use to culture them. The culturing process also needs to be adapted to the cells’ requirements. Blood has a pH of around 7.3 to 7.4 and we try to maintain similar values in the bioreactor. Blood also contains dissolved CO2, which is why we also supply the buffers and media we use with CO2. In addition, the reactors in which the CHO cells are cultivated run at a temperature of 37° C to mimic the body temperature of hamsters. In living tissue, cells are supplied with oxygen by way of haemoglobin that is contained in the blood. We pump oxygen into the medium, where it dissolves and is taken up by the cells. We do all this to keep the cells alive. The cells produce the sought-after protein as long as they are “happy” with the artificial conditions to which they are exposed. But it is worth pointing out that we need to genetically modify the cells in order to make them produce the sought-after recombinant proteins.
The first culture media were developed in the 1960s, shortly after the CHO cell line paper had been published. In their pragmatic endeavour to mimic the natural environment of the CHO cells, the researchers added serum based on the fact that the cells are supplied with blood, i.e. serum in the body. Initially, relatively high quantities of serum were used. The cell culture medium tended to contain 10 per cent serum. Serum contains many nutrients as well as many protective factors, proteins and growth factors. So the cells started to grow and divide. Over time, the use of serum turned out to be quite costly. The high cost of serum allied with new cell biological findings led researchers to reduce the proportion of serum in the culture media. In the 1990s, when I did my doctorate, we used to add about 2 per cent serum to the culture medium. However, it soon turned out that animal serum was associated with the risk of viral contaminants, prions for example. This is why further research focused on finding ways of developing culture media with no added serum.
Yes. At the beginning of the 1990s, public authorities stipulated that researchers and industry must avoid the use of animal constituents in culture media. Many companies put a great deal of energy into finding ways to produce culture media without serum. But when serum was no longer used, the cells grew much more slowly than before. However, they did not die.
Plant constituents, which do not contain viral contaminants, were added to the culture media in order to counteract the lower quantity of protein produced. Soy peptone, i.e. hydrolysed soy proteins and peptides, tended to be used as a substitute for animal serum. Many companies still use soy peptone today. However, the use of soy peptone is associated with the problem of non-reproducibility because plant peptones never have the same biochemical composition.
The regulatory authorities required companies to replace plant constituents with chemically defined media. Today, many such culture media exist, and the media we produce are among them. But a lot of research was needed to find out why soy peptone favoured the growth of cells. We needed to understand this in order to be able to develop media that consisted of the chemicals the cells required for growth.This is only a rough description of how modern culture media were developed. Many more factors had to be taken into account, for example the production of recombinant proteins and the growth-promoting proteins that were contained in serum. These proteins were isolated, expressed in E. coli and then added to the culture media. And this is still standard practice today. The development of culture media had to take into account many regulatory aspects, and it was decades before a satisfactory solution was found.
This question is relatively easy to answer. The decisive factor is the ratio of the nutrients contained in the culture media. I've often been asked how we manage to produce such effective media. But the only thing I am able to tell people is that we do not do anything new. We simply use a different ratio of nutrients. We have used different quantities of the 66 individual chemicals that are commonly used to prepare a culture medium. The ratio of these chemicals is key for the quantity of protein produced.
Both. We are aware of numerous biochemical pathways, but we only understand about half of them. We are trying to understand the rest through trial and error. This is an iterative process that can take months and even years.
When cells earmarked to be used for the production of proteins, we need to know whether they will remain genetically stable. This is important because the cells are engineered to produce a specific pharmaceutical protein. Some clones are genetically unstable although the gene has already integrated into the chromosome. When the cell starts to grow, the gene can either be switched off or deleted from the chromosome. This means that the cell continues to grow but no longer produces the desired protein. That would be a catastrophe. Before a clone is used for the production of protein, it is important to be sure that the clone will remain stable over a specific period of time, seven weeks is a standard measure.In addition, we need to find out whether the cell under investigation is suitable for up-scaling. We carry out many experiments on the small scale, involving 1-l- to 3-l reactors. But protein production takes place in huge tanks of between 1,000 and 10,000 litres. The chosen cell needs to have a certain stability and robustness in order for it to be able to grow under large-scale conditions. What I mean is that the cells need to be able to tolerate the shear forces that might occur in such big reactors.
Third, the production cell must be able to glycosylate proteins. Glycosylation conveys characteristics to proteins that make the proteins biologically active.
And fourth, the cell needs to be free of pathogenic contaminants such as viruses. Answers to all these questions need to be found when a particular clone is selected. The topic is highly complex. I have only mentioned the four major aspects.
Yes, the upstream process has to all intents and purposes got to the stage where no further improvements can be made. There is no reason to continue improving the productivity of the upstream process. Nowadays we are able to design processes that lead to yields of around 10 g antibody/litre. Let me give you the following example: a 1,000-l reactor produces ten kg protein in three weeks. About one kg is required for clinical phase I studies, three kg for clinical phase II studies and around five kg for clinical phase III studies. This makes a total of around nine kg. Therefore, one run is sufficient to provide almost all the material that is required for the clinical trials. And the clinical trials take around six to seven years to complete. And the same is basically true for routine production. Previously, with blockbuster proteins, 100 kg/year were required to supply the world market. This corresponds to ten runs in a 1,000-l reactor or one run in a 10,000-l reactor. Or in other words: I do not think that the further up-scaling of productivity is economical. It is necessary to adapt all the steps in the entire process chain, including the purification of the final product, to each other.
Nowadays we have 100% chemically defined, i.e. synthetic media that contain a well-balanced proportion of nutrients and supplements. We do not require any expensive and complicated recombinant proteins. We have shown that cells do not require such complex proteins to grow.
My dream is to develop a nutrient cocktail that will lead to higher product quality. I believe that culture media are able to change the quality, or in other words the characteristics of the protein. Some companies have an excellent approach, and that is what the future is all about.
What I have in mind is glycosylation, which affects cellular properties such as bioactivity, solubility, half-life etc. A protein’s glycosylation profile can be altered by exposing the cells that produce it to a specific nutrient cocktail. A protein consists of peptides, a chain of amino acids to which carbohydrate molecules are attached. These sugar molecules can be complete or incomplete, sometimes with two and sometimes with three side chains. These carbohydrate structures determine the activity, stability, solubility, etc. of the therapeutic proteins and their half-life in the blood. My dream is to make it possible to change the composition and distribution of the protein’s sugar molecules simply by altering the composition of the culture media. In other words, to interfere with the cellular metabolism.
Yes indeed. Up until now we have mainly been focused on the high-yield production of proteins, and on ways to make cells produce ever larger quantities of a specific protein. The Berlin-based company Glycotope is moving in a different direction. It is working on platforms for the development and high-yield production of therapeutic proteins with a slightly different glycosylation profile. And this can only be achieved when the metabolism of glycosylation is understood in detail.
We still lack in-depth knowledge about glycosylation. The challenge is to make cells produce specific recombinant proteins with a specific predefined quality. All other problems, for example those related to up-scaling, metabolic side products and productivity, have been solved. What still needs to be clarified and understood in greater detail is how molecules are modified on the post-translational level. If we knew in detail how post-translational modification worked, then we would be able to specifically alter the properties of a molecule produced. Although some issues have been solved, we are still far from understanding how the glycosylation process works in detail.
I believe it can, as long as we are patient and have enough money at our disposal.
Such cell lines are already available. They are human cell lines derived from human retinal tissue, for example. The reason for using human cell lines is quite simple: why not produce a glycosylated therapeutic protein with human rather than with animal cells, which has the same quality as human proteins?The interview was conducted by Walter Pytlik, BioRegionUlm.