The biotech revolution has led to increasing numbers of biological medicines reaching the market. These are usually complex, large molecules, which cannot be made easily by traditional chemical synthesis routes, so biological processes have to be used to make them instead – a very different proposition to standard small molecule manufacture.

The earliest biologics, such as the growth hormone EPO, were only needed in small volumes. These days, however, biologic medicines to treat common chronic diseases like arthritis are coming on stream, not to mention products that need to be administered in larger doses, so the ability to manufacture at bigger scales is becoming essential.

Many biologics are made using mammalian cell culture technology, which uses recombinant DNA technology to introduced relevant genes into cells. These cells are then used to make the desired protein, which is ultimately fished out of the complex soup at the end of and incubation period lasting weeks, or even months.

The first stage of the process, developing the cell line, involves the gene which codes for the desired therapeutic protein being inserted into the genome of the host’s cell. The recombinant cells that perform best are selected, and they can then be preserved by freezing them in ampoules. These can then be stored in a cell bank under liquid nitrogen at –180°C until they are needed.

The ‘business end’ of the process takes place in the bioreactors, which can range in size from small glass bottles to huge reactors with a capacity of thousands of litres. First, an ampoule of cells is taken from the cell bank, and its contents thawed. The cells are then multiplied in a suitable culture medium. Once there are sufficient cells, they are put in the bioreactor, and fed with a nutritious culture medium containing substances such as salts, sugar and vitamins, plus air to provide oxygen.

In the reactor, a number of different proteins are made by the cells, including the desired therapeutic protein. The state of the cell culture and the quality of the product are checked regularly. The mix containing the protein is then harvested and separated from the cells for purification. This involves several consecutive steps, including filtration and chromatography.

Purification starts with filtration. Typically, ultrafiltration is used to concentrate what comes out of the reactor to around a tenth of its original volume. Then chromatography is used to separate the different molecules. Various different forms of chemistry are used, including size exclusion, which acts a little like a sieve, allowing proteins to pass through a resin at various speeds according to their size, with the smallest molecules coming out first.

There are essentially two reactor technologies that are used in mammalian cell culture: the older perfusion technique, and the fed-batch technique that has been developed more recently.

In a perfusion bioreactor, the incubation process takes up to two or three months. The culture medium is fed into the bioreactor by perfusion technology over the months. In the newer fed-batch technique, the medium is fed to the bioreactor over a shorter period, typically three weeks, and the complete harvest is taken in one batch, with the entire reactor contents being separated using a centrifuge before purification. Perfusion reactors are up to around 300 litres in size, whereas fed-batch bioreactors are better adapted to larger scale production requirements.

One company that has been making significant investments in mammalian cell culture manufacturing is Serono. The Swiss based firm is the world’s third biggest – and Europe’s largest – biotech company. Its current major product, multiple sclerosis treatment Rebif (betaseron), reached blockbuster status last year, with sales topping a billion dollars. Rebif has been on the market since March 2002, when it gained approval for the treatment of relapsing forms of MS, having overcome a competitor’s orphan drug status based on superior efficacy in relapse reduction.

Serono has 10 manufacturing sites, including the Serono Biotech Centre at Corsier-sur-Vevey, near Lake Geneva in Switzerland. It is making substantial investments at the Vevey site, where it now has 5,000 litre fed-batch reactors, and has plans to install a dozen massive fed-batch reactors with a capacity of 15,000 litres each.

The expansion has become necessary because of the increasing move away from hormone type products, which are used in only very small doses, towards newer therapeutic proteins and monoclonal antibodies which are given in larger doses, and hence require larger manufacturing capacity to meet demand.

When the facility, on the site of an old tobacco factory, was first developed, it was designed to have plenty of space for expansion. The company began making Rebif there back in 2000, and it is now implementing a dual sourcing back up strategy. The aim is that all of its products will be manufactured at more than one site to ensure continuity of supply should the unthinkable happen and one of the manufacturing sites has a problem of some sort.

‘We have dozens of bioreactors at the SBC,’ says the company’s senior vice president, manufacturing, Michele Antonelli. ‘These range in size from 50 to 5000 litres. Some processes are still in roller bottles, where the process works well. We are looking to expand our bioreactor capacity to a total of 200,000 litres.’

‘When we opened the Serono Biotech Centre, only one third was developed, which means we are able to respond rapidly to market needs,’ claims Jonathan Barnsley, site director of the Vevey plant. Even after the recent capacity additions, a third of the building remains available for the introduction of the planned 15,000 litre reactors.

‘The molecules we are working with now have to be administered in larger quantities so we are making them on a larger scale,’ adds Antonelli. ‘Processes only work if they can be scaled up to a large scale. We need to minimise immunogenic reactions, so product purity is very important when you are developing new processes. We also have to work on production costs at an early stage of the process because has to be bearable and the products acceptable to patients.’

The SBC also currently has four 300 litre perfusor bioreactors. Rebif is made using a perfusor; some products can only be made using these types of reactors, especially if the products are very unstable, for example if they begin to degrade when they warm up. Rebif is one of these, and is made in 75 litre reactors. At any one time, the company is normally running between six and eight of these reactors.

Mammalian cell culture makes a lot of very dilute product; even the very best protein manufacturing processes make as little as 7 grams a litre, while most make much less, maybe only one or two grams. But yields are going up all the time. The practical aim has to be getting a few kilograms of product from a few thousand litres of reactor. Interferon is a substance made by cells in trouble, and Rebif is a very low concentration product – not even grams a litre – so large volumes are needed to manufacture sufficient for clinical needs.

The company carries out development work at the SBC as well as just production. This includes cell sciences and developing improved cell culture media, as well as investigating both upstream and downstream processes such as better purification methods. Purity is a big issue for therapeutic proteins, and the byproducts include incomplete protein chains, and components of the culture media. Another problem is that the product that is being made can interfere with the cells, slowing down the reaction and having a negative effect on the amount of yield the cells will make.

A further issue that slows down the process is the need for cleaning and validating the reactors between batches. For this reason, Serono is moving towards the use of disposables as much as possible. Throwing away a plastic bag liner is much quicker than having to worry about cleaning, and it makes the process more flexible, too.

Increasing production capacity is not just a case of installing more, and larger reactors. If you can manage to double the productivity of the cells, then twice as much product can be made without doubling the reactor size. ‘It’s more efficient to get the cells to improve the process for us!’ says Barnsley.

‘If you can reduce cycle times by, say, two days per run, this makes a big efficiency saving over the year,’ says Tim Clayton, manager, technical support at the SBC. ‘You need a good process to make the most of your cell line. But, of course, once the product is licensed, if you change the process enough it can change the product, and you may even need to do new clinical trials. Advantages can come from keeping the cells alive for longer.’

Ensuring the product is as pure as possible is essential, as is proving that there is very little variability from batch to batch. It is not as simple to ensure this as it is for traditional small molecule chemical manufacture, as so many other factors can influence the production process, such as the cell line, the media used and the temperature. So careful checking is essential to keep the process under control.

‘We sample at all the stages of the process,’ explains Barnsley. ‘The samples are analysed in our quality control labs to check their quality. And our quality assurance people check the quality of batches from production, to ensure they comply with the regulatory authorities’ specifications.’

Advances in genetics research mean that medicines like therapeutic proteins are only going to increase in importance in the future. And as more products reach the market that will be given to many millions of patients over protracted periods of time – treating chronic conditions ranging from arthritis to heart disease – it is going to become ever more important to have reliable, large scale manufacturing processes. Cell lines will be made more efficient, and culture media that encourage the cells to make more of the required proteins will be developed, as biotechnological manufacturing methods become as routine as small molecule chemical synthesis is today.

BOX – Identifying the genes involved in multiple sclerosis

Scientists at Serono have made a breakthrough in understanding the genetic basis of multiple sclerosis. It has been known for some time that, while susceptibility to multiple sclerosis runs in families, it is not an inherited condition.

Now, researchers at Serono have managed to identify and create a register of genes involved in multiple sclerosis. These 80 genes are involved in the inflammatory and neurodegenerative pathways of MS.

The chronic inflammatory condition of the nervous system is the most common non-traumatic neurological disease in young adults, and is thought to affect as many as two million people worldwide. Although symptoms vary, the most common include blurred vision, numbness or tingling in the limbs, and problems with strength and coordination. There are a number of different forms, but the relapsing ones are the most common.

The large scale association study was carried out in three populations, in France, Sweden and the US, totalling 900 people with MS and a similar number of healthy individuals. Scientists at the Serono Genetics Institute at Evry, outside Paris, scanned over 100,000 single nucleotide polymorphisms (SNPs) using Affymetrix GeneChip technology. These were compared between the patients with the disease and those without to pinpoint the genes involved in MS.

A total of 80 genes were found to be statistically significant in all three populations, including one of the seven genes previously thought to be involved in MS. ‘These are the ones we are rather sure are involved in MS,’ explains Daniel Cohen, Serono’s worldwide head of genetics.

The genes also provide potential new drug targets. ‘For our 79 new genes, at least half of them encode for proteins which can be modulated by therapeutics,’ Cohen adds. The genes are involved in several different pathways, notably inflammation and neurodegeneration, plus some novel pathways.

Thus far, a 40% scan of the genome has been carried out. The team now plan to apply the next generation of GeneChip technology to scan over half a million SNPs, with the aim of completing the whole genome scan at some point next year.

Tim Wells, Serono’s senior vice president, research, claims that one of the reasons so many biotech drugs fail is that they are too far abstracted from the disease process itself in humans. ‘We have a unique biotech approach that brings human genetics into all stages of the discovery process, increasing our success in the clinic,’ he says. ‘We are looking at real human disease.’