The Science

At the heart of modern biotechnology is the ability to manipulate DNA, the astonishing molecule that contains the genetic code of all life on earth. But fermentation and cell culture are also the practical art of keeping cells alive and growing in an artificial environment. As a result, the science of cell culture also includes a healthy focus on the natural processes that keep cells alive.

Recombinant DNA

Though human beings have long understood that certain characteristics could be passed from parent to child, until a half century ago they could only guess at how that worked. But then in February 1953, two young Cambridge scientists named James Watson and Francis Crick came up with the key when they worked out the structure of the DNA molecule.

Scientists already knew that DNA was composed of four "base" molecules: adenine, thymine, guanine, and cytosine (A, T, G, and C). What Watson and Crick determined is that the DNA molecule consists of two long strands of bases twisted around each other like a corkscrew staircase. More important, if you looked at the bases as they linked up across the "rungs" of the ladder, A was always linked to T and G to C and vice versa. The two strands were complementary: Pull them apart, and each could recreate the original double strand by adding the missing bases.

Watson and Crick's model showed how DNA could transmit information: information that was written in the four-letter language of G, A, T, and C. A crucial discovery came later—roughly 10% of the DNA molecules were genes that contained specifications for the production of proteins. These proteins are written in three-base "codons," each of which stands for a particular amino acid.

Throughout the 1960s, researchers studied DNA partly by breaking it into manageable bits using restriction enzymes. These enzymes recognize and bind to a specific pattern of base pairs and break the DNA molecule precisely at that point. Sometimes these enzymes leave blunt ends (as EcoRV does) and sometimes "sticky" ends (as EcoRI and TaqI do). Other enzymes called ligases bind those frayed ends together.

In 1972, John Morrow, Herb Boyer, and Stanley Cohen used these techniques to conduct an ingenious experiment. Using restriction enzymes and ligase, they pasted a bit of DNA from an African frog into a plasmid (a ring-shaped DNA molecule found in bacteria). When the plasmid was taken up by an E. coli bacterium, it became part of the organism's DNA: It was passed on from generation to generation — and it continued to produce frog RNA.

Morrow, Boyer, and Cohen used the plasmid, in the language of biotechnology, as a cloning vector. The job of the cloning vector was to carry foreign DNA into the E. coli without disturbing the organism's ability to reproduce. Plasmids were the first successful cloning vectors, but there were limits to how much DNA could be pasted into plasmids. A plasmid can typically accommodate DNA that is about 10,000 base pairs long. A cosmid, which consists of DNA taken from a virus that naturally lives in bacteria, can accommodate about 45,000. New vectors have been developed that will hold much larger fragments of DNA: A yeast artificial chromosome (YAC) can hold 100,000 base pairs. Mammalian and other artificial (or synthetic) chromosomes have been developed for even larger genes.

The African frog experiment contained the roots of the biotechnology revolution: It showed that altered or recombinant DNA could be passed on by a host organism, and that genes from one species could continue to function even when spliced into the DNA of a totally different species. From there it wasn't too great a leap to the insight that fast-growing bacteria or yeast loaded with recombinant DNA could be used to economically produce proteins and other substances in large quantities.

For example, human growth hormone was once available only from the pituitary glands of cadavers. It was very rare, expensive, and potentially contaminated with prions. In the 1980s, it even transmitted a rare neurological disease (Creutzfeld-Jakob disease) to patients before anyone was sure what kind of agent caused the disease. Now recombinant bacteria produce human growth hormone in large quantities that have made it safer and less expensive, and thus available to more patients around the world.

When those bacteria were genetically modified, the process had only just begun. They had to be cultivated in large enough numbers to produce therapeutic quantities of the protein. And that meant that it was essential to understand how they grew and reproduced.

Cell Metabolism

In broad outline, cell metabolism is basically the same in all living things. Oxygen and nutrients are brought inside the cell, where chemical reactions driven by proteins provide the cell with energy and raw materials. Waste products including water, carbon dioxide, and ammonia are expelled. These chemical facts of life point to three major concerns in fermentation and cell culture: nutrients, aeration, and the removal of waste products and heat. There are, however, some significant differences in the details of how different sorts of cells carry out the process of metabolism.

Bacteria, for instance, are prokaryotes, some of the oldest life forms on Earth. They are simple creatures, basically tiny capsules of watery cytoplasm in which float DNA and RNA, metabolic enzymes, food, and waste products. Prokaryote metabolism is relatively simple: They need simple nutrients (like sugar) to burn for energy.

Yeasts and plant and animal cells are a more complex form of life called eukaryotes. In eukaryotes, a nucleus protects the genetic material, which is stored on chromosomes. Ribosomes outside of the nucleus use RNA to translate that information for making proteins. Many other inclusion bodies or organelles are found inside the eukaryotic cell: microtubules, vacuoles, mitochondria, lysosomes, smooth and rough endoplasmic reticuli, ribosomes, and Golgi bodies.

In eukaryotic cells, mitochondria perform much of the metabolic work. They act just like little prokaryotes, employed by the big cell to do what they're best at: getting energy out of food.

Large nutrient molecules (such as proteins, polysaccharides, and fats) are broken down into smaller, usable pieces (amino acids, sugars, fatty acid, and glycerol). Those are broken down even further to make, among other things, two very important molecules: adenosine triphosphate (ATP) and nicotinamide adenine dinucleotide (NAD). Remaining nutrients are taken up by the mitochondria, which oxidize them to make more ATP and NAD.

ATP stores energy for all living things, and NAD is important in its production. ATP provides the power needed for most cell processes: molecular synthesis; transport of materials across the membrane; and cell replication, movement, and maintenance.

Cells for Bioproduction

Today's biotech and biopharmaceutical companies use a variety of bacterial, yeast, animal, and plant cells for production. The choice of which cell to use depends on a number of factors, some technical and some economic.

Bacteria and yeast, for instance, are relatively simple to grow. Each cell of a bacterium or yeast is an independent organism capable of its own metabolism. Yeasts and bacteria have fairly simple nutritional needs and grow well suspended in a liquid medium, even in big fermentors with a capacity of 1,000, 10,000, or more liters. The cell walls of bacteria and yeasts resist damage, even when the cells are packed closely together and stirred around by mechanical mixers.

Animal cells are much more fragile. They are often much larger than any microorganism and evolved to live in a collective as part of organs or tissues within complex anatomical systems. They are held together only by a delicate membrane, they are more difficult to grow in suspension, and they often have to grow attached to surfaces. These cells are complex, with systems of cellular machinery inside of them. Animal cells replicate slowly, require complex nutrients, and do not grow as well at high densities because of waste accumulation and oxygen stress. Animal cell culture is more complicated and thus more expensive than traditional fermentation.

TABLE 1: Approximate times from introduced gene to protein production at usable levels.

But ease of growing cells is not the only issue. Yeasts and bacteria may be easy to grow, but there are limits on the size of genes that can be planted in them. If it is necessary to clone a larger chunk of DNA, then it may be necessary to use animal or plant cells.

In addition, though various kinds of cells can all be made to express the same protein, a protein produced by a bacterium may have different effects than the same protein produced by animal cells. Why? After a protein is expressed in the cell, it goes through a process called posttranslational modification. Molecules of sugars and carbohydrates attach themselves to the protein (a process called glycosylation). The protein may fold itself into a different configuration, changing the surface available to attach to other molecules in the body. Folding and glycosylation have a great effect on the ability of a protein to be used for a particular process, and different types of cells perform these modifications in different ways. The choice of the right cell to culture may result in a protein that is more appropriate for use — or in the elimination of extra steps in the manufacturing process.

Some cells express a protein of interest and then keep it within their cellular membranes. Others secrete the protein, transporting it across those membranes into the liquid the cell is growing in. For biotechnology, secreted proteins are preferred because they are easier and more cost-effective to collect and purify.

What follows is a brief introduction to the types of cells currently used in biopharmaceutical production. A more detailed discussion of the pluses and minuses of particular cells and processes can be found.

But first, a note on terminology: In discussions of biotechnology, you will hear the process of growing cells referred to as both fermentation and cell culture. The terms are close in meaning, but the biopharmaceutical industry tends to distinguish between them, using fermentation to refer to the cultivation of single-celled organisms such as bacteria and yeast, and cell culture to describe a specific kind of fermentation used to grow cells that come from multicelluar organisms such as animals and plants.

Bacteria. Many biopharmaceuticals are produced by bacteria, especially the species Escherichia coli and Bacillus subtilus.

The DNA in bacteria occurs on a single circular chromosome and in small ring-like plasmids. These circular shapes keep the DNA safe from fraying at the ends, and they can be taken up and used by many kinds of cells, which makes them convenient for use as cloning vectors.

E. coli has been studied in microbiology laboratories for many years and was the first organism to have its entire genome mapped. It is common — certain strains live in our lower intestine all the time without causing a problem. It is cheap to cultivate, it replicates quickly, and it serves as a good model organism — that is, it provides an example of how other similar life forms will behave — how they grow and reproduce, what makes them deteriorate or die, and so on. When molecular biologists needed an organism to help them study genetics, they naturally turned to E. coli. Once safe strains had been engineered — most of which cannot even survive outside the optimal conditions of the lab — it only made sense to use them for further, more involved, experiments. When some researchers left academia to become biotech entrepreneurs, they took their knowledge of E. coli with them, and the bacterium became the workhorse of the biotechnology industry.

Yeasts. When yeasts were considered as a means of producing biopharmaceuticals, Saccharomyces cerevisiae was naturally the first candidate. Also known as brewer's yeast, it has been studied, characterized, and cultured over thousands of years. It and its similarly employed cousin Schizosaccharomyces pombe are our best understood species of yeast. The full genome of Saccharomyces cerevisiae was mapped and sequenced by molecular biologists in 1996.

Another species of yeast used in biotechnology is Pichia pastoris, which offers an interesting advantage for production. Some organisms tend to hold the protein they have produced inside their cell walls, which can make it more difficult to recover the protein. P. pastoris is one of the best of the yeasts at secreting protein into the liquid it grows in, which makes the product easier to purify. P. pastoris also is capable of posttranslational modifications that resemble those of human beings.

Animal cells. If historical precedent was important in the choice of E. coli and S. cerevisiae as tools for biotechnology, the same could be said of the choice of the most widely used animal cell: Chinese hamster ovaries (CHO).

Certain kinds of cells, particularly epithelial cells, are more robust than others and thus are easier to grow in culture. CHO cells are epithelial cells that were introduced to science in the 1950s. They multiply quickly, are relatively hardy, and grow well in culture.

In the 1960s, cancer researchers in Seattle discovered an interesting mutation in a particular line of CHO cells they were studying. It enabled the cells to grow in the presence of methotrexate, a chemical that kills cancer cells. That made them useful for later genetic engineering: When scientists add genes to a batch of cells, not every cell is modified. The trick is to separate cells that have taken on the new DNA from those that haven't. By adding the gene for methotrexate resistance to the new DNA package, scientists had a simple tool for distinguishing: If methotrexate was added to the culture, cells without the resistance gene would die off, and only the recombinant cells would survive. (A similar process is used in working with bacteria, but using a gene for resistance to antibiotics.)

By the 1980s, many molecular biologists were using CHO cells in their work, studying viruses among other things. So when those people went into the biotechnology industry, they took with them the knowledge of that particular cell line.

CHO cells are not the only cell line used for production of recombinant proteins by mammalian cell culture. Other commonly employed epithelial cell lines include human cervix (HeLa), African green monkey kidney (COS and CV-1), and baby hamster kidney (BHK) cells. In the realm of gene therapy, Per.C6, a cell engineered from certain cells from the human eye, is common. The abnormal cells created in cancers have also proven useful in biotechnology; hybridomas — "immortalized" cell lines from cancerous lymphocytes—are frequently used in the production of antibodies.

Every protein produced through fermentation or cell culture needs to be purified before use. For mammalian cell culture, an important part of that purification will be the removal of parasitic viruses and other agents that could infect users of the drug. The problem of viral contamination of animal cells dates to long before the birth of biotech. When the polio vaccine was first produced in the 1950s, many latent viruses were present in the monkey cell cultures used. Some of those (even some lethal viruses) were transmitted to humans. In the 1940s, millions of doses of live yellow fever vaccine produced in eggs were contaminated by endogenous chicken retroviruses and hepatitis B. In the 1980s, human-sourced products like growth hormone and blood products were contaminated with hepatitis, HIV, and other agents.

Many pathogens that infect one kind of mammal will happily reproduce inside another. So cells intended for cell culture must undergo the process of viral characterization — that is, testing for any possible adventitious agents that may be present with the cells. If there are too many, or the viruses are too dangerous, a cell line may be scrapped and other cells selected.

Contamination by an adventitious or endogenous agent does not necessarily render cell lines useless. Virus screening was instituted to avoid such problems. Safety decisions are based on detailed risk-benefit analyses that have given mathematicians (particularly statisticians) a whole new career choice.

Insect cells. When insect cells are cultured to produce proteins, the process is very different from that for mammalian cells. Insect cells are eukaryotic, too, and capable of doing many of the same complex posttranslational modifications to proteins that mammalian cells do. But the cells themselves are not genetically modified directly. Instead, a system called baculovirus expression vector system (BEVS) is used. Viruses are even simpler forms of life than bacteria. They cannot replicate by themselves but require a host. A virus doesn't even have a cellular structure, existing only as a particle made up of a protein shell that protects a bit of DNA or RNA inside. That genetic material is injected into the insect host cell, where it takes over the ribosomes to build more viruses.

Baculovirus is a particular type of virus that replicates only in the cells of lepidopteran insects, an order that includes butterflies and moths. It is harmless to all other creatures. Normally, when the virus infects an insect cell, it injects the cell with DNA instructions that cause it to produce large amounts of a protein that coats and protects the virus. When baculovirus is used in biotechnology, molecular biologists replace the gene coding for the protective protein with a gene coding for a desirable protein. The baculovirus infects the insect cells, and the cells express the protein.

Just because the baculoviral vector does not infect mammals doesn't mean that insect cell lines are completely free of potentially dangerous viruses. They must be characterized just like CHO and other mammalian cells because other adventitious agents may be present even if they don't harm the cells. Insect cell expression is still a new technology. Mammalian cells (particularly CHO cells) are still the most commonly used nonbacterial system.

Cell lines and cell banks. Biopharmaceutical companies do not necessarily obtain animal cells from their original sources — hamsters, monkeys, insects, or humans. Instead, they work with vendors that maintain specific strains of the cell lines. In the United States the American Type Culture Collection (ATCC) maintains more than 4,000 different mammalian, insect, and other cell lines and hybridomas. ATCC's European equivalent is the European Collection of Cell Cultures (ECACC).

Certified cell lines come with comprehensive data files that provide their full history (species of origin, strain, and colony). Tissue donors are tested for viruses such as hepatitis, papillomavirus, and the human immunodeficiency virus. Immunization regimes are also recorded, and gene fusion, DNA cloning, and cell selection procedures are detailed. Cell lines are characterized by tests that ensure freedom from microbial contamination and that confirm the cells' specific function and identity. DNA "fingerprinting" is often used to verify identity.

Once a biopharmaceutical company obtains a beginning cell bank from a cell culture collection, its scientists will do the work of genetically engineering those cells for the company's particular use. They may not only add the gene of interest (often "amplified" or repeated copies of that gene so that each cell produces more protein), but they may also make other changes. For example, an animal cell line that naturally prefers to grow attached to a surface can be adapted to grow suspended in liquid like bacteria or yeasts. Cells can be modified so that they perform posttranslational modification of the protein in more desirable ways.

When experimentation has produced an optimal version of the cell line, a master cell bank (MCB) is created. It will be maintained as the source of all cells used to produce the company's drug through preclinical and clinical testing and then into commercial sale. Working cell banks (WCBs) are created from the master cell bank for producing batches of product. Each batch will be made by seed stock that came from either another working cell bank or the master cell bank.

Cell lines and cell banking. Many factors are considered in choosing and maintaining a cell line, not just the type of protein being produced. The kinetics of protein production and yield, product stability during the manufacturing processes, and the sensitivity of cells to shear damage are also important. A cell line must produce the desired protein in amounts that warrant the expense and trouble of the manufacturing procedures involved.

Extensive records have to be kept on a master cell bank. The company's quality assurance department must document the origin of the DNA sequence that codes for a protein of interest, including the source from which it was obtained; the method used to prepare the DNA expression construct for genetic engineering; a detailed component map showing the number of copies of the gene of interest, all insertions and deletions, the sites on the DNA molecule where they took place, and the complete annotated sequence of the expression vector used; methods for transferring the expression construct into the host cell line and amplifying it for higher expression levels; and criteria for selecting cell clones used to create a master cell bank.

The company keeps records of all the reagents and media used with the cells, specifying storage conditions, determining the age of cells, and validating all methods and procedures involved. Data derived from the beginning cell line is used to determine the in vitro life span of WCB cells used in production, with additions and adjustments as development progresses through pilot and commercial scales.

Cell banks can be stored frozen (cryo-preserved) or, in the case of microbes, freeze-dried (lyophilized). In the industry, most microbial (bacterial and yeast) cell banks are stored frozen in liquid suspensions. Chemical additives (cryopreservatives) protect cells from ice crystal damage but can present a contamination risk. Most bacteria species used in biotechnology remain viable and productive if lyophilized or kept below –20°C. The oldest method of maintaining cell lines is through continuous subculture, in which a population of bacteria or yeast is kept growing in culture. That method is seldom used in biotechnology because of the danger of mutations.

Mutations are inevitable in a living, reproducing system. Perhaps one in a thousand genes is mutated in any given organism. These facts of life help evolution work, but they could wreak havoc in a fermentation process. Cell banking helps biotechnology companies avoid the problem of mutations occurring in their cell lines. There simply isn't enough time between thawing the MCB and running a fermentation batch to allow mutations to occur. But recombinant DNA is less stable than native DNA. Recombinant components can be lost in less time than it would take for a cell line to develop mutations.

On the Cutting Edge

Although fermentation and cell culture are the dominant expression systems in biotechnology, they are not the only methods of producing pharmaceutical proteins. Some people consider transgenics to be the next step in biotechnology. Transgenic mammals produce pharmaceutical proteins in their milk, prompting some devotees of that technology to refer to the mammary gland as "nature's perfect bioreactor." Transgenic hens lay the modern equivalent of the fabled goose's golden eggs, with valuable recombinant proteins inside. Transgenic plants are creating a need for a new kind of farming.

For companies that would rather stick to fermentation than switch to farming, new ideas in fermentation offer exciting possibilities. A Swiss company has developed a way to produce recombinant proteins in a slime mold, Dictyostelium discoideum. A primitive eukaryote that feeds on bacteria, it offers some of the advantages of yeast and some of the advantages of higher animal cells. Like yeast, it grows quickly in simple fermentation equipment and can be stored easily, but its cells do not produce endotoxins or possess cell walls. Like other eukaryotic cells, it can perform complex posttranslational modifications of proteins.

What does the future hold for protein expression systems? If the past is any indication, biopharmaceutical companies can expect to have many choices at hand, each with its own advantages and drawbacks. Bacteria, yeast, and mammalian cell culture may be the dominant choices now, but as therapeutic proteins become more complex and are needed in larger and larger quantities, it's evident that companies will need all these options in the future.