IVD Technology Magazine | IVDT Article Index

Originally published May, 1997

Pathogen-free human serum protein production using a hollow-fiber bioreactor system

Artificial production may eliminate the shortages, variability, and possible contamination that complicate use of human serum proteins.

James H. Kelly, Andrea L. Spiering, and Norman L. Sussman

Used in a wide variety of applications, human serum proteins are critical raw materials for the diagnostics industry.¹ Measurements of specific members of this group of proteins comprise some of the most common clinical chemistry tests. An obvious example is the test for alphafetoprotein (AFP), used universally in prenatal diagnosis. Albumin is an essential component of the serum-free cell culture media used to grow hybridomas for monoclonal antibody production and is also used as a stabilizer in their formulation.² Other human serum proteins with applications in IVDs include transferrin (used in hybridoma culture), the blood coagulation factors (measured to detect bleeding disorders), proteases, antiproteases, and other serum factors.

Traditionally, these proteins have been purified from pooled human plasma for use as antigens or manufacturing components. This source, however, is coming under heavier scrutiny by FDA because of the possibility of contamination by viruses and other infectious agents. Testing can lessen or eliminate the risk from known infectious agents, such as hepatitis B, but cannot detect the new agents that are continually entering the blood supply. Newly discovered strains of human immunodeficiency virus and the agent or agents responsible for the transmissible spongiform encephalopathies, such as Creutzfeldt-Jacob disease and mad cow disease, exemplify this problem.³

A solution to it is to divorce production of human serum proteins and other commercially important proteins from dependence on the blood supply. We have developed methods using a human liver cell line cultured in hollow-fiber bioreactors to produce human serum proteins in a controlled, closed system. There the risks of contamination by infectious agents can be virtually eliminated. This method can serve as a model for the manufacture of many other reagents, such as viral antigens, which expose the operator and user to unknown risks.

Protein Production in Hollow-Fiber Bioreactors

The major impediment to the use of cell culture systems for the production of bulk proteins is the relatively low efficiency of most systems. This inefficiency can be quantified in terms of their synthetic rate, the amount of protein produced per cell per unit time, and their synthetic efficiency, the volume of cell culture medium consumed per unit time.

Monoclonal antibodies, for instance, are generally produced in only 100- to 500-mg-per-day quantities, even in large bioreactors. This process consumes 7­10 L of cell culture medium per day. Given an average cost of $5 per liter of this medium, costs quickly expand beyond $1000 per gram when other materials are included.

Since serum proteins such as albumin and transferrin are used in very large quantities relative to monoclonal antibodies, protein production per cell must be maximized if it is to be cost-effective. Moreover, the proteins must be obtained for a cost that is competitive with purification from pooled plasma.

Hollow-Fiber Devices

Diagram of a hollow-fiber cartridge device showing the entire cartridge (A) and individual hollow fibers (B).

To grow the mass of cells needed for protein production, we have cultured C3A, a human liver cell line, in hollow-fiber culture devices.⁴ The cartridges are two-chamber units; the cells (and their secreted proteins) are separated from the large volume of circulating medium. Medium is circulated through the intracapillary space, and cells are grown in the extracapillary space (ECS). This system is most commonly used in the manufacture of monoclonal antibodies. It allows the production of highly concentrated secreted protein solutions in the ECS.

In this particular instance, we have used cellulose acetate hollow-fiber devices with approximately 2 m² of surface area for cell growth. Each device is composed of approximately 10,000 cellulose acetate fibers. In practice, growth medium is circulated through the intracapillary space at a high flow rate, delivering oxygen to the cells. Medium is pumped through the ECS space at selected intervals to collect the secreted product. While we have principally used a custom-built production system, large-scale systems--such as the Cellex AccuSyst series of instruments--are now commercially available.

A Human Hepatoblastoma Cell Line

C3A is a patented human liver cell line derived from a hepatoblastoma, a tumor of early childhood.⁵ C3A forms the basis of a liver-assist device that has been used to treat 23 patients with acute liver failure.^6,7 These cells have been shown to maintain a wide array of metabolic function characteristic of the human liver, including the following:

• Ammonia metabolism to urea and glutamine.
• P450-based drug metabolism.
• Aromatic amino acid metabolism.
• Serum protein synthesis.

In each instance, the cells have been shown to perform about as well as human primary hepatocytes on a per-gram basis.

Highly Characterized Cells. C3A cells have been subjected to the rigorous testing required of cell lines to be used in human subjects. With the exception of tumor formation in athymic mice, all results have been negative. The cells are stored in a qualified master cell bank (MCB). Production runs are prepared from the manufacturer's working cell banks (MWCBs).

Cross section of a mature bioreactor. C3A cells proliferate to fill the entire extracapillary space of the device. The circles are remaining hollow-fiber sections; in other areas, the fiber itself has been lost during fixation and sectioning.

High Levels of Metabolic Activity. Performance of the C3A bioreactor system is extremely high relative to that of most high-density cell cultures. Composed of liver cells, C3A is highly aerobic and metabolizes substrates efficiently. It has also been selected for strong contact inhibition. Cultures never overgrow their feed source and can be maintained indefinitely.

The environment in a hollow-fiber bioreactor is substantially different from that of a standard T flask. The nutrient environment is much richer, with better oxygen delivery and constant input of fresh cell culture medium. Cells are at a very high density and form three-dimensional structures, similar to those of tissues.

Wide Variety of Proteins Synthesized

The synthesis of serum proteins is a hallmark of the liver-specific phenotype, and C3A synthesizes a wide variety of these proteins. The cells synthesize an extraordinary amount of protein per day. Up to 4 g of human albumin can be obtained per day in a 2-m² bioreactor. Each of the proteins are synthesized in their appropriate ratios. C3A cells' only major deviation from normal adult hepatocyte synthesis is production of significant amounts of AFP, indicative of C3A's hepatoblastoma origin.

Consumption of Cell Culture Medium

Hollow-fiber systems are generally continuous, rather than batch. Fresh medium is continually pumped into the system, and an equivalent amount of waste medium is expelled. This exchange results in establishment of a steady-state system, in which nutrient and waste concentrations are balanced. This balance is in marked contrast to the inevitable buildup of waste products in other culture systems, which is usually the factor limiting production and lifetime.

The production of lactate and ammonia exemplify this contrast. Large hollow-fiber devices can easily consume 15 L of medium per day. This volume is necessary to supply the synthetic machinery with raw materials: glucose and amino acids. Of course, the majority of the components of cell culture medium are not consumed. Ions and vitamins are mostly recycled. Once cell growth has plateaued, cells should be able to survive with minimal medium, so long as oxygen, pH, glucose, and amino acid levels are maintained. Replacing only the consumed nutrients with a concentrated stock cuts medium consumption dramatically.

Supplementation with concentrated solutions has been suggested in other systems but is generally impractical because of the metabolic characteristics of the cells. For instance, if 80% of glucose is metabolized only as far as lactate, pH begins to drop. Base is added to maintain pH, and more glucose is added as well. New glucose produces more lactate, which then requires more base. Soon ionic strength problems arise. Addition of water to dilute the ionic strength lowers ion concentrations. In most real-world situations, replenishment with whole, fresh medium is the most expedient course.

C3A cells present a different situation. A study was conducted to track glucose use and lactate production across the growth curve of a single set of devices. Each device was inoculated with approximately 2 g of C3A cells. Over the course of about three weeks, the cells would multiply until, at maturity, each device contained about 150 to 200 g of cells.

Ammonia, another frequent problem in large-scale cultures, is metabolized to urea or glutamine by C3A cells. There is no ammonia buildup in the system. Ammonia toxicity is generally seen in the 2­5-mM range. Levels in the C3A system never approach these concentrations.

Economic Analysis

As mentioned earlier, serum proteins are currently produced from pooled human plasma, and any cell-based production system must be cost-competitive. At first glance, this challenge is daunting, since human serum albumin retails for between $5 and $10 per gram. Examination of the cost of other serum proteins is more encouraging; transferrin, for instance, retails for about $300 per gram. Since the cost of producing monoclonal antibodies in hollow-fiber systems is between $500 and $800 per gram, cell culture would not seem to be an economical source for serum proteins.

Two factors combine to make this a false impression. First, the human liver cell bioreactor system is extremely efficient, so the cost of production relative to that of monoclonals is dramatically reduced. Second, no one protein must bear the entire production cost. All of the serum proteins are obtained from the same extracapillary fluid. Hence, the production costs are amortized over as many different proteins as can be reasonably isolated from a single batch. Albumin, transferrin, fibrinogen, and AFP are all easily separated from the initial fluid using standard column chromatography.

Advantages

A number of competitive advantages derive from the production of serum proteins using cloned cells in a hollow-fiber system. The major advantage is that proteins manufactured in this way can be guaranteed specific-pathogen-free. The cell lines can be subjected to a battery of tests to ensure that they are completely free of any known contaminants. Master cell banks and manufacturers' working cell banks can then be prepared from these tested and documented cells. All production is carried out from these banks, eliminating concern about hepatitis viruses, prions, and other pathogens.

Protein production in these systems is not dependent on human plasma collection and so is not subject to shortages and recalls. The supply of human serum albumin has always been limited by the availability of plasma.

Only liver-derived serum proteins are present in the harvest from the hollow-fiber system. Plasma contains up to 25% nonliver proteins, which complicates purification. Since the starting material in a cell system is less complex than plasma, higher purity of the final product can be obtained with fewer purification steps.

Finally, production of serum proteins in this system is cost-competitive. The resulting products are of equal or better quality and lower risk than those obtained from plasma.

This system can serve as an example of the use of hollow-fiber systems for both large- and small-scale production. Adaptation of cells, both suspension- and anchorage-dependent, to hollow-fiber culture is relatively straightforward. Experienced contract manufacturers, such as Cellex (Minneapolis) and Unisyn Technologies (Boston), can facilitate cell adaptation.

The use of hollow-fiber bioreactors is frequently more economical and almost always less labor intensive than the use of large numbers of roller bottles and the collection of large volumes of supernatant fluid. As these systems become more widely available and used, they should be seen as a straightforward alternative to roller bottle culture or, in cases such as described here, to potentially dangerous human plasma for protein isolation.

Return to the IVDT home page

References

1. Bowman BH, Hepatic Plasma Proteins, New York, Academic Press, 1993.
2. Peters T, Jr., All About Albumin: Biochemistry, Genetics and Medical Applications, New York, Academic Press, 1996.
3. Collinge J, Sidle KCL, Heads J, et al., "Molecular Analysis of Prion Strain Variation and the Etiology of 'New Variant' CJD," Nature, 383:685­690, 1996.
4. Oka MS, and Rupp RG, "Large-Scale Animal Cell Culture: A Biological Perspective," in Lubiniecki AS (ed), Large-Scale Mammalian Cell Culture Technology, New York, Marcel-Dekker, 71­92, 1990.
5. Kelly JH, "Permanent Human Hepatocyte Cell Line and Its Use in a Liver Assist Device (LAD)," U.S. Patent No. 5,290,684, 1994.
6. Sussman NL, Gislason GT, Conlin CA, et al., "The Hepatix Extracorporeal Liver Assist Device: Initial Clinical Experience, Artif Organs, 18:390­396, 1994.
7. Sussman NL, and Kelly JH, "The Artificial Liver," Sci Am: Sci Med, 2:68­77, 1995.

---

James H. Kelly, PhD, is president, Andrea L. Spiering is laboratory manager, and Norman L. Sussman is executive vice president, Amphioxus Cell Technologies, Inc. (Houston).

ÇYÐÇY