Medical Device Link.

Originally Published IVD Technology June 2003

ASSAY DEVELOPMENT

Improved manufacturing techniques enable bovine serum albumin to shed its low-tech image.

Michael O. Budnick and Randal S. Fitzgerald

Continuous production of "heat shock" bovine albumin takes place within a computer-monitored closed system of tanks at the Proliant Biologicals fractionation plant (Boone, IA). Each tank performs a different function and is designed to draw specific fractions of plasma and further process them into other plasma-derived products. Most existing BSA purification processes destroy everything but BSA.14

Few diagnostic assay components have received as much blame for variable product performance as the protein workhorse bovine serum albumin (BSA) (see Figure 1 and Table I). Background interference, loss of signal, conjugate instability, and component degradation are only some of the common problems that can induce troubleshooters to point to BSA as a scapegoat.

Not that this reputation is entirely undeserved. The performance variability that is attributed to BSA is in no small part a consequence of the way it was originally manufactured. The methodology and equipment used in BSA production were developed for a completely different industry. BSA manufacture was an offshoot of a human serum albumin (HSA) production process designed in the wartime 1940s for plasma volume expansion. The production goals for manufacture of that type of albumin were not congruent with the current needs of the diagnostic industry, which requires virtually 100% pure albumin with even trace amounts of potentially interfering analytes removed.

One of the major in vivo functions of BSA is to serve as a carrier protein. It carries well—in fact, it even carries numerous unwanted constituents all the way through its traditional manufacturing processes. BSA itself is a highly stable and resilient protein.^1–3 And minute quantities of interfering analytes are generally of little concern in therapeutic applications. But variability in diagnostic BSA attributable to these copurifying, bound, and generally confounding hitchhikers has been a persistent problem.

Fortunately, much has happened over the years to improve the performance and consistency of bovine albumin for diagnostic manufacturing. Development of new technology, an industry shake-up, a growth in understanding of albumin's role, and a sharpened focus on critical issues in albumin production have all led to greater BSA quality and choice for the manufacturers of diagnostic assays and components that depend on BSA. This article examines present and potential future best manufacturing methods in the light of BSA's production history.

Cohn Fractionation and Fraction V

At a circulating concentration of 35–50 g/L, albumin is the largest protein component of blood plasma.⁴ The need for a component-stable blood substitute that would save the lives of Second World War trauma patients motivated Harvard biochemist Edwin Cohn at that time to develop methods of breaking down plasma into its individual components (see sidebar). Cohn exploited the differential solubility of plasma proteins at various temperatures, pH ranges, and salt and organic solvent concentrations to identify five fractions of human plasma that yielded fibrinogen, immunoglobulins, transferrin, lipoproteins, and other components that would later have commercial significance. The fifth fraction produced by his process was albumin, which is why the protein is commonly referred to as Fraction V.^5,6 This name has been a universal synonym for albumin whether manufactured via the original five-step Cohn fractionation process or not.

Figure 1. The BSA amino acid sequence.14

The original Cohn methodology involved manipulating the concentration of ethanol at low temperature and varying pH to isolate albumin from plasma. In its most basic form, ethanol fractionation of albumin produces a crude preparation that is 80–90% pure. Manufacturing process improvements soon added dialysis, ion exchange, and various filtration techniques to the Cohn method to further purify the product. However, the very properties that make albumin such an important protein to the organism worked against the achievement of a high level of purification for in vitro applications.

The ligand-binding properties of the transport protein albumin include affinities for fatty acids, trace metals and ions, hormones, and other endogenous and exogenous substances in blood plasma (see Table II). Binding may exhibit very high affinity, but can also show a high rate of dissociation under the right conditions.^2,7–12 The affinity is typically hydrophobic, but may also be covalent or ionic in nature.² While the prevalence of bound ligands has little effect on the performance of albumin as a plasma-volume expander—indeed, bound fatty acids can increase its stability—they can cause difficulties in such higher-purity applications as immunoassays and clinical chemistry.^9,13

Human plasma for albumin manufacturing was in limited supply and expensive. Cohn's team originally thought that the fractionation process could be adapted for the economical and plentiful bovine plasma available from the beef industry. In vivo results achieved with BSA proved less than satisfactory, however. Bovine albumin was isolated, and it was effective in plasma volume expansion, but its use was potentially deadly.

Albumin's structure is highly conserved among various species of mammals, with the sequence homology between BSA and HSA being 76%. Nonetheless, it appeared that the differences between BSA and HSA were great enough to make the molecule dangerously allergenic in a small percentage of infused patients. Purification of BSA from bovine plasma consequently became a method in search of an application, a moribund project awaiting the emergence of modern diagnostics and cell culture.

A Role for BSA

With the arrival of a medical diagnostics industry based on clinical chemistry and immunoassay technology, bovine albumin became the economical component of choice for a variety of applications. The uses for BSA in diagnostics are as numerous as those in its living host. It can serve as a protein standard and a diluent; as a blocking and coating agent in immunoassays; as a buffer component to stabilize proteins, antibodies, and conjugates; and even as a cell-culture supplement for the production of monoclonal antibodies in hybridoma culture.

Figure 2. The heat-shock process isolates albumin by means of heating plasma in the presence of an excess of caprylic acid, which displaces native fatty acid species and is in turn removed by subsequent dialysis and filtration. (click to enlarge)

Beyond Cohn Fractionation. Various modifications of the Cohn procedure have been used to produce higher-purity BSA. Still, many manufacturing methods are rooted in Cohn fractionation—even employing ethanol or other organic solvents—and result in a product that can be as frustrating as it is indispensable. But this will change. Considerable research into albumin in recent decades has revealed new information about its stability, conformation, and ligand-binding characteristics.

Use of a heat-shock method, in which the plasma is heated in the presence of a caprylic acid stabilizer to denature nonalbumin proteins, has come to be accepted as a way to achieve a greater degree of purity (see Figure 2).^2,4,13 The typical residual fatty acid content of Cohn ethanol BSA is much higher than that of a standard heat-shock grade (see Table III). Manufacturers of Cohn ethanol products typically advertise that their products are rich in fatty acids and cofactors in order to appeal to customers preparing cell and microbial cultures. By contrast, many diagnostic detection methods involve the use of enzymatic components, analogs of which may be present in plasma and its fractionated proteins. Contaminants, even at very low levels, can activate chromagenic or other substrates, resulting in high background levels. Many of these enzymes need to be heat-inactivated, which can only be achieved through an aggressive heat-shock manufacturing procedure (see Figure 3).

Subsequent extensive diafiltration, charcoal filtration (for delipidation), and ion exchange also can help remove interfering analytes.^2,4,15

Figure 3. The heat stability characteristics of a BSA contaminant that activates the chromagenic tetramethylbenzidine (TMB) substrate used in horseradish peroxidase–based enzyme-linked immunosorbent assays (ELISAs), graphed as residual colorimetric enzyme activity in response to heat treatment. (click to enlarge)

An Adverse Event Opens the Door. Efficient BSA production methods developed in the past 5 to 10 years have finally effected significant improvement in the consistency and reliability of bovine albumin. The changes in methodology were instigated by industry upheaval as well as by basic research in albumin and fractionation chemistry. A useful shakeout among BSA manufacturers was precipitated, interestingly, by problems in the human blood fractionation industry.

Production of human blood plasma products came under increased scrutiny by FDA in the 1980s in the wake of incidents of HIV transmission by plasma and clotting factors. FDA continued to keep a close eye on the industry even after many production processes were validated for the removal of pathogenic viruses. By the mid-1990s the agency was zeroing in on human albumin fractionation to verify that good manufacturing practices (GMPs) were being strictly followed.

In August 1996, an adverse event involving patient infection was traced to therapeutic human albumin manufactured by Centeon (Kankakee, IL). Scrutiny of all human plasma fractionators was thereafter intensified. A series of FDA audits of Centeon resulted in a Form 483 report detailing a number of GMP violations by that manufacturer. Production was shut down, product recalls ensued, and manufacturing resumed only many months later under a consent decree.

Although little noticed outside the industry, Centeon was also an important manufacturer of BSA, producing the material for Intergen Co., formerly its biochemical division called Armour Biochemicals. Its BSA production was interrupted for a time during the plant shutdown. BSA manufacturing at this time was dominated by two companies, both offshoots of Armour Pharmaceuticals in Kankakee, an original collaborator with Cohn. Soon after the Centeon shutdown, therefore, there was a worldwide shortage of BSA and a doubling in market price. The shortage continued when Centeon cut back its bovine operations and eventually eliminated BSA production.

The Industry Reacts. The field was then open for new manufacturers whose plants were not tooled for 1950s-era BSA production techniques. These newcomers could develop albumin processes from the ground up in order to address specific needs of the diagnostic and other industries. What was required first of all was a more consistent, reproducible, and analyte-free product that would still be economical to use.

Companies forced to reevaluate and requalify their BSA vendors began scientifically researching the question of albumin performance. Previously, few had much knowledge of what made one BSA vendor or lot superior to another. Wide variation, both between vendors and among individual lots from the same vendor, was apparent but somewhat mysterious. In addition, some material seemed to perform well in certain assays but fail miserably in others. Necessity being the mother of invention, this situation began to change in a hurry.

A flurry of collaborative applications research projects got under way. These identified some problematic properties of commercial albumin that were causing industry headaches and inducing companies to look for new manufacturers with the flexibility to address them. Among these common issues were residual enzymatic activity, protease-related degradation, residual IgG, low-molecular-weight contaminants, and the presence of calcium and phosphate leading to precipitation problems.

New BSA Products and Methods

A bovine plasma collector and fractionator, AMPC Inc. (Ames, IA), had been investigating more-economical, efficient, and regulatorily compliant methods of producing higher-purity BSA for diagnostic applications since the early 1990s. With the BSA shortage in full swing, AMPC, now renamed Proliant Inc., began to roll out BSA products that benefited from many new manufacturing twists.

Improvement of Production Controls. Although Proliant relied on the tried-and-true heat-shock manufacturing process as its core technology, the company discarded the open vats and filter presses historically common in animal product manufacturing. These were replaced by a closed, computer-monitored system with continuous-flow in-line centrifuges and modern filtration skids using the latest membrane technology. In addition, the company introduced several proprietary steps designed to efficiently pull albumin and its bound analytes out of the purification stream. Statistical process control, raw-material control, and large-scale continuous-flow columns have been key elements of a process that has enabled the company to produce albumin having an unprecedented level of consistency.

Manufacturers have traditionally sold albumin in many different grades that reflect pass-fail status relative to certain analytical requirements. Customers have lot-selected for such parameters as low-endotoxin and protease-free, taking material from individual lots displaying such performance in tests. But adequate control of raw materials and manufacturing processes makes possible the establishment of low parametric values as standard specifications for all BSA produced. The Proliant process allowed albumin to be engineered from the start to be low-protease, low-endotoxin, and essentially fatty acid–free in every lot, in conformance with six-sigma statistical methods for determining compliance to specifications (see Table IV).

Critical to achievement of such consistency was gaining control over the raw material. To minimize variability in an inherently heterogeneous biological material like plasma, the company had to do three things: introduce complete ownership and traceability of the material and its collection; standardize the shortest possible time from collection to processing; and establish a rapid, just-in-time manufacturing process taking place in a closed, chilled environment. This was done and the result was an albumin with reproducibly low bioburden and endotoxin levels. Not only does the product quality indicate the effectiveness of the biological production control, but it is an important step toward reducing the presence of bacterial enzymes.

Chromatography for Purification. This and other newly available plasma protein purification technologies will continue to change the fractionation industry for animal-derived products. For example, column chromatography combined with traditional fractionation methods can serve very well to purify plasma proteins.^2,17

Chromatography can refine albumin and remove specific analytes to enable extremely low sensitivity in certain diagnostic assays. A disadvantage of chromatography is that it costs more to produce such specialty BSA in a market that is highly sensitive to price. Ion-exchange chromatography allows recovery of albumin at high purity with mild conditions, but it is generally feasible only on a small scale. Most chromatographic processes for albumin recovery involve an anion exchange based on diethylamino ethanol, usually at pH 8 on filtered plasma, followed by application of the fraction to a carboxymethyl or sulfopropyl cation exchanger near pH 5. These methods are suitable for producing albumin for high-value, low-volume, and highly sensitive applications as in molecular biology.¹⁹

In large-scale commercial production, downstream use of batch-sized columns may further increase the efficiency of ligand and fatty acid stripping. Great skill and expertise is required to ensure the absence of ligand leakage that leads to ligand rebinding to albumin during large-scale processes. Proliant has managed to achieve residual nonesterified fatty acid levels of under 0.005% with packed large-scale resin columns undergoing continuous flow at appropriate pH and temperature conditions. Such preparations also tend to produce very low to nondetectable steroid and serine-protease residuals.

Chemical Innovations. As an alternative to increased use of chromatography, the chemical environment of plasma proteins can be controlled to achieve chromatography-like results at a fraction of the cost. Manipulating temperature, pH, and dialysis conditions in the light of today's better knowledge of albumin chemistry can result in a product that offers excellent analytical and performance characteristics in even the most sensitive assays. This modern approach to classic fractionation methods avoids the regulatory, financial, and cleaning challenges that can accompany scale-up to large-scale chromatographic production.

An additional example of an innovative technology comes from ProMetic Life Sciences Inc. (Montreal), which has developed synthetic ligands through combinatorial chemistry. These ligands are specifically engineered and screened to offer both high selectivity for specific proteins and optimal adaptability for industrial-scale production. ProMetic's technology can be adapted to capture proteins from crude plasma fractions, thereby resulting in improved recovery and purity.²¹

The gene for human albumin has been isolated and cloned, and several companies have produced albumin recombinantly in relatively small quantities, primarily using yeast as an expression system.² Human albumin has also been expressed in the milk of cloned cows. The focus in this commercialization has been on the production of an economical and safer alternative to human albumin for therapeutic use.

The economy has not yet been realized, with material costing as much as $10,000/kg for nonsterile bulk powder. This compares with $2000–$3000/kg for current therapeutic-grade human albumin and $200–$400/kg for bovine albumin in the diagnostic market. Customers who use BSA primarily for in vitro assays are unlikely to pay much of a premium for nonhuman or animal-derived products, since therapeutic safety is not as great an issue. Clearly, recombinant albumin for IVD use has a long road to feasibility as an alternative to naturally purified BSA.

Conclusion

Many of the bovine serum albumin products manufactured by the altogether new or modified Cohn methods discussed here have been embraced by the industry and gained market share. But the cost and labor involved in changeover have been persistent and pervasive barriers to acceptance of some improved BSA products. Even when new production methods have resulted in a more economical, consistent, and dependable BSA, barriers to adoption can be considerable. Regulatory submissions, vendor requalifications, and the need to verify or tweak established diagnostic products in order to accommodate new raw materials all stand as potential roadblocks to market entry for new bovine albumin products. However, as more new diagnostic products come to incorporate modern BSA variants during early R&D of diagnostic products, it is likely that assay developers will discover that BSA has passed from its decades of variability into a new era of reliability.

References

1. OJ Bos, "Albumin, the "Jack of All Trades" in the Protein World" (PhD diss., University of Utrecht, The Netherlands, 1989).

2. T Peters, All About Albumin (San Diego: Academic Press, 1996).

3. KO Johanson et al., "Refolding of Bovine Serum Albumin and Its Proteolytic Fragments. Regain of Disulfide Bonds, Secondary Structure and Ligand Binding," Journal of Biological Chemistry 256 (1981): 445–450.

4. FW Putnam, ed., The Plasma Proteins, vol. IV (New York: Academic Press, 1984).

5. EJ Cohn et al., "Preparation and Properties of Serum and Plasma Proteins IV. A System for the Separation into Fractions of the Protein and Lipoprotein Components of Biological Tissues and Fluids," Journal of the American Chemical Society 68 (1946): 459–475.

6. EJ Cohn, WL Hughes Jr, and JH Weare, "Preparation and Properties of Serum and Plasma Proteins XIII. Crystallization of Serum Albumins from Ethanol-Water Mixtures," Journal of the American Chemical Society 69 (1947): 1753.

7. JA Hamilton et al., "Locations of the Three Primary Binding Sites for Long-Chain Fatty Acids on Bovine Serum Albumin," Proceedings of the National Academy of Sciences USA 88 (1991): 2051–2054.

8. XM He and DC Carter, "Atomic Structure and Chemistry of Human Serum Albumin," Nature 358 (1992): 209–215.

9. B Honore, "Conformational Changes in Human Serum Albumin Induced by Ligand Binding," Pharmacological Toxicology 66, supp. 2 (1990): 7–26.

10. U Kragh-Hansen, "Molecular Aspects of Ligand Binding to Serum Albumin," Pharmacological Review 33 (1981): 17–53.

11. U Kragh-Hansen, "Structure and Ligand Binding Properties of Human Serum Albumin," Danish Medical Bulletin 37 (1990): 57–83.

12. AA Spector, "Fatty Acid Binding to Plasma Albumin," Journal of Lipid Research 16 (1975): 165–179.

13. PD Boyer et al., "The Combination of Fatty Acids and Related Compounds with Serum Albumin I. Stabilization against Heat Denaturation," Journal of Biological Chemistry 162 (1946): 181–188.

14. T Peters, All About Albumin (San Diego: Academic Press, 1996): 11.

15. RF Chen, "Removal of Fatty Acids from Serum Albumin by Charcoal Treatment," Journal of Biological Chemistry 242 (1967): 173–181.

16. T Peters, All About Albumin (San Diego: Academic Press, 1996): 25.

17. WF Weinbrenner and MR Etzel, "Competitive Adsorption of Alpha-Lactalbumin and Bovine Serum Albumin to Sulfopropyl Ion-Exchange Membrane," Journal of Chromatography 662 (1994): 414–419.

18. T Peters, All About Albumin (San Diego: Academic Press, 1996): 77.

19. LE Trujillo et al., "A Protocol for the Purification of Bovine Serum Albumin Free of Deoxyribonuclease Activity," Biotechniques 9 (1990): 620–622.

20. T Peters, All About Albumin (San Diego: Academic Press, 1996): 303.

21. T Burnouf, "Plasma Protein Purification Technologies: What Next?," Transfusion Today, no. 42 (2000).

Michael O. Budnick is director of biologicals and Randal S. Fitzgerald is director of operations for biologicals at Proliant Inc. (Ames, IA). The authors can be reached at michael.budnick@proliantinc.com and randy.fitzgerald@proliantinc.com, respectively.

Photo Courtesy Proliant

Copyright ©2003 IVD Technology