Albumin is the most abundant serum protein. It serves several functions in vivo: e.g., binding and transport of fatty acids, hormones, and metal ions; maintenance of osmotic pressure and pH; and binding of exogenous toxins and products of lipid oxidation (1). Over time, development of large–scale purification methods have translated those functions into diagnostic, cell culture, and microbiological applications. It is important to note, however, that purification procedures can promote molecular changes and thereby add to the already complex nature of albumin. That increase in heterogeneity can cause unpredictable performance in vitro. Consequently, it is important for suppliers to understand how the process modifies the end product — and for buyers to consider the potential effects that process-induced modifications can have on the performance of their process. Relying strictly on published product specifications is often insufficient when making a decision to purchase.

PRODUCT FOCUS: ALL BIOLOGICS
PROCESS FOCUS: PRODUCTION
WHO SHOULD READ: MANUFACTURING, PROCESS DEVELOPMENT, QA/QC
KEYWORDS: CONTAMINATION, MEDIA, ANALYTICAL METHODS, TSE, RAW MATERIALS
LEVEL: INTERMEDIATE

In Part One, we focused on the molecular structure of bovine albumin. The following brief descriptions of albumin fractionation methods and processing will provide a sense of the impact a selected purification method has on the final recovered albumin, how purified albumin can be further modified post fractionation, and how the processing has the potential to alter the final product.

Manufacturing

Three classical industrial methodologies are used to purify albumin from serum or plasma: cold-organic solvent fractionation, heat shock, and ion-exchange chromatography. Each will produce ≥96% electrophoretically pure product; however, the potential for process perturbations to increase albumin heterogeneity is a serious consideration. Consequently, both manufacturer and user should focus on product performance and analytical specifications that directly relate to performance.

Most albumins are manufactured to meet standard specifications; however, what is not typically measured or reported may be critical to the protein’s overall performance. This includes the molecule’s SH groups, polymer content, and contaminants such as apolipoproteins, nucleases, peptidases, phospholipases, and alkaline phosphatase. Peptidases are especially problematic when albumin is used as an immunoassay carrier (3). Another factor to consider is deamidation of the amino acid residues asparagine and glutamine caused by acetylation. The process apparently occurs in-vivo (7), but it can also occur at a pH >10 or at high temperatures. Deamidation promoted by copper in an oxidative environment also can cause albumin fragmentation (17).

A strong indication that the manufacturing process affects albumin is discussed in a study by Janatova et al., who found that the dimers recovered by anion-exchange chromatography from freshly prepared or aged albumin differ in their reactivity with thioglycolate. The population from aged albumin was significantly more resistant to reduction to a monomeric state, suggesting that the type of dimer present in a given albumin preparation may be directly related to prior procedural steps. Disulfide bond formation may not be the only albumin polymer-forming mechanism (4). Consequently, the state and storage of purified albumin before sale and/or use should be considered. Liu et al. found that both the moisture content of a freeze-dried preparation and its storage temperature can profoundly affect polymer generation (18).

Cold-organic solvent fractionation of proteins is based on the following variables: temperature, pH, ionic strength, organic solvent concentration, and protein concentration. By altering those variables, groups of proteins are selectively precipitated. The process conditions applied to human plasma were published by EJ Cohn in 1946 (19). He defined five protein groups by their respective manipulation of five variables: fraction I (fibrinogen and other clotting factors), fraction II–III (gamma globulin, lipoproteins), fraction IV-1 and IV-4 (beta globulin, alpha globulins), and fraction V (albumin). Typically, recovered albumin is highly monomeric and rich in fatty acids. The Cohn process has been used to purify bovine albumin for use in both mammalian cell culture and microbiological nutrient media. Because the process does not completely separate the alpha, gamma, and beta globulins from albumin, it is important to identify such contaminants in the final preparation — in particular, levels of IgG and transferrin. Cohn Fr V is typically specified at ≥96% pure.

Heat-shock purification is based on albumin’s thermal stability compared with most other plasma proteins. When stabilized with 0.02–0.04 M caprylic acid, albumin can be recovered from serum adjusted to pH ~5 at a temperature of ~60 °C. Other serum proteins are denatured and rendered insoluble under such conditions. Harvested albumin may be concentrated either by ultrafiltration or precipitation. Typical purity of a heat-shock albumin is ≥98%, with a caprylic acid level generally ≤0.5 mol/mol albumin (which may run higher). Users interested in albumin substantially free of fatty acids rely on acid-charcoal treatment of this albumin, which can reduce them to 3, 5).

Ion-Exchange Chromatography: Of all purification methods, chromatography may be the best candidate for minimizing process-generated albumin heterogeneity. Curling described albumin purification from plasma by sequential anion, cation, and gelfiltration chromatographic procedures (20). The purified product was reported to be 99% pure as measured by cellulose acetate electrophoresis and >97% monomer, with a final yield >90%. Curling also reported that IgG, IgA, IgM, prealbumin, alpha 2 glycoprotein, alpha 2 macroglobulin, transferrin, haptoglobin, hemopexin, and alpha 1 lipoprotein were immunochemically undetectable. Importantly, the absence of alpha 1 lipoprotein improved the final product stability. He also was able to show significantly more final-product color measured at 403 nm than with Cohn-derived albumin (20).

Gamma Irradiation: Another process input that affects albumin molecules and should be noted is gamma irradiation. For some time, irradiation has been used to “sterilize” albumin in a dry or frozen liquid state. Generally the effects of ionizing radiation on this protein have not been addressed by suppliers or users, with the possible exception of quantifying the level of polymerization — or in some cases ensuring that a particular property (e.g., efficacy in cell culture) was not destroyed. If the irradiated product does not perform to standard, however, tracing root causes would not typically involve the investigation of molecular changes.

Consequences were addressed by Alexander and Hamilton, who were able to demonstrate a linear reduction in several amino acid residues when albumin was exposed to gamma irradiation (9). They were also able to show generation of a new amino acid (alpha-amino-N-butyric acid), an increase in the number of reactive SH groups, the formation of carbonyl groups, and changes in chemical reactivity related to SH groups and ultraviolet light absorption (9). In a later study, Zakrzewski reported formation of carbonyl groups when albumin was irradiated in solution and a decrease in SH groups when the protein was irradiated dry (10). The apparent discrepancy in SH results is covered by Alexander’s further claim that the SH groups are very labile and prone to oxidation (21). In his study with Borrelia burgdorfori, Blevins showed that irradiation had no measurable effect on either the albumin’s ability to support growth or expression of the microbe’s outer-surface protein C (22).

Albumin Polymerization: Commercially available albumin normally contains dimer, trimer, and traces of higher polymeric forms.The formation of polymers is an inadvertent byproduct of specific steps in their manufacturing process such as an acidic pH (7,8), high temperatures (8,9), radiation (9,10), dehydration caused by exposure to an organic solvent (11), and lyophilization or spray drying. Tattini et al. state that lyophilization can cause protein unfolding, a decrease in the alpha-helix structure, an increase in the beta sheet, and the formation of disulfides (23).

Gao concluded that polymerization results from three events: changes in secondary structure, covalent disulfide bond formation and, peptide bond formation. It was further suggested that thermal unfolding could generate mixed protein oligomers (heterogeneous polymers) (8). In fact, it is common to find commercially available albumins with a range of high–molecular-weight albumin polymers. Peters recommends that, unless the thiol group is required, it should be blocked to prevent polymerization from occurring over time, especially if albumin will be in an alkaline environment >pH 7.4 (3).

Polymerization can also be promoted chemically on purpose. Two commonly used polymerizing agents are 1-ethyl-3-(dimethylaminopropyl)-carbodiimide (C₈H₁₇ N₃•HCl, EDC) and glutaric acid dialdehyde (glutaraldehyde, C₅H₈O₂). EDC causes the peptide bond formation between acids (e.g., the carboxyl end terminal of a protein, COOH) and amines (e.g., the amino end of a protein, NH₂) (24). Glutaraldehyde can also cross-link albumin. Payne demonstrated the glutaraldehyde/albumin reaction conditions that caused polymerization up to decamers of the protein while maintaining its solubility (25). In addition, he found that the reaction did not alter albumin’s pI or optical rotary dispersion. The process maintained the protein’s native conformation and showed no significant disruption of its amino acid profile (25).

However, care must be taken when exposing albumin to polymerization conditions because excessive polymerization (aggregation) can result in a protein’s insolubility. Tanaka demonstrated the effective use of glutaraldehyde (0.06%) and formaldehyde (0.1%) to polymerize the molecule for use in red blood cell immunohematology. The reaction produced 24% dimer, 12% trimer, and 8% higher–molecular-weight polymers (26). Regardless of the means used, polymer-rich albumin is very effective in enhancing agglutination of red blood cells and accelerating the rate of reaction.

Over the past decade, certain fractionation techniques have been identified as robust reduction steps for transmissible spongiform encephalopathies (TSEs). In 1998, Blum et al. reported on a series of bovine albumin purification procedures with total clearance of 16.2 log₁₀. Their process included two heating steps and two passes through ion-exchange and charcoal filtration (27). In 2000, Foster et al. reported on the effectiveness of several common human plasma protein fractionation techniques in reducing the presence of TSEs by ~3–5 log₁₀. Their procedures included cold-ethanol precipitation, depth filtration, and anion- and cation-exchange chromatographies (28).

In 2006, Thyer et al. also demonstrated the effectiveness of both anion- and cation-exchange chromatographies individually or sequentially in reducing the presence of prions in purified human albumin and immunoglobulins (29). Individually, prion reduction was ≥4 and ≥3 log₁₀, respectively; with both, the reduction was ≥5.6 (29). Finally, a developing strategy available to users of bovine albumin is to source the protein directly from geographical BSE risk category I (GBR-I) countries such as New Zealand and Australia or from US manufacturers that derive their albumin from GBR-I serum or plasma.

Discussion and Implications

In nature, albumin’s structure gives it the capacity to perform several functions. It would be reasonable to assume, then, that a manufacturing procedure that keeps the protein’s secondary and tertiary structures in place would be best suited to produce albumin for in vitro use. The exception may be when the product is used in blood banking, which requires an aggregated or polymerized albumin. In that case, precisely managing polymerization would be essential.

Decades of research have shown that albumin is a complex protein susceptible to modifications that can alter its performance. So we suggest that in making a manufacturing or purchase decision, special attention should be given to the structure and heterogeneity of commercially available albumin. Albumin has often been viewed as a commodity, but it is an undeserved label. If anything, experience has taught us that commercially available albumins are frequently very different from one another. Even subtle differences often cause product performance variances, not only among manufacturers, but also among lots from the same manufacturer. The common reaction to those differences is to test lots until one works, a tactic that is costly to both suppliers and users.

So our message is two-fold: The industry needs to augment its understanding of which aspects of albumin are critical to performance quality (e.g., alpha-helix and beta-sheet conformations, levels and types of contaminants, and degrees of fragmentation/polymerization) and design manufacturing procedures that produce albumin that consistently meets its critical quality attributes (CQAs). W hat does this all mean? We should be looking to a process that best fits our intended applications. If heterogeneity is inevitable, we can at least identify the factors that add value to a product and design a manufacturing procedure that controls those factors.

The absence of contaminants is not necessarily required. Some may work cooperatively with albumin in certain circumstances. Consequently, best-case in-process manufacturing and finished-goods tolerance levels should be determined and optimized to deliver a total albumin product that performs consistently. For example, if albumin is used in cell culture or microbiological media, what best-practice manufacturing procedures will deliver not only an albumin that performs its binding and transport functions best and consistently, but also that consistently delivers the “contaminants” that augment its performance?

If albumin is used as a component in diagnostic reagent kits, its molecular state is also critical. What is the SH/albumin molar ratio? Should free SH groups be blocked? W hat is the fatty acid composition? W hat is the polymerization? Is the pI within acceptable limits? Does the albumin have normal near- and far-UV spectra, demonstrating intact secondary and tertiary structures? What is its degree of hydrophobicity?

Thus, in making a purchase or manufacturing decision, it is not sufficient to consider only the typical published product specifications (e.g., percent purity, percent protein, pH, sodium and chloride ion content, IgG contamination, endotoxin level, and proteolytic activity). Those are important to overall product quality, but they are only part of what’s necessary to determine whether or not a product will perform in a given user’s systems. Both supplier and customer should also consider the protein characteristics listed here (e.g., albumin/SH molar ratio, N-F transition, secondary and tertiary structures, degree of polymerization, polymer profile, heterologous and homogolous polymers, immunochemical protein contaminant profile, pI, fatty acid and lipid profiles, and hormone profile). This will help create a correlation matrix of product performance measured against all quality metrics.

Our comments and questions are meant to start a dialogue between suppliers and consumers. It is insufficient to look at the typical analytical specifications as sole product quality indicators. Ultimately, a product must perform as required. Performance may in part depend on control of one or more typical analytical specifications; however, it may also rely on factors that are not usually measured or identified. Consequently, once an albumin has been validated for use by a customer, the manufacturer must precisely control all processing steps that led to the production of that successful product. Manufacturers must be cognizant of how process variabilities will affect the factors noted here, particularly if they in turn affect performance consistency.

REFERENCES

1.) Bertucci, C, and E. Domenici. 2002. Reversible and Covalent Binding of Drugs to Human Serum Albumin: Methodological Approaches and Physiological Relevance. Curr. Med. Chem. 9:1463-1481.

2.) Peters, T. 1985. Serum Albumin. Adv. Prot. Chem. 37:161-245.

3.) Janatova, J. 1968. The Heterogeneity of Bovine Albumin with Respect to Sulfhydryl and Dimer Content. J. Biol. Chem. 243:3612-3622.

4.) Chen, RF. 1967. Removal of Fatty Acids from Serum Albumin By Charcoal Treatment. J. Biol. Chem. 242:173-181.

5.) Foster, JF. 1977.Albumin Structure, Function, and Uses, Pergamon Press, Oxford:53-84.

6.) Gao, Y. 2001. Interchain Disulfide Bonds Promote Protein Cross-Linking During Protein Folding. J. Biochem. 129:179-183.

7.) Alexander, P. 1960. Irradiation of Proteins in the Solid State, 2: Chemical Changes Produced in Bovine Serum Albumin. Rad. Res. 13:214-233.

8.) Zakrzewski, K. 1973. Cross-Linkages in Human Serum Albumin Polymerized By Gamma Irradiation in Solution. Rad. Res. 53:124-133.

9.) Yari, F. 2007. Simultaneous Purification and Polymerization Method for Bovine Serum Albumin Preparation. Ital. J. Biochem. 56:163-165.

10.) Marx, G, and M. Chevion. 1986. Site-Specific Modification of Albumin By Free Radicals: Reaction with Copper(II) and Ascorbate. Biochem J. 236:397-400.

11.) Liu, WR, R Langer, and AM. Klibanov. 1991. Moisture-Induced Aggregation of Lyophilized Proteins in the Solid State. Biotechnol. Bioeng. 37:177-184.

12.) Cohn, EJ. 1946.. J. Am. Chem. Soc. 68:459.

13.) Curling, JM. 1980.Plasma Protein Fractionation, Academic Press, London:77-91.

14.) Alexander, P, and LD. Hamilton. 1960. Iradiation of Proteins in the Solid State. Radiation Res. 13:214-233.

15.) Blevins, JS. 2009. A Readily Available Source of BSA Consistently Supports Borrelia burgdorferi Cultivation and Differential Gene Expression. BioProcess Int. 7:26-34.

16.) Tattini, V. 2005. Effect of Lyophilization on the Structure and Phase Changes of PEGylated-Bovine Serum Albumin. Int. J. Pharmaceut. 304:124-134.

17.) Tedder, JM. 1972.Chapter 6Basic Organic Chemistry, John Wiley & Sons, London:305-342.

18.) Payne, JW. 1973. Polymerization of Proteins with Glutaraldehyde: Soluble Molecular-Weight Markers. Biochem. J. 135:867-873.

19.) Tanaka, K. 2001. Isolation of Bovine Plasma Albumin By Liquid Chromatography and Its Polymerization for Use in Immunohematology. Brazil. J. Med. Biol. Res. 34:977-983.

20.) Blum, M. 1998. A Bovine Spongiform Encephalopathy Validation Study for Aprotinin and Bovine Serum Albumin. BioPharm Int.:28-34.

21.) Foster, PR. 2000. Studies on the Removal of Abnormal Prion Protein By Processes Used in the Manufacture of Human Plasma Products. Vox Sang. 78:86-95.

22.) Thyer, J. 2006. Prion-Removal Capacity of Chromatographic and Ethanol Precipitation Steps Used in the Production of Albumin and Immunoglobulins. Vox Sang. 91:292-300.