Affinity Chromotography - from Textile Dyes to Synthetic Ligands by Design, Part 2

The first part of this review (BioPharm International July 2004) discussed the development of synthetic ligands with a focus on integrated rational design, combinatorial synthesis, and solid phase screening for ligand discovery and design. This article deals with the properties of synthetic ligands used in affinity chromatography of biopharmaceuticals with emphasis on the safety and the use of triazinyl ligands.

John Curling

PROPERTIES

Initially, dye ligands were pursued not only for their biomimetic function but also because they were readily available from major chemical companies; they were cheap and they were stable when attached to a support matrix. Very significant progress has been made since the mid-1970s when dye ligands were introduced, and the major differences and improvements are given in Table 1.

Table 1. Comparison of Synthetic Biomimetic Ligands with Earlier Dye Ligands.

The advantages of synthetic ligands compared to their biological counterparts have been summarized by Lowe et al.¹ and Boyer and Tsu.² Table 2 is an adaptation of their view, modified to encompass specifically designed ligands.

Because they are based on triazine chemistry, the synthetic ligands under discussion have some similarities with their dye predecessors. These ligands, therefore, display significant stability and resistance to both chemical and biological degradation. This property also confers long-term stability and re-usability, and synthetic ligand adsorbents can be used for at least 100 cycles.³ In this context it may be remembered that the colorfast, reactive (triazine) dyes developed during the 1950s were designed for covalent bonding to cellulosic fabrics (in contrast to earlier animal and synthetic fibers) under mild conditions with minimum leaching.

Table 2. Comparison of Biomimetic Ligands with Conventional (Biological) Ligands.

These affinity adsorbents display dynamic protein-binding capacities in the range 5 up to 40 mg/mL. Most absorbents used in bioprocessing have a capacity of 15 to 20 mg/mL at a linear flow rate of 100 cm/hour. They are comparable with Protein A adsorbents and most ion exchangers. They may be regarded as a particular compound family similar to the families of ion exchangers, and they certainly show similar characteristics in bioprocessing.

Figure 1. Triazine Structure with Two Substitution Positions and a Spacer Arm to the Matrix

LEACHABLES

As has been previously mentioned, ligand leakage was noted when Blue dextran was coupled to agarose and significantly reduced when the ligand was coupled directly. Eketorp

⁴

discussed the leakage of ligand from cyanogen bromide-coupled products, which have now generally been discarded in favor of oxirane coupling through a far more stable epoxy linkage. However, as with all adsorbents used in chromatography, leaching with or without degradation of the matrix, is an important consideration.

As use was increased in the early '90s, particularly of Cibacron Blue/CI Reactive Blue adsorbents, leachates and potential toxicity came under scrutiny. Initial studies focused on the development of immunochemical quantification, since the sensitivity of a spectrophotometric determination at 608 nm (for Cibacron Blue) was only about 1 g/mL dye.⁵ The free dye is non-antigenic, and antibodies can only be raised when the dye is conjugated with, for example, mannosylated BSA or Keyhole Limpet Hemocyanin (KLH).⁶ Using this latter method, a 1,000-fold sensitivity increase was attained in a competitive inhibition assay. Even further improvement was achieved in an enzyme-linked immunosorbent assay (ELISA) capable of measuring CI Reactive Blue at concentrations down to 10 ppm (10 nM), representing a 3,000-fold higher sensitivity over direct spectrophotometric analysis.⁷ This assay formed the basis for a commercial product, and similar assays for other ligands have subsequently been developed. For example, Novo Nordisk has reported the stability of a symmetrical synthetic ligand developed for the purification of coagulation Factor VIIa (NovoSeven).⁸ In this study, a reverse phase HPLC assay was developed with a sensitivity corresponding to a quantification limit of 6.2 ng/mL with 100 L injections. The response was linear over a 0 to 500 ng range for leachates, both in the presence and in the absence of the protein. Leakage during storage in 20% ethanol for 24 hours at 23°C was < 0.002%, and leakage significantly declined at +5°C. Cleaved ligand adsorbent conjugate, after regeneration with 0.5 M sodium hydroxide, could be washed out with 2.5 column volumes, and it was estimated that 3,000 batch runs could be performed if a 1% ligand density decrease (caused by alkaline regeneration) was accepted.

TOXICITY

As shown in Figure 1, the triazines are symmetrical hexameric rings of alternating carbons and nitrogens. Chlorotriazine derivatives such as atrazine are common herbicides and have been used since 1959 to control the growth of annual grasses and broadleaf weeds. In 1990, over 64 million acres of cropland were treated in the US. Triazines inhibit photosynthesis in plants and only show low-level toxicity in non-photosynthetic organisms. In a 1999 review, Eldridge et al. concluded that "triazines pose no reproductive or developmental hazard in animals. Triazines have also been assessed in more than 40 mutagenicity-genotoxicity tests using

in vivo

markers as well as prokaryotic and eukaryotic cells, and a complete weight of evidence analysis concluded that the herbicides are neither genotoxic nor mutagenic."

⁹

The same review concluded that "triazines have no intrinsic hormone activity and cannot support carcinogenesis on their own." Other studies report that "carcinogenic effects [of triazines] in rodents have no relevance for humans,"

¹⁰

and that they are not causally linked to mortality among triazine herbicide manufacturing workers.

¹¹

Safety and toxicological information on triazine herbicides is available from the CDC Agency for Toxic Substance and Disease Registry (ATSDR) (

www.atsdr.cdc.gov

) and the Environmental Protection Agency (

www.epa.gov

). After a re-registration review, ATSDR concluded that "atrazine is not a likely human carcinogen." It should also be noted that typical leachate levels under operational conditions are at picogram to nanogram levels — if leaching occurs at all. These levels are between six to nine orders of magnitude lower than the levels at which triazine compounds are known to be safe.

Clearly, information on herbicides is not a substitute for toxicology testing of leached ligands. However, the large body of information supporting the safety of triazine derivatives is a reasonable starting point, which, together with completed studies, indicates an adequate safety profile for the triazine-derived ligands tested to date. Such studies are normally carried out by adsorbent manufacturers at the request of clients and form part of a regulatory support file package supplied confidentially to the client. Santambien et al. looked at the in vitro toxicity of Reactive Blue 2 and Reactive Red 120,¹² but later, confidential toxicity studies were conducted. It should be noted that during ligand synthesis, the active chlorines of tri- and di-chlorotriazines are eliminated, and R-groups that could contribute to toxicity are discarded as candidate substituents.

SAFETY

Chromatographic adsorbents, eluents, and cleaning solutions that contact or potentially contact target biopharmaceutical products can potentially contaminate the final product. Synthetic ligand adsorbents can eliminate the potential hazards of animal- and microorganism-sourced ligands,

¹³

minimizing running, cleaning, and validation costs. However, the same care must be exercised with potential contaminants from eluent media and the risks from adventitious organisms. A significant advantage of synthetic ligand adsorbents is their tolerance to strong alkaline solutions, normally to 0.5 M, but for short periods to 1 M sodium hydroxide, as well as iso-propanol alkalis.

APPLICATIONS

Dye ligands such as Cibacron Blue quickly found application in investigations of plasma fractionation and the recovery of albumin, the industry driver at the time, from Cohn Fraction IV.

¹⁴

Subsequently, a synthetic analogue adsorbent was developed, and has been applied to the commercial production of yeast recombinant albumin, Recombumin, now used as an excipient.

^15,16

Procion Blue and Red triazine derivatives have been compared in the purification of α1-proteinase inhibitor with the finding that a Procion Red Fractogel was more effective.

¹⁷

Lysine-agarose was used to purify plasminogen from plasma, and the ligand was later adapted (L-lysine linked through 6-aminohexanoic acid to agarose) to the large-scale recovery of the zymogen plasminogen, leading to an active plasmin product, from Fraction II + III of the Cohn process.

^18-20

A case study by Datar et al. reports on extension of lysine-agarose use to the commercial recovery of tissue plasminogen activators (tPA) expressed in

E. coli

.

²¹

tPA has also been fractionated on a lysine derivative.

²²

A mimic of Protein A has been constructed through rational design and subsequent synthesis on a triazine backbone.^23-25 Two variants have been commercialized as MAbsorbent A1P and A2P, and A2P has been applied to the isolation of polyclonal IgG from ethanol precipitates of the Cohn and Kistler-Nitschmann processes.^26-28

Following the rapid introduction of recombinant DNA technology into biopharmaceutical development, affinity chromatography was applied to cytokine purification, tissue necrosis factors, and, as mentioned previously, tPA purification.^29,30 Interferon-α was purified on Blue-agarose, and later work describes the application of a "Mimetic Green" adsorbent to interferon-α.³¹ A simpler synthetic ligand, phenylboronate, has been used as a capture step for erythropoeitin.³² Phenylboronate adsorbents are otherwise used for purification of glycoproteins containing cis-diols, as reviewed by Liu and Scouten. ³³

Figure 2. Ligand Structure for Capture of Amediplase, a tPA-urokinase Fusion Protein. PuraBead® 6HF is a new "High Flow" agarose intermediate.

The biopharmaceutical industry, an adsorbent vendor, and academia collaborated in designing a novel synthetic ligand to recover the recombinant human insulin precursor MI3 from the crude fermentation broth of S. cerevisiæ.³⁴ In similar work, a novel, synthetic ligand was designed to replace a monoclonal antibody immunoaffinity step in the process for Factor VIIa (NovoSeven).^35,36

Like tPA, urokinase has been purified using lysine derivatives. In addition, isolation of urokinases from urine has been reported using p-aminobenzamidine derivatives.^37,38 This and earlier work on tPA and plasminogen suggest that ligand structures containing positively charged groups, for example lysine and benzamidine, are candidates for improved downstream processing of a tPA-urokinase fusion protein — Amediplase. However, and surprisingly, screening general synthetic ligand libraries revealed ligand candidates (and a lead ligand through secondary screening), which contained strongly negatively charged groups, as shown in Figure 2. This led to the development of an entirely new synthetic ligand adsorbent for the tPA-urokinase fusion.³⁹

The synthetic adsorbent developed for this product showed a capacity of 11.4 mg/mL, with ≥ 99% purity at 92% recovery from cell culture supernatant (5 mg/mL target protein) when operated at a column flow rate of 300 cm/hr. The adsorbent can be sanitized with 0.5 M sodium hydroxide and shows excellent stability in 1 M sodium hydroxide, 30% propan-2-ol/0.2 M sodium hydroxide, and in 20% ethanol. This demonstrates the benefits of developing synthetic ligands by screening combinatorial libraries and the major advance from the existing lysine-adsorbent (1 mg/mL capacity) chromatography, which produced a < 80% pure protein and could only be used for < 50 cycles.

FUTURE PERSPECTIVES

It is clear that what was originally referred to as dye-affinity or pseudo-affinity chromatography has been superseded by the far more exact science of synthetic ligand chromatography. Even "biomimetic" may be a misleading term if it refers to a mimic of a dyestuff. The later ligands have all been developed from combinatorial libraries, frequently after computational chemistry study of the protein's target binding epitopes. The triazine backbone with diverse amine substitution has been a versatile and successful chemistry for synthetic ligand design and attachment through stable epoxy linkages to agarose and other matrices, such as those based on methacrylates.

This article has described the move towards dedicated, custom design of synthetic ligand adsorbents that are specifically designed and manufactured for the protein purification task at hand. General bioseparation unit operations of ion exchange chromatography and membrane steps will always retain their workhorse position in the biopharmaceutical industry. However, with ever increasing demands on purity, yield, and reliability — repeatability and consistency — synthetic ligand chromatography is set to assume a significant position in the bioseparations armory.

REFERENCES

1. Lowe CR, Burton SJ, Burton N, Stewart DJ, Purvis DR, Pitfield I, et al. New developments in affinity chromatography.

J Mol Recognit

. 1990 Jun; 3(3):117-22.

2. Boyer PM, Hsu JT. Protein purification by dye-ligand chromatography. Adv Biochem Eng Biotechnol. 1993; 49:1-44.

3. More JE, Hitchcock AG, Price S, Rott J, Harvey MJ. Dye-protein interactions: developments and applications. In: Vijayalakshimi MA, Bertrand O, editors. London: Elsevier; 1989. p. 265.

4. Eketorp R. Affinity chromatography in industrial ethanol fractionation of human plasma. In: Curling JM, editor. Methods of plasma protein fractionation. London: Academic Press; 1980. p. 175-188.

5. Hulak I, Nguyen C, Girot P, Boschetti E. Immobilized cibacron blue — leachables, support stability and toxicity on cultured cells. J Chromatogr. 1991; 539(2):355-62.

6. Santambien P, Girot P, Hulak I, Boschetti E. Immunochemical quantification of procion red HE-3B used as ligand in affinity chromatography. J Biochem Biophys Methods 1992 Jun; 24(3-4):285-95.

7. Stewart DJ, Purvis DR, Pitts JM, Lowe CR. Development of an enzyme-linked immunosorbent assay for C.I. Reactive Blue 2 and its application to a comparison of the stability and performance of a perfluorocarbon support with other immobilised C.I. Reactive Blue 2 affinity adsorbents. J Chromatogr. 1992; 623:1-14.

8. Christensen J., Mollerup I, Breinholt J. Documentation of leakage from a triazine based mimetic affinity matrix developed for purification of coagulation factor VIIa. Poster presented at the European Congress of Biotechnology 10, 1999 July 11-15; Brussels, Belgium.

9. Eldridge JC, Wetzel LT, Stevens JT, Simpkins JW. The mammary tumor response in triazine-treated female rats: a threshold-mediated interaction with strain and species-specific reproductive senescence. Steroids 1999;64(9):672-8.

10. Stevens JT, Breckenridge CB, Wetzel L. A risk characterization for atrazine: oncogenicity profile. J Toxicol Environ Health A 1999 Jan 22; 56(2):69-109.

11. MacLennan PA, Delzell E, Sathiakumar N, Myers SL. Mortality among triazine herbicide manufacturing workers. J Toxicol Environ Health A 2003 Mar 28; 66(6):501-17.

12. Santambien P, Sdiqui S, Hubert E, Girot P, Roche AC, Monsigny M, et al. In vitro toxicity assays for dye ligands used in affinity chromatography. J Chromatogr B Biomed Appl. 1995 Feb 3; 664(1):241-6.

13. Behizad M, Curling JM. Comparing the safety of synthetic and biological ligands used for purification of therapeutic proteins. Biopharm 2000; 13(9): 42-46.

14. Harvey MJ. The application of affinity chromatography and hydrophobic chromatography to the purification of serum albumin. In: Curling JM, editor. Methods of plasma protein fractionation. London: Academic Press; 1980. p. 189-200.

15. Burton SJ. Considerations in the use of synthetic dye-ligand adsorbents for the manufacture of blood proteins. Biotechnol Blood Proteins 1993; 227:19-24.

16. Balance DJ. Yeast-derived recombinant human albumin (Recombumin). Presented at Biological Safety and Production 99. 1999 April 19-23; Boston, USA.

17. Gunzer G, Hennrich N. Purification of alpha 1-proteinase inhibitor by triazine dye affinity chromatography, ion-exchange chromatography and gel filtration on Fractogel TSK. J Chromatogr. 1984 Jul 27; 296:221-9.

18. Summaria L, Spitz F, Arzadon L, Boreisha IG, Robbins KC. Isolation and characterization of the affinity chromatography forms of human Glu- and Lys-plasminogens and plasmins. J Biol Chem. 1976 Jun 25; 251(12):3693-9.

19. Blomqvist I, Ekström, Gustavsson J, Westergren H, Mattson A, Erikhans M, et al. Characterisation of ECH-Lysine Sepharose Fast Flow — a new affinity chromatography support. Poster presented at Plasma Product Biotechnology Meeting; 2003 April 22-26; Curacao, Netherlands Antilles.

20. Alonso WR, Dadd C, Rebbeor J, St Peter M, Yuziuk J, Korneyeva M, et al. Plasmin process development. Presentation at Plasma Product Biotechnology Meeting; 2003 April 22-26; Curacao, Netherlands Antilles.

21. Datar RV, Cartwright T, Rosen C-G. Process economics of animal cell and bacterial fermentations: a case study analysis of tissue plasminogen activator. Bio/Technology 1993 Mar; 349-357.

22. Mori K, Dwek RA, Downing AK, Opdenakker G, Rudd PM. The activation of type 1 and type 2 plasminogen by type I and type II tissue plasminogen activator. J Biol Chem. 1995 Feb 17; 270(7):3261-7.

23. Li R, Dowd V, Stewart DJ, Burton SJ, Lowe, CR. Design, synthesis, and application of a Protein A mimetic. Nature Biotechnology 1998 Feb; 16:190-195.

24. Teng SF, Sproule K, Hussain A, Lowe CR. A strategy for the generation of biomimetic ligands for affinity chromatography: combinatorial synthesis and biological evaluation of an IgG binding ligand. J. Mol. Recogn. 1999; 12:67-75.

25. Teng SF, Sproule K, Husain A, Lowe CR. Affinity chromatography on immobilised "biomimetic" ligands. Synthesis, immobilisation and chromatographic assessment of an immunoglobulin G-binding ligand. J Chromatogr B. 2000; 740:1-15.

26. Russell C, Baines D, Burton SJ. Application of a synthetic adsorbent for the purification of immunoglobulins. Presented at Recovery of Biological Products IX Conference; 1999 May 23-28; Whistler, Canada.

27. Baines D, Burton M, Burton S, Curling J. Synthetic ligand affinity adsorbents for highly selective purification of human plasma proteins. Presented at Recovery of Biological Products X Conference; 2001 June 3-8; Cancun, Mexico.

28. Curling JM, Baines D, Russell C, Watson K, Ward E, Pollard H, Burton S. Recovery of IgG from ethanol precipitates of the Cohn-Oncley and Kistler-Nitschmann fractionation schemes. Poster presented at Plasma Product Biotechnology Meeting; 2003 April 22-26; Curacao, Netherlands Antilles.

29. Zoon KC, Smith ME, Bridgen PJ, zur Nedden D, Anfinsen CB. Purification and partial characterization of human lymphoblast interferon. Proc Natl Acad Sci USA 1979 Nov; 76(11):5601-5.

30. Werner RG, Berthold W. Purification of proteins produced by biotechnological process. Arzneim-Forsch/Drug Res. 1988; 38(1 Nr. 3):422-428.

31. Swaminathan S, Khanna N. Affinity purification of recombinant interferon-a on a mimetic ligand adsorbent. Protein Expr. Purif. 1999; 15:236-242.

32. Zanette D, Soffientini A, Sottani C, Sarubbi E. Evaluation of phenylboronate agarose for industrial scale purification of erythropoietin from mammalian cell cultures. J Biotechnol. 2003; 101:275-287.

33. Liu XC, Scouten WH. Boronate affinity chromatography. Methods Mol Biol. 2000; 147:119-28.

34. Sproule K, Morrill P, Pearson JC, Burton SJ, Hejnæ KR, Valore H, et al. New strategy for the design of ligands for the purification of pharmaceutical proteins by affinity chromatography. J Chromatogr. B 2000; 740:17-33.

35. Mollerup I, Lowe CR, Sproule K, Morrill P, Burton S, Pearson J, et al. Design and development of ligands for affinity purification of rFVIIa. Presented at Recovery of Biological Products IX Conference; 1999 May 23-28; Whistler, Canada.

36. Christensen J, Mollerup I, Lowe C, Sproule K, Morrill P, Burton S, et al. Comparison of different types of affinity ligands for the purification of Factor VIIa. Presented at Recovery of Biological Products IX Conference; 1999 May 23-28; Whistler, Canada.

37. Male KB, Nguyen AL, Luong JHT. Isolation of urokinase by affinity ultrafiltration. Biotechnol. Bioeng. 1990; 35:87-93.

38. Takahashi R, Akiba K, Koike M, Noguchi T, Ezure Y. Affinity chromatography for purification of two urokinases from human urine. J Chromatogr B Biomed Sci Appl. 2000 May 26; 742(1):71-8.

39. Burton SJ, Sawkins R, Pearson J, Collins S. Development of a robust affinity adsorbent for primary capture and purification of a tPA-uokinase fusion protein. Presented at Recovery of Biological Products XI; 2003 Sep 14-18; Banff, Canada.