Implications of Trace Levels of Redox-Active Metals in Drug-Product Formulation

Protein pharmaceuticals are vulnerable to chemical and physical instability during synthesis/biosynthesis, isolation, purification, and storage (1–3). The chemical and physical stability of proteins are often interrelated, and oxidation is one of the degradation pathways that affect chemical stability. Although oxidation represents only one subcategory of the chemical degradation pathways of proteins, it is often observed that oxidative modification of proteins leads to covalent intermolecular cross linkages, which may then result in physical instability (4–7).

One of the contributing factors for oxidation in protein pharmaceuticals is the presence of redox-active transition metals as impurities (8). Redox-active transition metals are present in most buffers, excipients, and/or incorporated from the manufacturing process of protein pharmaceuticals. These residual metals are found to affect the processing and/or storage of biologic pharmaceuticals and are capable of generating radicals by redox reactions with organic, inorganic substrates, and molecular oxygen (9). The major mechanisms of oxygen activation by metal ions involve Fenton/Haber-Weiss chemistry and autoxidation. The substrates are molecular oxygen, superoxide (O₂-•) and hydrogen peroxide (H₂O₂). Reaction 1 illustrates the conversion of H₂O₂ into hydroxy radical (OH) via the oxidation of a metal cation (M):

The metal cation that has just been oxidized may be turned back to Reaction 1 by reduction with O₂-•:

The balance of these two reactions is:

Thus, the two oxidation states of a metal cation (Mn⁺ and M⁽ⁿ⁺¹⁾⁺) can form a catalytic electron-transfer (redox) couple. The substrates may originate from cell metabolism and/or autoxidation of some metal ions with molecular oxygen:

The yield of O

₂

-• and H

₂

O

₂

in

Reactions 4

and

5

strongly depends on the pH and/or presence or absence of chelators.

Reactions 2

and

5

indicate that, depending on the reaction conditions, O

₂

-• may serve as either the reductant or oxidant of a metal ion. Metal compounds interacting with H

₂

O

₂

produce not only free OH radicals but also other strong oxidants, such as singlet oxygen, and metallo-oxo and peroxo species, all capable of damaging proteins.

The formation of metal associated oxidants at sites of metal binding allows for the explanation of the often-observed site specificity of metal-mediated oxidative damage. Figure 1 illustrates the sequence of events in metal-catalyzed oxidation (MCO) (1, 10) of proteins, whereby the reduced form of the metal generated in Reaction 1 generates reactive oxygen species (ROS) upon reaction with oxygen or peroxide. These ROS are then free to interact with native protein, thereby oxidizing them and regenerating the oxidized form of the metal. Many amino acids, such as cysteine, methionine, tyrosine, lysine, tryptophan, and histidine, are susceptible to MCO within a protein sequence (11). MCO usually occurs site specifically, where amino acids ligating the redox-active transition metal are targets for oxidation (12, 13).

The typical pharmaceutical industry practice of screening for oxidative susceptibility of a molecule, from a drug-product development perspective, has been to quantify the potential amount of trace metals present in multiple lots of excipients and buffer components to be used in the drug-product formulation, and to identify the potential for degradation from metals to the drug of choice by spiking the formulation with relatively high amounts (10-100 ppm) of redox-active metals. This practice, although effective with small-molecule drugs, can prove to be ineffective with biologics because of the availability of multiple residues for oxidation and the conformational structure of proteins altering the degradative susceptibility of these residues in proteins (14, 15). It is apparent that the sensitivity of trace-metal quantification methods is not absolute below 0.25 ppm, and hence, there is a significant chance that a mixture of metals below the quantification level can be present in the drug-product formulation, which can cumulatively lead to a final concentration of these metals in the low ppm range. Formulations of biologics are extremely sensitive to the presence of metals, in which case, small quantities of redox-active metals can lead to a significant amount of aggregation (16) or fragmentation (17).

Another aspect of protein formulations is that for proteins/biologics susceptible to degradation in the presence of low ppm concentration of redox-active metals (~1 ppm), it is imperative that there is no lot-to-lot variability in the specifications of metals in the various excipients used in the formulation. This aspect is almost impossible to guarantee throughout the life of the product. It is, therefore, important to study this aspect and incorporate appropriate controls, such as metal chelators, to account for the variability in excipients. Incorporation of a metal chelator, purely as a counter measure for unforeseeable process changes, especially to decrease protein aggregation/fragmentation from free radicals generated by residual and unquantifiable amount of metal impurities in excipients and buffers, is not a widely used practice in the biopharmaceutical industry. The authors demonstrate that there can be advantages in incorporating typical metal chelators, such as polyamine carboxylate derivatives (e.g., ethylenediaminetetraacetic acid [EDTA] and diethylene triamine pentaacetic acid [DTPA]), in an attempt to avoid metal-catalyzed processes.

MATERIALS AND METHODS
Biologic A is a chimeric antibody expressed in Chinese hamster ovary (CHO) cells, and purified by standard chromatographic techniques. Manganese chloride (MnCl₂), copper chloride (CuCl₂), and iron chloride (FeCl₂) were purchased from Alfa Aesar. EDTA salt solution (0.5 M) was purchased from Sigma-Aldrich and DTPA was purchased from Mallinckrodt. Novex 4-20% bis-tris gels, tris-glycine-sodium dodecyl sulfate (SDS) sample buffer, 10X tris-glycine-SDS running buffer, and Mark12 molecular weight standard were purchased from Invitrogen. GelCode blue-stain reagent was purchased from Pierce.

Effect of metal ions and chelating agents on mAb stability
Samples were prepared at a final concentration of 10 mg/mL Biologic A in 10 mM succinate buffer at pH 5.0, containing 0.4% w/v poloxamer 188 and spiked with either the metal ion and/or metal ion in the presence of EDTA or DTPA as described in the following. Because the purpose of the study was to evaluate only the effect of residual quantities of trace metal ions, the final concentration of the individual redox-active metal ion was kept to 1 ppm. In this investigation, MnCl₂, CuCl₂, and FeCl₂ were used as the redox-active transition metals and were added to a target concentration of 1 ppm. Stock solutions of metal chelators (disodium salt of EDTA and DTPA) were prepared in the same buffer solution that contained Biologic A, and spiked into the formulation solutions at required concentrations. The samples were then sterilized by filtration through 0.22 μm syringe filters (Millipore) and stored at 5 °C, 25 °C, or 45 °C. Samples were collected at designated time points and analyzed for appearance, pH, protein concentration (A₂₈₀), and percentage of high-molecular-weight (HMW) species by size-exclusion high-performance liquid chromatography (SE-HPLC). Samples stored at -80°C were used as controls. Experiments were set up such that the each metal-spiked sample with a chelator had a positive control without any chelator and a negative drug-product control that contained neither the metal spike nor the chelator.

Size-exclusion chromatography
Size-exlusion chromatrography (SEC) was used to quantify the percentage of HMW species that were formed at regular time points. The analysis was conducted using a Tosohaas TSK-gel column (7.8 mm x 30 cm, particle size 5 μ, G3000SWXL) containing a Tosohaas TSK-gel guard column (6.0 mm x 40 cm, particle size 7 μ, Guard SWXL). The mobile phase comprised of 10 mM potassium phosphate buffer with 0.9% sodium chloride, pH 6.8. The samples were analyzed at a flow rate of 1 mL/min for 20 min and monitored at a wavelength of 280 nm. The samples were maintained at 4 °C and the column was maintained at room temperature throughout sample analysis.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
The extent of antibody aggregation and/or fragmentation was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing and non-reducing conditions using Novex 4%-20% bis-tris gels. Prior to electrophoresis, the samples were diluted to a protein concentration of 1 mg/mL using purified water (Milli-Q, Millipore) and then heated at 80 °C for 2 minutes in SDS sample buffer, with or without reducing agent dithiothreitol (DTT) for reducing and non-reducing gels, respectively. Mark12 molecular weight standard was used as molecular weight marker and protein bands were visualized by Coomassie blue stain (GelCode Blue, Pierce). After fixing the gels with a solution containing 50% methanol and 7% acetic acid, the gels were scanned using a densitometer (SI Amersham Biosciences). The intensity of protein bands was quantified using an image-density analysis software (Quantity One, Bio-Rad).

RESULTS AND DISCUSSION
Quantitation of trace metals in excipients such as poloxamer 188 and sucrose, prepared in 10 mM succinate buffer at pH 5.0 at concentrations identical to the drug-product formulation. Formulation samples with no metal chelator have shown that redox-active metal ions, such as ferrous (Fe⁺²), cuprous (Cu⁺²) and manganese (Mn⁺²) ions, account for approximately 1 ppm of trace metals in Biologic A drug product. It was obvious that trace metal analysis is inconclusive in determining the amounts of trace metals if their quantities are ≤ 1 ppm. Evidence from literature (9) suggests that such metals can be detrimental to the long-term stability of protein formulations, even at such low concentrations. Manufacturing experiences with a proprietary mAb at Bristol-Myers Squibb have also shown that low amounts of trace metals due to contamination of plant water can result in significantly lower long-term stability of this mAb. Based on these observations, metal-spiking studies were initiated with a final concentration of individual metal at 1 ppm to represent realistic pharmaceutical manufacturing conditions. The stability of the formulations was then evaluated using the methods described earlier.

Table I: Stability of “Biologic A” in the presence of metals (Fe²⁺, Mn²⁺, and Cu²⁺ at 1 parts per million [ppm]) and metal chelator (ethylenediaminetetraacetic acid [EDTA] or diethylene triamine pentaacetic acid [DTPA] at 50 mM) using size-exclusion chromatography. Control [C] is 10 mg/mL “Biologic A” in 10 mM pH 5.0 succinate buffer with 0.4% w/v poloxamer 188.

Effect of 50 mM metal chelator on  “Biologic A” after exposure to 40 °C.

Samples

% change in high-molecular-weight species after 4 weeks

% change in  high-molecular-weight species after 8 weeks

Control [C]

5.27

9.53

[C] + Fe²⁺

9.88

17.39

[C] +  Fe²⁺ + EDTA

8.42

14.71

[C] +  Fe²⁺ + DTPA

4.0

8.7

[C] +  Mn²⁺

15.64

27.51

[C] +  Mn²⁺ + EDTA

8.30

18.05

[C] +  Mn²⁺ +  DTPA

2.5

7.6

[C] +  Cu²⁺

14.52

24.48

[C] +  Cu²⁺ +  EDTA

4.01

7.90

[C] +  Cu²⁺ + DTPA

2.1

4.5

Spiking studies performed with Biologic A drug-product formulation (containing 10 mM succinate buffer, pH 5.0, 0.4% w/v poloxamer 188, with various transition metals at 1 ppm concentrations under accelerated conditions) have indicated that the presence of transition metals leads to significant drug-product instability, resulting in increase in the percentage of HMW species as shown in Tabel I. The data have been further confirmed by SDS-PAGE analysis under reducing (Figure 2) and non-reducing conditions (data not shown). It is apparent from SDS-PAGE (under reducing conditions after exposure to accelerated temperature of 40°C for 8 weeks) that the HMW species (Bands 1-3 in Figure 2b), is a mixture of dimers and higher-order species. SDS-PAGE under reducing conditions in Figure 2 also revealed that in addition to the HMW species, there is fragmentation of the intact mAb, evident by the appearance of bands 4 and 5 (Figure 2b) after 8 weeks at 40 °C. Fragmentation of the mAb could happen either by peptide-bond hydrolysis or disulfide-bridge cleavage in the hinge region of the mAb (17). Peptide bonds between aspartate-proline, asparatate-threonine, and histidine-lysine in the vicinity of the hinge region are prone to cleavage. Multiple factors, such as pH and trace metal impurities, could enhance the cleavage of these bonds (18). Importantly, the aggregates bands are not present in control samples. Similarly, the fragment bands are at a lower intensity in control sample, indicating that these observations are not purely temperature-dependent effects or aberrations based on SDS-PAGE sample preparation. Image density analysis of these bands (n=3) also proves that there is significant difference in the relative percentage area of these bands.

EDTA (50-500 μM) and DTPA (25-100 μM) were evaluated as potential metal chelators to be incorporated into the Biologic A drug-product formulation. Based on the preliminary metal-spiking studies, it was observed that Fe⁺², Cu⁺², and Mn⁺² were most detrimental to the stability of Biologic A at 1 ppm. Therefore, the effect of chelators (EDTA and DTPA) was evaluated in the presence and absence of these metals at a final concentration of 1 ppm in Biologic A drug-product solution prepared in 10 mM succinate buffer, pH 5.0 containing 0.4% w/v poloxamer 188. Experiments with chelators in the presence of metal spiking were essential not only to determine the effectiveness of individual chelator, but also for comparing the two chelators (i.e., EDTA and DTPA).

Based on the results shown in Table I for various sample prototypes, at elevated temperatures of 40ºC up to 8 weeks, it is evident that the presence of chelators is beneficial in decreasing protein aggregation (% HMW species is an indicator for protein stability). It is also evident that the affinity of EDTA and DTPA is different for various metals studied, as reflected in their differential effectiveness in preventing the formation of HMW species due to metal spiking. It was also observed that DTPA has significantly higher potential in preventing the formation of HMW species due to the presence of transition metals. As a result, further studies were performed with various concentrations of DTPA (50-100 μM) to determine the optimal concentration of DTPA to be incorporated in the final drug-product formulation (see Figure 3).

The enhancement in stability by DTPA compared with EDTA appears to be due to the higher-stability constant due to the additional carboxylate group of the metal-chelator complex that is formed in the presence of the respective chelator. Another aspect that seemed vital in the decision for the incorporation of a metal chelation came from the observed results in Table I and Figure 3. The control prototype had higher % HMW species at all time points when compared to all prototypes that were spiked with both metal and chelator, considering the fact that trace metal analysis of the control sample indicated the absence of any significant amount of redox-active trace metals. This observation gave a good indication that the chelators were effective in preventing the formation of HMW species due to residual and unquantifiable amounts of transition metals that were incorporated in the formulation due to excipients.

Results from Figure 3 indicate that the presence of higher concentration of DTPA did not correlate to an equivalent increase in protection from transition metals, and hence, a reduction in % HMW species. However, it appeared that the degree of protection obtained from higher concentration of DTPA is also dependent on the type of transition metal present in the formulation, as evident in the case of manganese chloride in which a higher concentration of DTPA offered better protection from the formation of HMW species. Based on this finding, a decision was made to incorporate 100 μM of DTPA in the final formulation.

The increase in thermal stability of protein solutions due to the presence of carboxylic acids, such as DTPA, has also been demonstrated earlier. The presence of succinic acid and EDTA in protein solutions enhanced the stability of these solutions due to the increase in surface free energy of the solvent medium (19). A separate study (see Table II) was conducted to review the effect of 100 μM DTPA in the formulation. It was observed that although the enhancement of stability is not as evident at 5°C, the presence of DTPA provides protection from trace-metal impurities. The final concentration of chelators to be included in the formulation needs to be evaluated carefully. One needs to keep in mind the allowable limits of these materials in different pharmacopeia to avoid challenges during regulatory filing in countries where the drug will be marketed.

Table II: Stability of “Biologic A” in the presence of 100 mM diethylene triamine pentaacetic acid (DTPA) at different temperatures and time points using size-exclusion chromatography. Control [C] is 10 mg/mL “Biologic A” in 10 mM pH 5.0 succinate buffer with 0.4% w/v poloxamer 188.

Effect of 100 mM metal chelator on “Biologic A”

Samples

% high-molecular-weight species after:

Initial

4 weeks 40°C

6 months
25°C

24 months
5°C

Control [C]

2.36

5.74

3.72

3.16

[C] + 100 µM DTPA

2.37

4.89

3.10

2.9

A new paradigm in biopharmaceutical industry is to provide the drug substance to drug-product manufacturing sites as a “formulated drug substance” that is ready to fill into drug-product vials at the commercial site after sterile filtration. The purpose is to enhance the stability of drug substance on shelf and also to minimize the number of processing steps at the manufacturing site, which is less equipped to handle complicated biological processing steps. Facilitating the preparation of “formulated drug substance” solutions leads to increasing processing times, and hence, potential for protein degradation at the drug-substance manufacturing site. One such degradation is the development of color in mAb solutions (20) during their biosynthesis, which was primarily dependent on the length of processing time and holding in batching vessels before affinity purification and protein isolation. This degradation is probably due to the formation of indole derivatives of tryptophan, such as 5-hydroxy tryptophan or N-formylkynurenine (see Figure 4) that are oxidative degradants of tryptophan, which give the solutions a yellowish brown hue. Presence of a metal chelator also had significant improvement in the appearance of the “formulated drug substance” of Biologic A. These solutions were clear to colorless, unlike the drug substance solutions without chelator, which appeared brown after six months at recommended storage conditions of 2 °C-8 °C.

CONCLUSION
This study demonstrated that even in the presence of tight controls with stringent specification on the amount of metal impurities from all the process components and the process itself, prevention of oxidative
degradation of proteins due to residual metals can be challenging. To maintain the stability of the drug product through its lifetime and account for any unforeseeable process changes, it is beneficial to incorporate minute quantities of metal chelators in protein formulations.

Acknowledgments
The authors wish to acknowledge Venkatramana M. Rao and Rajesh Gandhi for their helpful discussions.

References
1. E.R. Stadtman, Free Radic. Biol. Med. 9 (4) 315-325 (1990).
2. M.C. Manning, K. Patel, and R.T. Borchardt, Pharm. Res. 6 (11) 903-918 (1989).
3. J.L. Cleland, M.F. Powell, S.J. Shire, Crit. Rev. Ther. Drug. Carrier Syst. 10 (4) 307-377 (1993).
4. K.J. Davies, J Biol. Chem. 262 (20) 9895-9901 (1987).
5. K.J. Davies, M.E. Delsignore, and S.W. Lin, J Biol. Chem. 262 (20) 9902-9907 (1987).
6. K.J. Davies and M.E. Delsignore, J Biol. Chem. 262 (20) 9908-9913 (1987).
7. K.J. Davies and S.W. Lin, J Biol. Chem. 262 (20) 9914-9920 (1987).
8. X. Huang et al., J Biol. Inorg. Chem. 9 (8) 954-960 (2004).
9. S.W. Hovorka et al., J Pharm. Sci. 90 (1) 58-69 (2001).
10. E.R. Stadtman, and B.S. Berlett, Drug Metab. Rev. 30 (2) 225-243 (1998).
11. E.R. Stadtman, Annu Rev. Biochem. 62, 797-821 (1993).
12. V. Sadineni, N.A. Galeva, and C. Schoneich, Anal. Biochem. 358 (2) 208-215 (2006).
13. V. Sadineni and C. Schoneich, J Pharm. Sci. 96 (7) 1844-1847 (2007).
14. H. Pan et al., Protein Sci. 18 (2) 424-433 (2009).
15. D. Liu et al., Biochemistry 47 (18) 5088-5100 (2008).
16. S. Li et al., Biochemistry 34 (17) 5762-5772 (1995).
17. B.A. Salinas et al., J Pharm. Sci. 99 (7) 2962-2974 (2010).
18. G. Gaza-Bulseco and H. Liu, Pharm. Res. 25 (8) 1881-1890 (2008).
19. J.K. Kaushik and R. Bhat, Protein Sci. 8 (1) 222-233 (1999).
20. A.B. Magill, Abstracts of Papers, 236th ACS National Meeting, (Philadelphia, PA, 2008).

About the Author
Vikram Sadineni* is senior research investigator, Department of Drug Product Science & Technology R&D, Bristol-Myers Squibb, New Brunswick, NJ 08903; Sangita Chandrasekharan is global liaison senior manager, Novartis, Building No. 6, Mind Space, Hitech City, Hyderabad, Andhra Pradesh 500081, India; and Munir N. Nassar is an independent pharmaceutical development consultant.