Proteins removal by immunodepletion

Immunodepletion with a ProteoPrep^® 20 Plasma Immunodepletion Kit was selected to remove interfering proteins, based on various criteria, including: (i) the ability to remove interfering proteins, (ii) Oxyglobin^® recovery, (iii) CE and CE-MS compatibility, (iv) cost, and (v) process length. This kit is reported able to remove 20 of the most abundant proteins from human plasma or serum, via a high density conjugation process using small recombinant immunoaffinity ligands and conventional antibodies. The system was tested with blank plasma between Oxyglobin^® analyses and no carry-over was observed. The kit also permitted up to 2 loadings of the same plasma sample on the spin column to increase the degree of protein depletion. In this study, the samples treated with one vs. two loadings were compared on the basis of protein depletion efficiency (DE), Oxyglobin^® recovery (R1 and R2), and sample processing time.

Immunodepletion performance was evaluated by monitoring the 280 nm absorbance of blank plasma before and after immunodepletion. After one loading on the ProteoPrep column, 50%

of total proteins were removed. The second loading removed an additional 21% of proteins, resulting in a total DE of 71%.

Similarly, the ProteoPrep^® performance was evaluated in terms of Oxyglobin^® recovery, without (R1) or with (R2) the matrix, and process efficiency (PE) by UV/Vis absorbance at 415 nm (Figure 1). After one loading on the spin column, a PE of 72% was attained, along with R1 and R2 values of 82% and 75%, respectively. Overall, these results were satisfactory for the present application. After two loadings, R2 was found to be 60%, R1 was 66%, and

Acquisition Time (min) mAU

100%

0 4 8 12 16

A

B

C

Hb A₀

Oxyglobin^®

*

Oxyglobin^®

PE of 59% was observed. The preparatory procedure itself was thus the main source of Oxyglobin^® loss during the sample preparation (R1), because of adsorption of Oxyglobin^® on the cartridge medium and the Minicon^® system. On the other hand, the matrix had only a limited effect (6-7%, R2) on Oxyglobin^® recovery, regardless of the number of loadings on the spin column. Since a very low matrix effect was observed for UV/Vis measurements, PEs were found to be similar to R2s.

While a single loading on the immunodepletion column resulted in a 50% depletion of interfering proteins, with 75% recovery of Oxyglobin^®, the second loading increased the amount of proteins depletion (71%), but also lowered the Oxyglobin^® recovery to 60%. On the basis of these data, a single loading on the column was considered optimal, as it provided appropriate levels of depletion, without requiring any coating or post-washing steps.

Moreover, a single loading provided adequate Oxyglobin^® recovery, and reduced the overall processing time by 35%.

Figure 6 shows the electropherograms obtained following immunodepletion. In contrast to the “dilute and shoot” procedure, immunodepletion avoided adsorption problems.

Furthermore, the migration time repeatability was satisfactory (RSD <3%, N=5), and no capillary coating or post-washing procedures were required, resulting in reduced analysis time and greater MS-compatibility. The ghost peak next to the Oxyglobin^® peak remained present (Figure 6A) and was not found to be due to interfering proteins, as immunodepleted blank plasma did not exhibit such a peak (Figure 6B). Similarly, no peak appeared when an aqueous Oxyglobin^® standard was treated by immunodepletion (Figure 6C).

An important point concerned the removal of haptoglobins (Hp), because these proteins have a great affinity for circulating Hb. As previously mentioned, the presence of Hb in plasma samples could not be avoided due to potential hemolysis as a result of mechanical stress and/or sampling conditions. The Hb-Hp complex exhibited the same mobility as Oxyglobin^® under experimental conditions (Figure 6D), and demonstrated significant UV/Vis absorbance at 415 nm, due to the Hb component of the complex. The Hb-Hp complex also dissociated during the ionization process and individual Hb globin chains were detected. In case of Oxyglobin^® uptake, the discrimination was achieved because of the mass differences. False positive results could thus be excluded thanks to MS measurements. In presence of HBOCs made of human polymerized hemoglobin, no discrimination could be achieved by MS.

Therefore, Hb-Hp complex had to be removed to eliminate false positive results. Since the ProteoPrep^® 20 Plasma Immunodepletion Kit contains specific antibodies for human Hp, its potential for Hb-Hp complex removal was evaluated based on maximal circulating Hp concentrations (2360 mg·mL^-1). Because the matrix had a limited effect on immunodepletion performance, a water sample containing 2500 mg·mL^-1 of Hp and 1920 mg·mL^-1 of Hb was treated by the immunodepletion procedure (an Hp/Hb ratio of 1.3 was chosen on the basis of the manufacturer’s information). Because no peak was detected at 415 nm, it was concluded that the immunodepletion procedure was sufficiently effective at removing Hb-Hp complexes for the present application.

Figure 6: CE-UV electropherograms after immunodepletion procedure. Trace A: Oxyglobin^® (0.3 g·dL^-1) in plasma, *: ghost peak. Trace B: blank plasma. Trace C: Oxyglobin^® (0.2 g·dL^-1) in water. Trace D: Hb-Hp complex (0.192 and 0.250 g·dL^-1, respectively) in water. UV/Vis 415 nm online. See text for experimental conditions.

Thus, the immunodepletion sample procedure removed 50% of plasma proteins, while its gentle conditions allowed direct injection of processed samples into the CE system, and permitted Oxyglobin^® to be analyzed in its intact form. Furthermore, the degree of protein depletion permitted the repeated use of a bare-fused silica capillary, which was readily compatible with ESI-MS detection. A concentration range of 0.10 g·dL^-1 to 1.20 g·dL^-1 was tested and detection limits (estimated with a signal-to-noise ratio of 3) of 0.20 g·dL^-1 and 0.45 g·dL^-1 were achieved in plasma for CE-UV/Vis at 415 nm and CE-ESI-TOF/MS, respectively.

This methodology thus appears suitable for implementation as a doping control screening method for Oxyglobin^® analysis.

Acquisition Time (min) mAU

100%

0 4 8

A

B

C

D Oxyglobin^®

Oxyglobin^®

*

Hb-Hp

4. CONCLUDING REMARKS

In the context of doping analysis, the ability to detect HBOC doping agents such as Oxyglobin^® is critical. In this study, CE-UV/Vis and CE-ESI-TOF/MS methods were developed for the analysis of Oxyglobin^® in plasma. Optimal CE conditions were obtained at basic pH, resulting in full resolution of Hb and Oxyglobin^® peaks without the need for time-consuming capillary coating and washing procedures. Additional selectivity was gained from ESI-TOF/MS detection due to protein dissociation in the ESI source, which resulted in the ability to distinguish bovine and human hemoglobin by the differences in their monomeric chain molecular weights. Thus, the differentiation of Oxyglobin^®, a polymerized form of bovine hemoglobin, from human hemoglobin was possible. UV/Vis absorbance at 415 nm offered another selectivity level, since only hemoproteins absorb at this wavelength. Sample preparation was identified as a key aspect of the method: removal of plasma proteins by immunodepletion was necessary to reduce adsorption onto the capillary walls and eliminate Oxyglobin^® peak degradation and signal suppression. This procedure also achieved excellent haptoglobin removal, thus eliminated Hb-Hp complexes with the potential to interfere with Oxyglobin^® detection.

5. ACKNOWLEDGMENTS

The authors wish to thank the Swiss Federal Council of Sports (Magglingen, Switzerland) and Dr Matthias Kamber, Director of Antidoping Switzerland (Bern, Switzerland) for financial support. The World Anti-Doping Agency (Montreal, Canada) is also acknowledged for supporting the "forensic approach of fight against doping" project.

The authors also wish to thank Dr Neil Robinson and Dr Lidia Avois-Mateus for helpful and valuable comments.

6. ABBREVIATIONS

DE depletion efficiency

HBOC hemoglobin-based oxygen carrier

Hb hemoglobin

Hp haptoglobin

PE process efficiency

R1 recovery without matrix

R2 recovery with matrix

7. REFERENCES

[1] Thevis, M., Schänzer, W., Mass Spectrometry Reviews 2007, 26, 79-107.

[2] Chang, T.M.S., Blood substitutes: Principles, Methods, Products and Clinical Trials, Karger Landes Systems, Basel 1997, pp. 9-18.

[3] Thevis, M., Ogorzalek Loo, R.R., Loo, J.A., Schänzer, W., Anal. Chem. 2003, 75, 3287-3293.

[4] Audran, M., Varlet-Marie, E., Sciences & Sports, 2004, 19, 1-7.

[5] Buehler, P.W., Alayash, A.I., Biochimica et Biophysica Acta 2008, 1784, 1378-1381.

[6] Lasne, F., Crepin, N., Ashenden, M., Audran, M., De Ceaurriz, J., Clin. Chem. 2004, 50, 410-415.

[7] Varlet-Marie, E., Ashenden, M., Lasne, F., Sicart, M.-T., Marion, B., De Ceaurriz, J., Audran, M., Clin. Chem. 2004, 50, 723-731.

[8] Simitsek, P.D., Giannikopoulou, P., Katsoulas, H., Sianos, E., Tsoupras, G., Spyridaki, M.-H., Georgakopoulos, C., Analytica Chimica Acta 2007, 583, 223-230.

[9] Sahin, A., Laleli, Y.R., Ortancil, R., J. Chromatogr. A 1995, 709, 121-125.

[10] Shihabi, Z.K., J. Chromatogr A 2004, 1027, 179-184.

[11] Cao, P, Moini, M., J Am Soc Mass Spectrom 1999, 10, 184-186.

[12] Boys, B.L., Konermann, L., J Am Soc Mass Spectrom 2007, 18, 8-16.

[13] Schappler, J., Veuthey, J. L., Rudaz, S., Coupling CE and Microchip-Based Devices With Mass Spectrometry, Academic Press, San Diego 2008, pp. 477-521.

[14] Staub, A., Schappler, J., Rudaz, S., Veuthey, J.-L., Electrophoresis 2009, 30, 1610-1623.

[15] Tirumalai, R.S., Chan, K.C., Prieto, D.A., Issaq, H.J., Conrads, T.P., Veenstra, T., Mol. Cell. Proteomics 2003, 2.10, 1096-1103.

[16] Kim, M.-R., Kim, C.-W., J. Chromatogr. B 2007, 849, 203-210.

[17] Marchi, I., Rudaz, S., Veuthey, J.-L., J. Pharm. Biomed. Anal 2009, 49, 459-467.

[18] Telford, R.D., Sly, G.J., Hahn, A.G., Cunningham, R.B., Bryant, C., Smith, J.A., J.

Appl. Physiol. 2003, 94, 38-42.

[19] Grant, M.S., J. Emerg. Nurs. 2003, 29, 116-121.

[20] Peng, J., Mandal, R., Sawyer, M., Li, X.-F., Clin. Chem., 2005, 51, 2274-2281.

[21] Salplachta, J., Rehulka, P., Chmelik, J., J. Mass Spectrom. 2004, 39, 1395-1401.

[22] Troxler, H., Neuheiser, F., Kleinert, P., Kuster, T., Heizmann, C.W., Sack, R., Hunziker, P., Neuhaus, T.J., Schmid, M., Frischknecht, H., Biomed. Biophys. Res.

Comm. 2002, 292, 1044-1047.

I I . 6 I d e n t i f i c a t i o n e t q u a n t i f i c a t i o n d e f o r m u l a t i o n s d ’ i n s u l i n e à l ’ a i d e d e l a t e c h n i q u e d ’ i n j e c t i o n s m u l t i p l e s e t p a r C E - TO F / M S

Dans un contexte de contrôle qualité de formulations pharmaceutiques, il peut être intéressant d’avoir une méthode analytique permettant une quantification et une identification simultanées. Pour les deux aspects, le TOF apporte une solution. L’identification est rendue possible grâce à l’exactitude de masse de cet analyseur et par comparaison des ions majoritaires multichargés du produit à identifier avec ceux d’un standard de référence.

L’aspect quantification paraît aussi possible en utilisant ces mêmes ions majoritaires, mais le problème du choix d’un standard interne reste posé.

Les standards internes les plus employés dans le domaine de la CE-MS sont ceux marqués isotopiquement ainsi que les analogues de structure. Dans le cas des protéines, leur emploi est compliqué car ces derniers sont difficiles à obtenir et/ou très onéreux. L’article VI propose une technique alternative de quantification basée sur l’injection multiple. Pour cela, un lot de la formulation de protéine à quantifier est choisi comme matériel de référence et injecté en premier dans le système. Une seconde injection est ensuite effectuée lors du même run avec la formulation à identifier et quantifier. Les tests de faisabilité ont été effectués avec l’insuline comme protéine modèle. L’identification est faite sur les spectres de masse extraits du pic du standard et de la protéine « inconnue ». Par comparaison des masses exactes des ions majoritaires, l’identité est confirmée ou infirmée. La quantification est effectuée sur l’électrophérogramme ionique extrait en faisant le rapport de l’aire du pic de la protéine sur l’aire du pic du standard. Il s’est avéré que la correction d’ionisation n’était pas suffisante (CV>8%). Ainsi, un standard d’injection a été ajouté aux deux échantillons protéiques (standard et « inconnu »). L’aire de ces standards a été acquise via la détection UV en ligne et les aires des protéines obtenues en TOF corrigées par les aires respectives de leur standard d’injection. La combinaison des deux corrections a permis de diminuer le CV à moins de 2% et la méthode de quantification a ensuite pu être validée. Les performances quantitatives sont satisfaisantes et le profil d’exactitude est compris dans les limites de spécification de ± 5% autour de la valeur cible de 100% exigée pour les formulations pharmaceutiques. Finalement, quatre lots de formulations d’insuline ont été analysés avec la méthode. Trois provenaient du marché régulier tandis qu’un autre a été obtenu sur internet sans ordonnance. Les quatre lots ont donné des résultats conformes tant en termes d’identité que de concentration.

Par la suite, cette approche méthodologique devra être confrontée à des protéines plus complexes et testées pour le contrôle des contrefaçons.

II.6.1 Article VI

A. Staub, S. Rudaz, J.L. Veuthey, J. Schappler. Multiple injection technique for the determination and quantitation of insulin formulations by capillary electrophoresis and time-of-flight mass spectrometry, J. Chromatogr. A 2010, 1217, 8041-8047.

Multiple injection technique for the determination and quantitation of insulin formulations by capillary electrophoresis and time-of-flight mass spectrometry

Aline Staub^1,2,, Serge Rudaz^1,2,,Jean-Luc Veuthey^1,2,, Julie Schappler^1,2,*

1School of pharmaceutical sciences, University of Geneva, University of Lausanne, Bd d’Yvoy 20, 1211 Geneva 4, Switzerland

2Swiss Centre for Applied Human Toxicology (SCAHT), University of Geneva, CMU, Rue Michel-Servet 1, 1211 Geneva 4, Switzerland

Journal of Chromatography A 2010, 1217, 8041-8047.

ABSTRACT

This paper describes an efficient CE-UV-ESI-TOF/MS method for the determination and quantitation of intact insulin (INS) in a pharmaceutical formulation. The CE conditions were optimized to avoid the adsorption of proteins onto the capillary wall. Particular attention was paid regarding the choice of the internal standard (IS). A strategy based on multiple injections was selected and the methodology was validated according to international guidelines. The optimized method was applied with success to the analysis of INS formulations obtained from regular and parallel markets.

KEYWORDS

Capillary electrophoresis, intact protein, multiple injection, quantitation, time-of-flight mass spectrometry

1. INTRODUCTION

In the pharmaceutical area, recombinant proteins produced by biotechnology have grown considerably with a market evaluated at over $70 billion per year by a 2010 estimate [1].

These proteins comprise antibodies, hormones, biological response modifiers to stimulate cell growth, enzymes, and vaccines [2,3]. During the biopharmaceutical development process, several parameters are needed for regulatory purposes, regarding the identity, quantity (concentration), quality, and purity of the products [4,5]. Determining the identity and concentration of the therapeutic proteins is also important after their release on the market from a quality control perspective. Since unofficial channels exist to obtain these products without prescription or without extensive evidence for quality control, therapeutic protein analysis is also relevant for the parallel market. Moreover, biopharmaceuticals available in the parallel market can be counterfeit drugs. These include products with or without the correct ingredients, without the active ingredients, with insufficient or too much active ingredients, or with fake packaging [6]. Consequently, from public health perspective, it is essential to develop analytical methods to quickly monitor the identity, quality, and quantity of these biopharmaceuticals.

For the identification and quantitation of protein formulations, the analysis of proteins in their intact form is a promising approach because no tedious sample preparation, such as a digestion step, is required. Various methods for determining the quantity of intact protein exist, and the choice of the assay mainly depends on parameters such as the quantity of protein available or the required throughput [7]. Commonly, these assays are based on UV-VIS spectroscopy (e.g., UV absorbance at 280 nm, Bradford protein assay, Lowry assay) or fluorescent detection after derivatization with a fluorescent probe (e.g., fluorescamine, 3-(4-carboxybenzoyl)quinoline- 2-carboxaldehyde) [7,8]. The lack of specificity is the main bottleneck of these assays. In the contrary, mass spectrometry (MS) allows for a higher level of selectivity; often, confirmation of the product’s identity is obtained through the accurate determination of its molecular mass, when high resolution mass analyzers are used [4]. To perform the simultaneous identification and quantitation of the active protein in its intact form, whether in a pharmaceutical formulation or in another matrix, it is necessary to couple a separation technique to an appropriate detector. Therefore, the hyphenation of capillary electrophoresis (CE) and MS via an electrospray ionization (ESI) source is an attractive option [9,10]. CE offers high speed, great efficiency, and low solvent and sample consumptions, while MS provides selectivity, sensitivity, and specificity.Due to its high mass range and mass accuracy, the time-of-flight (TOF) analyzer is particularly well suited for the detection of intact proteins that are multi-charged as a result of ESI [11].

Capillary zone electrophoresis (CZE) is widely used given its versatility and compatibility with ESI-MS. However, the analysis of proteins by CZE is often impaired by the tendency of the proteins to adsorb onto the negatively charged surface of fused silica (FS) capillaries [12,13], thus degrading CE performance. The evaluation of protein adsorption and its prevention must be considered during the analytical method development, particularly when accurate quantitation is attempted. The choice of the internal standard (IS) is an important point to consider in quantitative analysis. Even if matrix effects seem negligible in the case of pharmaceutical formulations analysis involving good separation of active ingredient(s) and

excipient(s), stable isotopically labeled (SIL) compounds and structural analogues remain the gold standards. However, for intact proteins, SIL compounds are not commonly available and/or could be very expensive [14,15]. Structural analogues differ from the intact protein by an exchange or removal/addition of amino acids, or a small modification in one or more side chains. These analogues are not easy to obtain for all proteins, can be expensive, and may present a different ionization behavior than that of the protein of interest. An alternative methodology to the IS concept was adapted from the multiple injection technique [16].

Initially developed to reduce the analysis time, this technique could be used to overcome the lack of satisfactory IS for intact proteins. In this approach, two injections are performed in the same analytical run, the first one with a standard of the protein of interest at a known concentration and the second one with the protein to be quantified. Therefore, the IS would be a standard of the protein, considered as the reference material.

In this study, a CE-UV-ESI-TOF/MS method was developed for the analysis of a recombinant human insulin (INS) as a model protein. INS was selected because of the numerous pharmaceutical formulations available on the market. Furthermore, since 1999, INS has been prohibited in sports for athletes who do not suffer from diabetes mellitus [17]. In addition, patients with this chronic disease often buy INS online without prescription because of the potentially lower cost. Due to these misuses, the risk of finding counterfeit drugs on the parallel market has increased dramatically. In the context of public health, analytical methods for quality control of these pharmaceutical formulations are needed. INS was already analyzed by CE [18,19,20], also coupled with MS detection [21], but never with identification and quantitation by MS. Quantitation was here attempted, using a multiple injection technique based on the successive injection of a reference standard of INS and the sample in one single run. The complete methodology was fully validated according to the guidelines of the International Conference of Harmonization (ICH) and applied to pharmaceutical formulations obtained in pharmacies and on the web without a formal prescription.

2. MATERIAL AND METHODS