innovative strategies for stabilization formulations · rijksuniversiteit groningen innovative...

InnovatIve strategIes for stabIlIzatIon of therapeutIc peptIdes In aqueous

formulatIons

Christina Avanti

The research presented in this thesis was performed within the framework of project D6-202 of the Dutch Top Institute Pharma.

Paranimfen: Milica Stankovic Leviny Zachreina

Printing of this thesis was supported by generous contributions from: University of Groningen Faculty of Mathematics and Natural Sciences of the University of Groningen

ISBN: 978-94-6182-122-5

© Copyright 2012 C. Avanti

All rights reserved. No part of this thesis may be reproduced, stored in a retrieval system or transmitted in any form or by any means, mechanically, by photocopying, recording or otherwise, without the written permission of the author.

Cover design: Bao Tung Pham and Christina Avanti

Layout & printing: Off Page, Amsterdam

RIJKSUNIVERSITEIT GRONINGEN

InnovatIve strategIes for stabIlIzatIon of therapeutIc peptIdes In aqueous

formulatIons

Proefschrift

ter verkrijging van het doctoraat in deWiskunde en Natuurwetenschappenaan de Rijksuniversiteit Groningen

op gezag van deRector Magnificus, dr. E. Sterken,in het openbaar te verdedigen op

maandag 2 juli 2012om 16.15 uur

door

Christina Avanti

geboren op 3 april 1968 Kota Baru, Indonesië

Promotor: Prof. dr. H.W. Frijlink

Copromotor: Dr. W.L.J. Hinrichs

Beoordelingscommissie: Prof. dr. G.M.M. Groothuis Prof. dr. A.J.M. Driessen Prof. dr. W. Jiskoot

contents

PART ONE INTRODUCTION

Chapter 1 General Introduction 9

Chapter 2 Current Strategies for Stabilization of Therapeutic Peptides in Aqueous Formulations 13

PART TWO THE USE OF DIVALENT METAL IONS AND CITRATE BUFFERS TO STABILIZE OXYTOCIN IN AQUEOUS SOLUTION

Chapter 3 A New Strategy to Stabilize Oxytocin in Aqueous Solutions : I. The Effects of Divalent Metal Ions and Citrate Buffer 35

Chapter 4 A New Strategy To Stabilize Oxytocin in Aqueous Solutions: II. Suppression of Cysteine-Mediated Intermolecular Reactions by a Combination of Divalent Metal Ions and Citrate 51

PART THREE THE USE OF DIVALENT METAL IONS AND ASPARTATE BUFFER TO STABILIZE OXYTOCIN IN AQUEOUS SOLUTION

Chapter 5 Insight into the Stability of the Zinc-Aspartate-Oxytocin Complex 79

Chapter 6 Aspartate buffer and divalent metal ions affect the oxytocin conformation in aqueous solution and protect it from degradation 95

PART FOUR THE USE OF EXTREMOLYTES TO STABILIZE PROTEIN IN AQUEOUS SOLUTION

Chapter 7 Extremolytes: Are There Universal Stabilizers for Proteins in Aqueous Solution? 121

Summary, Concluding Remarks, and Global Perspective 137

Samenvatting, Conclusies, Aanbevelingen, en Mondiaal Perspectief 145

Acknowledgments 153

for all the women in existence… for a better chance to live the life with the new born …

Every year 166 000 women die of bleeding after child birth, and more than 50% of these deaths occur in sub-Saharan Africa

(Clyburn et. al., 2007)

1

general IntroductIon

1 general IntroductIonAlthough successful developments in the field of peptide synthesis increased the availability of peptide drugs [1], a significant part of the world’s population is still facing serious problems associated with insufficient access to several essential drugs which are peptide in nature. The major cause for this problem is the lack of stable formulations that withstand transport, storage and distribution, particularly in rural and remote areas of developing (tropical) countries that lack a cold chain.

An important active pharmaceutical peptide is oxytocin, a nonapeptide which is the drug of first choice to treat bleeding after childbirth or post-partum hemorrhage (PPH) [2]. Although the availability of this drug has greatly declined maternal mortality rates in the developed world, PPH remains a leading cause of maternal mortality elsewhere [3]. Current commercial formulations of oxytocin are insufficiently heat-stable to withstand tropical conditions [4,5]. Cleavage of the disulfide bridge was found to be the major degradation pathway [6,7].

The aim of the present thesis was to develop heat-stable formulations for polypeptide drugs, in particular oxytocin and investigate the mechanism of stabilization based on degradation products formed during thermal stress before and after formulation.

This thesis is divided in four parts. In part 1 (Chapter 2) a general introduction is given on the stability of peptide drugs in aqueous solution and current stabilization technologies. Part 2 (Chapter 3 and 4) describes the use of a combination of divalent metal ions and citrate buffer to improve stability of oxytocin in aqueous solution, whereas part 3 (Chapter 5 and 6) is about the effect of zinc ions to improve stability of oxytocin in aspartate buffer. Finally, part 4 (Chapter 7) describes an attempt to find the universal stabilizer for polypeptides by investigating various extremolytes using lysozyme and insulin as model peptides.

In Chapter 2, we review the most common degradation pathways of peptides, such as hydrolysis, deamidation, isomerization, racemization, oxidation, disulfide exchange, dimerization, and further aggregation that cause loss of potency of pharmaceutical peptides. In this chapter we also review different strategies to overcome these instabilities, such as pH optimization, the use of buffers, antioxidant and other additives.

In Chapter 3, we describe an investigation on the effect of monovalent (Na+ and K+) and divalent (Ca2+, Mg2+, and Zn2+) metal ions in combination with citrate and acetate buffers at pH of 4.5 on the stability of oxytocin in aqueous solution. The effect of combinations of buffers and metal ions on the stability of aqueous oxytocin solutions was determined by reversed-phase high performance liquid chromatography (RP-HPLC) and size exclusion chromatography (HP-SEC) after 4 weeks of storage at either 4°C or 55°C. We also measured the interaction between oxytocin and Ca2+, Mg2+, or Zn2+ in citrate buffer in comparison with acetate buffer by using isothermal titration calorimetry (ITC).

In Chapter 4, we identified various degradation products of oxytocin in citrate-buffered solution after thermal stress at a temperature of 70 °C for 5 days and the differences in degradation pattern in the presence and absence of divalent metal ions. Degradation products of oxytocin in the citrate buffer formulation with and without divalent metal ions were analyzed using liquid chromatography−mass spectrometry/mass spectrometry (LC−MS/MS).

In Chapter 5, we investigated the effects of various metal ions (Ca2+, Mg2+ and Zn2+) on the stability of oxytocin in aspartate buffer pH 4.5 and determined their interaction with the peptide in aqueous solution. The effect of combinations of various metal ions on the stability of oxytocin in aspartate buffer solutions was determined by RP-HPLC and HP-SEC

10

1

General IntroductIon

after 4 weeks of storage at either 4°C or 55°C. We also investigated which degradation products of oxytocin were formed in the aspartate buffer formulation with and without divalent metal ions using LC−MS/MS and determined the interaction between oxytocin and Ca2+, Mg2+, or Zn2+ in aspartate buffer by using ITC.

In Chapter 6, we further explored the mechanism of stabilization of oxytocin by the combination of Zn2+ and aspartate buffer. Therefore, we investigated the conformation of oxytocin in aspartate buffer in the presence of Zn2+ in comparison with Mg2+ using 2D NMR spectroscopy, i.e. NOESY, TOCSY, 1H-13C HSQC and 1H-15N HSQC with neither 13C nor 15N enrichment.

In Chapter 7, we describe a study on the effects of extremolytes on the stabilization of two model proteins (the larger peptides) lysozyme and insulin in aqueous solutions. The effects of different extremolytes (betaine, hydroxyectoine, trehalose, ectoine, and firoin) on the stability of lysozyme were determined by Nile red Fluorescence Spectroscopy and a bioactivity assay. Insulin stability was determined by RP-HPLC and HP-SEC. The effects of extremolytes on the unfolding temperature of the proteins were analyzed using a thermal shift assay for lysozyme and liquid differential scanning microcalorimetry for insulin. The interaction between extremolytes and protein was studied by isothermal titration calorimetry (ITC).

references1. B.A. Moss, Peptide Synthesis, in: S. Frokjaer,

L. Hovgaard (Eds.), Pharmaceutical Formulation Development of Peptides and Proteins, CRC Press, United States of America, (2000) 1-12.

2. R.A. Dyer, D. van Dyk, A. Dresner, The use of uterotonic drugs during caesarean section, Int. J. Obstet. Anesth. 19 (2010) 313-19.

3. P. Clyburn, S. Morris, J. Hall, Anaesthesia and safe motherhood, Anaesthesia 62 (2007) 21-5.

4. H.V. Hogerzeil, G.J.A. Walker, M.J. De Goeje, Stability of injectable ocytocics in tropical climates, World Health Organization, Geneva WHO/DAP/93.6. (1993).

5. A.N.J.A.D. Groot, T.B. Vree, H.V. Hogerzeil, G.J.A. Walker, WHO Action Programme on Essential Drugs: Stability of Oral Oxytocics

in Tropical Climates : Results of Simulation Studies on Oral Ergometrine, Oral Methylergometrine, Buccal Oxytocin and Buccal Desamino-Oxytocin, World Health Organization, Geneva, (1994).

6. A. Hawe, R. Poole, S. Romeijn, P. Kasper, R. van der Heijden, W. Jiskoot, Towards heat-stable oxytocin formulations: analysis of degradation kinetics and identification of degradation products, Pharm. Res. 26 (2009) 1679-88.

7. C. Avanti, H.P. Permentier, A.V. Dam, R. Poole, W. Jiskoot, H.W. Frijlink, W.L.J. Hinrichs, A new strategy to stabilize oxytocin in aqueous solutions: II. Suppression of cysteine-mediated intermolecular reactions by a combination of divalent metal ions and citrate, Mol. Pharmaceutics 9 (2012) 554-62.

11

Christina Avanti1, Wim Jiskoot2, Wouter L. J. Hinrichs1, and Henderik W. Frijlink1

1 Department of Pharmaceutical Technology and Biopharmacy, University of Groningen, Groningen, The Netherlands

2 Division of Drug Delivery Technology, Leiden/Amsterdam Center for Drug Research, Leiden University, Leiden, The Netherlands

current strategIes for stabIlIzatIon of therapeutIc peptIdes In aqueous

formulatIons

2

2

abstractParenteral administration is one of the most used routes to obtain systemic delivery of peptide drugs. Although most of therapeutic peptides are formulated in the dry powder for reconstitution, from economic point of view aqueous liquid formulations are preferred. However, therapeutic peptides are often unstable in the aqueous formulation of the injection. This review will focus on the formulation strategies that can be applied to stabilize therapeutic peptides in aqueous solution. We have organized this review as follows: first the main peptide stability problems in solution are described followed by a discussion on the known strategies to reduce peptide degradation, including recent research on improving peptide stability in liquid formulations.

2

StabIlIzatIon of therapeutIc peptIdeS

1. IntroductIonThere has been a rapid expansion in the use of peptides as potential drugs since the successful chemical synthesis of oxytocin by duVigneaud in 1953 [1]. The following decades witnessed the discovery of a large number of peptides as active pharmaceutical ingredients, and this process is likely to continue in the future. Currently over a hundred peptide drug candidates are in development for a wide variety of diseases such as several forms of cancer and metabolic diseases [2]. Examples of several peptides in the Dutch market are described in Table 1.

Peptides differ from proteins in that they are smaller and typically lack a defined tertiary structure, but a clear distinction is difficult to make and several arbitrary definitions exist. For instance, Malavolta [3] defined peptides as molecules containing fewer than 40 amino acid residues, while proteins contain 50 residues or more, with in between a category called polypeptides (40-49 residues). In another source, amino acids joined together in chains of 50 amino acids or less are defined as peptides, 50-100 amino acids are defined as polypeptides, and over 100 amino acids are defined as proteins [4]. Although also other definitions of the term peptide have been proposed, in this review we will focus on peptides containing less than 50 amino acid residues. Within this definition peptides can range in size from dimers containing two (modified) amino acid residues, such as enalapril and lisinopril [5], through small peptides, such as oxytocin [6] and octreotide [7] containing fewer than 10 residues, to relatively large polypeptides, such as calcitonin (32 residues) [8].

A number of hormones, enzymes, antitumor agents, antibiotics and neurotransmitters are peptides. Peptides regulate many physiological processes, acting at some sites as endocrine or paracrine signals and at other sites as neurotransmitters or growth factors. Nowadays, peptides are used as therapeutic agents in diverse disease areas such as neurological, endocrinological and hematological disorders [9].

Therapeutic peptides pose a number of challenges for pharmaceutical scientists regarding their formulation and delivery. The sensitivity of many peptides to enzymatic breakdown (e.g. in the gastro-intestinal tract) [10] and their poor ability to pass absorbing membranes typically results in a poor bioavailability following non-parenteral administration [11]. Moreover, the lack of physical and chemical stability may lead to significant degradation during processing and storage of the (aqueous) formulations. Enalapril and lisinopril do not have the typical delivery issues and oral delivery is well possible. Therefore we do not discuss those modified dipeptides in this review.

The lack of oral efficacy of peptides has prompted the examination of other non-invasive routes for peptide drug administration [4,12]. These routes include buccal [13-15], rectal [16,17], vaginal [18,19], percutaneous [20], ocular [21,22], transdermal [23,24], nasal [25-27] and the pulmonary route [28,29]. However, many of these routes of administration are still under investigation and they may be insufficiently efficient, especially when a rapid effect is desired. Therefore, for delivery of therapeutic peptides formulations the invasive parenteral routes are still often preferred [30]. Intravenous administration is the most efficient way to deliver peptide drugs directly into the systemic circulation, since no absorption process is involved. Intramuscular routes can also be used, however, the absorption is slower than after intravenous administration. Both intravenous and intramuscular routes do not easily allow self-administration and patient experiences pain after injection. The subcutaneous route can be employed when a more sustained action is acceptable or required and this route is more suitable for self-administration.

15

2

Table 1 Several therapeutic peptide-based products in the Dutch market in 2010-2011

Generic names

Trade names Principal

Dosage form

Self-life and storage cond.

pH(adj. agent)

Thyrotropin Releasing Hormones

ProtirelinRelefact TRH® Sanovi-Aventis Liquid inj. 3 yrs at 15-25°C 6.5 (Phosphate)

Antibiotic peptides

Daptomycin Cubicin® Novartis Powder for inj. 3 yrs at 2-8°C 4-5 (NaOH)

Teicoplanin Targocid® Sanofi-Aventis Powder for inj. 4 yrs at 2-8°C 3.8 to 6.5

Platelet aggregates inhibitors

Eptifibatide Integrilin® GlaxoSmithKline Liquid for infusion 3 yrs at 2-8°C 5.35 (citrate)

Somatostatin analog

Octreotide Sandostatin® Novartis Liquid inj. 3 yrs at 2-8°C 4.2 (lact/carb.)Vasopressins and analog

Desmopressin DDAVP® Ferring Liquid inj. 4 yrs at 2-8°C 4-5

Octostim® Ferring Liquid inj. 4 yrs at 2-8°C 4 (HCl)

Minrin® Ferring Liquid inj. 4 yrs at 2-8°C 4 (HCl)

FelypressinCitanest 3%

Octapressin® Densply Liquid inj. 3 yrs at 15-25°C 3.2

Oxytocins

Oxytocin Syntocinon® Defiante Liquid inj. 4 yrs at 2-8°C 4 (acetate)

carbetocin Pabal® Ferring Liquid inj. 2 yrs 2-8°C 3.8 (acetate)Oxytocin antagonist

Atosiban Tractocile®® Ferring Liquid inj. 4yrs at 2-8°C 4.5 (HCl)Gonadotropin Releasing Hormone (GNRH)/Luteinizing Releasing Hormone (LHRH) agonists

Goserelin Zoladex® Div Liquid inj. 3 yrs < 25°C

GonadorelinRelefact LH-RH® Sanofi-Aventis Liquid inj. 15-25°C

TriptorelinDecapeptyl

-CR® Ipsen

Powder and solvent for solution

for inj.

3 yrs at 2-8°C

nafarelin Synarel® Pfizer Liquid nasal spray 2 yrs < 25*C 5-7 (acetate)

Leuprolide Eligard ® Sanofi-Aventis


for inj.

2 yrs at 2-8°C

Cetrorelix Cetrotide® Serono Powder for inj. 3 yrs at 15-25°C

Source: http://www.geneesmiddelenrepertorium.nl/repertoriumcontinued next page

16

2


Table 1 continued Several therapeutic peptide-based products in the Dutch market in 2010-2011

Generic names

Trade names Principal

Dosage form

Self-life and storage cond.

pH (adj. agent)

Non-Steroidal Anti-Inflammatory Drugs

Ziconotide Prialt® Elan Liquid inj. 3 yrs at 2-8°C 4-5 (HCl/NaOH)Calcitonins

Salmon Calcitonin

Calcitonin Sandoz® Novartis Liquid inj. 5 yrs at 2-8°C 3.3-3.7

Human Parathyroid Hormone [hPTH (1–34)

Teriparatide Forsteo® Eli Lily Liquid inj. 2 yrs at 2-8°C 4 (acetate)Fusion inhibitor of Human Immunodeficiency Virus (HIV) with Cluster Difference 4 (CD4) cells

Enfuvirtide Fuzeon® Roche


for inj.

4 yrs at 2-8°C 9-9.5 (carbonate)

Adrenocorticotropin Hormone (ACTH) and derivatives

CorticorelinCRH -

Ferring® Ferring Powder for inj. 3 yrs < 25°C

Source: http://www.geneesmiddelenrepertorium.nl/repertorium

Because of their instability, most peptide drugs have to be stored and transported at low temperatures. This has a big impact on the access to pharmaceutical active peptides, particularly in rural and tropical areas that lack a cold chain [31,32]. Therefore, an urgent need exists for new strategies to tackle the problems associated with peptide instability, especially for aqueous injectable formulations which are preferred over freeze-dried formulations. Although freeze-dried formulations seem an ideal strategy to maintain the peptide´s structural integrity [9], unfortunately, this strategy is economically not ideal. Particularly for developing countries, the lyophilization process may be too expensive. Production costs are high and the products that have to be stored and transported have a volume and mass up to twice as large as that of a liquid formulation, since the storage and transport includes both the vials containing the lyophilized powder and sterile water for reconstitution. Therefore, liquid formulations are preferred.

Stability of peptide drugs in liquid formulations is the most important factor to be considered in the design and development of liquid parenteral peptide formulations. Table 1 displays the shelf-life and storage conditions of the various marketed peptides, showing that the majority of the formulations is to be stored under refrigerated conditions. This is a clear indication for the poor stability of many products. Loss of potency of a peptide commonly finds its origin in physical degradation processes such as adsorption [33], aggregation and chemical degradation pathways (e.g. hydrolysis, oxidation, deamidation [34]. This process is not only specific for small peptide or protein but for peptides in general. The chemical instability of peptides poses a problem in their development as active pharmaceutical ingredients. Therefore, a better understanding of the underlying mechanisms of instability of a certain peptide is essential to design rational strategies that can be investigated during the pharmaceutical development process in order to optimize the stability of the peptide in the final formulation [9].

17

2

2. peptIde InstabIlIty and the possIble causes of degradatIonDuring formulation, processing, storage and transportation of a peptide drug product, the peptide is exposed to conditions that can have significant effects on its chemical and physical integrity and many peptides have to be transported and stored under a “cold chain” regime [35]. There are several degradation pathways peptides may undergo when they are formulated as an aqueous solution. The major pathways of peptide degradation are broadly divided into: chemical and physical instability.

Chemical instability can be defined as any process involving modification of the peptide by covalent bond formation or covalent bond cleavage, generating new chemical entities [36]. Chemical instability pathways include hydrolysis, oxidation, racemization, isomerization, deamidation, disulfide exchange and β-elimination [37]. Physical instability refers to any changes to the higher order structure (dimerization and further aggregation), i.e. changes in noncovalent bonds, not necessarily involving covalent bond modifications. This physical instability can result in denaturation which may lead to adsorption to surfaces, aggregation and precipitation [9]. In Table 2, reported degradation pathways of various peptides are presented.

2.1 Hydrolytic pathwaysHydrolysis is one of the potential degradation pathways of peptides. In aqueous solution peptide bonds can undergo chemical reactions, usually through an attack of an electronegative atom (nucleophile) on the carbonyl carbon, resulting in cleavage of peptide bonds. In an alkaline aqueous environment, hydroxyl ions are better nucleophiles than polar molecules such as water. Under acidic conditions, the carbonyl group becomes protonated, which leads to a much easier nucleophile attack [58].

2.1.1 Acid/base catalyzed hydrolysisThe stability of peptides in aqueous solution strongly depends on pH. It was reported that at pH range of 1-3 gonadorelin and triptorelin mostly undergo acid-catalysed hydrolysis via deamidation of the C-terminal amide to form free acid deamidated products. In the pH range 5−6, the peptide backbones of gonadorelin and triptorelin are hydrolyzed at the N-terminal side of the serine residue probably involving nucleophilic addition of the serine hydroxyl group to the neighboring amide bond forming a cyclic intermediate resulting in fragmentation [47,59]. At pH values over 7, base- catalyzed epimerization is the main pathway of degradation. Serine is most likely involved in base-catalyzed epimerization through a carbanion intermediate. Its ability to form a relatively stable six membered intermediate with a hydrogen bridge explains the relative high rate of racemization of the L-serine residue compared to other amino acid residues. Parallel to the epimerization, base-catalysed hydrolysis of gonadorelin and triptorelin occurs. Epimerization of serine is considered the most important degradation reaction for hydroxyl-catalyzed degradation of gonadorelin and triptorelin [47,48,60].

Acid/base catalyzed hydrolysis has also been observed in the degradation of somatostatin analog octastatin in aqueous formulations. This hydrolysis is also affected by the buffer species. Jang et. al. [43] reported that the degradation rate of octastatin was higher in phosphate-buffered solution than in glutamate buffers. Increasing buffer concentrations resulted in a greater degradation of octastatin in phosphate-containing buffer, presumably

18

2


because of catalytic effects of phosphate ions. In contrast, in the glutamate-buffered solution, increasing buffer concentrations caused greater stabilization, presumably due to a stabilizing effect of glutamic acid by both ionic and hydrophobic interactions between octastatin and glutamate ions.

2.1.2 Deamidation of Asn and Gln residuesDeamidation is regarded as the most common chemical degradation pathway for peptides and proteins. Peptides containing asparagine (Asn) and glutamine (Gln) residues are known to undergo spontaneous deamidation to form aspartic acid (Asp) and glutamic acid (Glu) residues under physiological conditions. Under acidic conditions (pH below 3), deamidation of Asn residues is reported to proceed through direct hydrolysis of the Asn side chain amide to form Asp. Similarly, Gln residues are converted to Glu by acid catalyzed direct hydrolysis [36].

At alkaline and neutral conditions, the deamidation of Asn residues predominantly occurs through a cyclic imide intermediate formed by an intermolecular reaction in which the carbonyl carbon in the side chain of Asn residue is attacked by the nitrogen in the amino acid residue following Asn. Therefore, the rate of deamidation via this pathway depends on

Table 2. Reported degradation pathways of different peptides

Peptide

Number of amino acid

residues Degradation pathways Reference

Thyrotropin Releasing Hormones 3 Hydrolysis [38]Ceftazidime 5 Hydrolysis [39,40]

Eptifibatide 6 Hydrolysis, isomerization, deamidation, oxidation, and dimerization [41]

Octreotide 8 Hydrolysis [42,43]

Oxytocin 9 Hydrolysis, deamidation, oxidation, beta-elimination, and dimerization [31,44]

Desmopressin 9 Beta-elimination, deamidation, disulfide exchange, racemization, and oxidation [45]

Leuprolide 10 Hydrolysis, isomerization, oxidation, and aggregation [46]

Goserelin 10 Acid-base catalyzed hydrolysis [47]Gonadorelin 10 Acid-base catalyzed hydrolysis [47,48]Triptorelin 10 Acid-base catalyzed hydrolysis [47,48] Somatostatin and analogues 14 Hydrolysis, [42,43]

Calcitonin 32 Deamidation, oxidation, and aggregation/fibrilation [8,49,50]

Human Brain Natriuretic Peptide [hBNP (1–32)] 32 Aggregation,

deamidation, and oxidation [51]

Human Parathyroid Hormone [hPTH (1–34) 34 Aggregation, deamidation, and oxidation [51]

Enfuvirtide 36 Deamidation and aggregation [52,53]Adrenocorticotropin Hormone (ACTH) 39 Deamidation of Asn residues, and

racemization [54]

Corticotropin Heleasing factor 41 Oxidation [55]

Amyloid-β (Aβ) Peptides 36-43 Metal-catalyzed oxidation,dimerization and aggregation [56,57]

19

2

the nature of the amino acid residue on the carboxyl side of the Asn residue [54,61]. At the same condition, deamidation of Gln residues is less common than for Asn, because cyclization of Asn residues leads to a five-membered ring. With Gln, the corresponding intermediate is a six-membered ring, which formation is thermodynamically less favorable than the smaller, five-membered, ring [36].

The high rate of deamidation is caused by the high degree of peptide chain flexibility. The rate of deamidation also depends on the amino acid sequence in the peptides [62]. The amino acid residues following asparagine (Asn) such as glycine (Gly), alanine (Ala), serine (Ser), and aspartic acid (Asp) may seriously increase the reaction rate in the degradation of therapeutic peptides [63]. Catak et al. [64] reported that deamidation in aqueous solution can occur either through direct hydrolysis or through succinimide-mediation. These two reactions are competitive even in the absence of acid or base catalysis (Figure 1).

Adrenocorticotropic hormone (ACTH), a 39-amino acid polypeptide with a single Asn residue, was shown to be degraded via deamidation of Asn residues [54,65]. The deamidation of Asn or Gln residues was also observed in the degradation of salmon calcitonin [49]. The nonapeptide oxytocin is another example of a peptide that can undergo deamidation and involves the hydrolysis of Asn5 and Gln4 side chain amides. It was also reported that Gly-NH2 of oxytocin undergoes deamidation at a pH of 2. [31].

2.1.3 Isomerization of Asp residuesDeamidation of the Asn residue in neutral conditions proceeds through the formation of a cyclic succinimide intermediate by the intermolecular attack of the backbone nitrogen atom on the carbonyl group of the Asn side chain [66]. Hydrolysis of the succinimide intermediate generates the formation of isoaspartic (isoAsp) and aspartic acid (Asp)

Figure 1. Deamidation pathways of Asn residue via A. direct hydrolysis and B. succinimide mediation [64].

20

2


residues. Furthermore, direct dehydration occurs in a transformation of aspartic acid (Asp) into its isoform isoAsp through the same cyclic succinimide intermediate [67]. Additionally, L-succinimide may be racemized to D-succinimide to form the D-Asp and D-isoAsp enantiomers [61,68]. Thus, the mechanisms for the deamidation and the isomerization reactions are similar since they both proceed via an intra-molecular cyclic succinimide intermediate. The formation of succinimide intermediate is the rate-limiting step in both the deamidation and the isomerization reactions at physiological pH [69].

Under acidic conditions, cleavage at the Asp residue is the most important degradation pathway of recombinant human parathyroid hormone (rhPTH). While at pH values above 5 the Asn deamidation is the major degradation route [70].

2.2 Oxidation pathwaysOxidation is another primary chemical degradation pathway that can occur in a peptide. Oxidation of organic molecules is defined as an increase in oxygen or a decrease in hydrogen content [71]. Alternatively, oxidation can be defined as a reaction that increases the content of more electronegative atoms in a molecule, in which the electronegative heteroatoms are generally oxygen or halogens.[72]. Any peptide that contains cysteine (Cys), methionine (Met), histidine (His), tyrosine (Tyr) and tryptophan (Trp) residues can potentially be damaged due to their high reactivity with various reactive oxygen species (ROS). Cys and Met are at risk for oxidation because of their sulfur atoms, and His, Tyr and Trp because of their aromatic rings [73]. Figure 2 illustrates the oxidation reaction of Met and His.

Figure 2. Oxidation reaction of A: methionine to methionine sulfoxide in acidic condition (HA) and B: histidine through an oxometallacycyclic intermediate to form various degradation products, such as 2-oxomidazoline, asparagine, and aspartate (adapted from Li, et. al, 1995)

21

2

Oxidation can be induced by contaminating oxidants and light exposure. Oxidation can be catalyzed by the presence of (traces of) transition metal ions during processing and storage. Oxidation of peptides may furthermore be influenced by pH, temperature, and buffer composition. [63].

2.2.1 Auto-oxidationA rapid oxidation reaction of a peptide may occur when the ROS is able to access the side chain, particularly for peptides with Met residues autooxidation may be a major degradation pathway [36]. Adrenocorticotropic hormone (ACTH) and human atrial natriuretic peptide (ANP) are examples of peptides that undergo spontaneous Met-oxidation [74].

Cysteine residues may also undergo spontaneous oxidation to form the molecular byproducts sulfinic acid and cysteic acid in the presence of metal ions or nearby thiol groups. Cysteine oxidation involves a nucleophilic attack of thiolate ions on disulfide bonds, generating new disulfide bonds and other thiolate ions. The newly formed thiolate ions can subsequently react with another disulfide bond to form cysteine [75].

2.2.2 Metal ion-catalyzed oxidationMetal-catalyzed oxidation occurs when a redox active metal ion binds to the peptide. The amino acid residues that are most susceptible to metal ion catalyzed oxidations are His, Arg, Lys, Pro, Met, and Cys. Of these amino acids, His and Cys are sensitive to oxidative damage, as the ROS generated at the metal center does not have to diffuse very far before reacting with the peptide [36,76]. Several metals, such as iron and copper have a strong catalytic power to generate highly reactive hydroxyl radicals if reacted with hydrogen peroxide (Fenton reaction). The hydroxyl radicals may react with His residue to form 2-oxo-His[77]. Metal ion-catalysis has been observed in the oxidation of amyloid-β (Aβ) peptides which mediates the neurotoxicity of Alzheimer’s disease. The amino acid residue involved in this pathway is His [56].

2.2.3 Light-induced oxidationIn 2007, Kerwin and Remmel [78] summarized light-induced degradation in biopharmaceuticals. These reactions may occur at several points from production to delivery of the products. Light-induced oxidation is initiated when a compound absorbs a certain wavelength of light, which provides energy to raise the molecule to an excited state. The excited molecule can then transfer that energy to molecular oxygen, converting it to reactive singlet oxygen atoms. This is how tryptophan, histidine, and tyrosine can be modified under light in the presence of oxygen [37,79]. Tyrosine photo-oxidation can produce mono-, di-, tri-, and tetrahydroxyl tyrosine as byproducts [80]. Aggregation is observed in some peptides due to cross-linking between oxidized tyrosine residues [81].

2.3 β-elimination reactionsDisulfide-bridge peptides might undergo destruction to form free thiol groups through β-elimination. This reaction is commonly observed when peptides are incubated at higher pH values or elevated temperature [82,83]. At neutral and alkaline pH levels peptides can undergo a disulfide exchange reaction which is catalyzed by thiol groups [84]. When cystine-containing proteins are heated at 100°C, even at a neutral pH, they undergo destruction to form free thiols [85]. Salmon calcitonin (sCT) degrades via β-elimination at a disulfide bridge between the cysteine residues at positions 1 and 7. The

22

2


insertion of an extra sulfur to form a trisulfide bridge has also been reported as a result of a β-elimination reaction [8].

2.4 Disulfide exchange reactionsDisulfide exchange reactions contribute to the formation of dimers and larger aggregates. These reactions may occur at the cysteine–cysteine link. In a study on the degradation of sCT, dimeric products were found to be formed via disulfide exchange reactions, however, the disulfide-linked dimers may undergo further disulfide exchange reactions that will eventually regenerate sCT monomers [8]. Disulfide interchange at aqueous acid conditions proceeds through sulfenium ions, which arise from the acid hydrolysis of the disulfide bond [86]. Several studies on the disulfide exchange reaction and the importance of disulfide-bridges for the stability of peptides have been reported [8,82,86,87].

2.5 Dimerization, aggregation, and precipitationThe Cys residues in the oxytocin molecule are able to form disulfide bonds due to thiol exchange between two oxytocin molecules [31,44]. Thiol exchange occurs at neutral and alkaline conditions via a nucleophilic attack of free thiolate on one of the sulfur atoms within the disulfide bridge. At lower pH values the disulfide exchange progresses via a sulfenium cation, formed following protonation of the disulfide bridge [36,86]. In either case, disulfide exchange can result in dimerization and progressive aggregation [31]. As mentioned above, peptide molecules are also able to form dimers due to light or metal induced oxidation of tyrosine residues to form dityrosine-linked dimers [31,88,89]. Aggregation may be induced by several stress condition, such as heating, freezing or agitation. Aggregates can form either via covalent bonds, such as disulfide bonds, ester, or amide linkage or non-covalent bonds occurring via hydrophobic interaction or charge-charge complexation. The relative weakness of non-covalent protein bonds can lead to aggregate disruption during the analytical process [90].

There is no single pathway by which peptides can form aggregates [36]. Both mechanisms can occur simultaneously leading to the formation of either soluble or insoluble aggregates [91]. Furthermore higher concentrations of the peptide in solution will bring aggregates more easily to a later stage of physical instability i.e. precipitation [12]. Calcitonin, β-amyloid peptide, leuprolide, and deterelix have been reported to form gel-like aggregates. The gel formation is a function of peptide concentration, temperature, time, pH, and agitation. However the chemical stability and dimerization of leuprolide was not affected by gelation [92].

3. strategIes to Improve peptIde stabIlIty In aqueous formulatIons Many efforts have been made to improve the stability of peptides in aqueous solution. Examples are: the use of buffers, metal ions or organic solvents, and oxygen removal. However, an optimal use of these strategies requires first of all knowledge of the peptide´s structure and thorough understanding of the possible degradation pathways of the specific peptide.

Examining the amino acid sequence of a peptide can provide an insight into how the molecule may degrade. This will be of real use for peptides of which the amino acid sequence is generally exposed on the aqueous environment and therefore able to undergo chemical

23

2

degradation. The degradation pathways and involved amino acid sequences as well as the formulation strategies to inhibit the specific degradation pathway are summarized in Table 3.

3.1 Optimizing hydrolytic stability

3.1.1 pH optimization and buffer speciesStability of peptides depends on the pH, therefore the common strategy to avoid instability of peptides in aqueous solution is adjusting the pH using buffers. In table 1 the pH values and the pH adjusting agent or buffer of several marketed products are given. Considering modern formulation development strategies it may be assumed that these values represent a value which is (close to) the optimum pH stability of the different peptides. The acceptable pH range for (slow) intravenous administration is 3-10.5 and 4-9 for other parenteral routes to minimize discomfort at the injection site [93,94]. Therefore, it is important to study the pH stability of a peptide in the range of pH 3-10 with different buffers in the early stages of formulation development [95,96]. Hydrolysis of the side chain amide on glutamine and asparagine residues could be reduced by formulating at a pH below neutral. However, the pH should not be lower than 3 to prevent direct hydrolysis of the Asn and Gln side chains and to minimize hydrolytic fragmentation [36]. Oxytocin was described to have the highest stability at pH 4.5 [31]. Calcitonin undergoes hydrolysis at basic pH, but no such degradation is observed at a pH of 7 even at room temperature [97]. Other mechanisms for stabilization by buffers have also been reported: In some cases they can act as radical scavengers [98]. Even more important is the fact that some buffers are able to bind directly to peptides, thereby increasing their conformational stability [36]. Citric acid has been reported to bind with oxytocin forming N-cytril oxytocin which increases the stability of oxytocin in the presence of divalent metal ions [44].

The pH of the formulation can also substantially affect deamidation. Deamidation is known to be sensitive to both the pH and the composition and concentration of the buffer. In general, formulations with a pH in the range from 3 to 5 minimize peptide deamidation [65,98,105]. Buffer species can also affect deamidation. It has been reported that the rate of deamidation is faster in phosphate and bicarbonate buffers than in acetate and pyruvate buffers [96]. Isomerization is greatest at low pH values for asparagine and glutamine residues [106].

3.1.2 Ionic strengthThe total concentration of dissolved electrolytes might affect the rate of hydrolysis. The increase in ionic strength could have a stabilizing or destabilizing effect on a peptide, depending on the nature of the charge–charge interactions within the peptide [107-109]. However, in another study, no significant effect of the ionic strength on the rate of deamidation or hydrolysis in small peptides was found [110].

3.1.3 CosolventCosolvents such as low molecular weight polyethylene glycol (PEG) were shown to reduce the aggregation of several peptides [9,12,111]. Peptide deamidation can be modestly inhibited in an aqueous solution by the addition of glycerol [100], or ethanol [112]. Addition of organic solvents decreases the dielectric constant of an aqueous solution. Reduction of the solvent´s dielectric strength leads to significantly lower rates of isomerization and deamidation [69]. The use of polyols includes polyhydric alcohols and carbohydrates

24

2


might be used to lower degradation. Examples of polyols used for cosolvent for hydrolytic stabilization of peptides are glycerol and propylene glycol [113]. The hydrolytic stability of the cyclic heptapeptide eptifibatide increased by formulating it in 10% ethanol, 40% propylene glycol and 50% 0.025 M citrate buffer at a pH of 5.75 [41]. Poloxamer 407 (Pluronic® F127) retarded the rate of deamidation of the Asn residue in a model peptide Val-Tyr-Pro-Asn-Gly-Ala [103].

3.1.4 ViscosityThe effect of solution viscosity on the deamidation rates has been explored in some peptides using various concentrations of glycerol and polyvinylpyrrolidone (PVP). It has been shown that a high concentration of PVP reduces the Asn deamidation rate of the Asn-hexapeptide deamidation. However, it is unclear whether the decreased deamidation rate was due to the increased viscosity or due to a decreased dielectric constant of the solution by the high PVP concentration or a combination of both [99,100].

3.2 Optimizing oxidative stabilityOxidation can be activated by specific pH values and the presence of oxygen, light and/or metal ions. Therefore, oxidative damage can be decreased by expelling oxygen from the solution,

Table 3. Degradation pathways and possible stabilization strategies for peptides including amino acid residues involved in the degradation.

Degradation pathway Stabilization strategyAmino acid

residue(s) involved Ref.

Chemical Instability

Acid/base catalyzed hydrolysis pH, buffer species, co-solventsSer TrpAsn-ProAsn-Tyr

[47,60][42]

Deamidation pH 3-5, increased solvent viscosity Asn, Gln [62,69][99,100]

Isomerization pH 6<pH<8 Asp [69]Racemization pH below 5 Asp [69]β-elimination Buffer type, divalent metal ions Cys-Cys [31,44]

Oxidation pH < 7, air exclusion, antioxidants Trp, Met, Cys, Tyr, His [101]

Light induced oxidation Protect from light Trp [37,79]

Metal induced oxidation Chelating agents, polyols His, Cys, Arg, Pro, Met

[56,76][102]

Disulfide exchange Surfactans, polyols and sugarsBuffer and divalent metal ions Cys-Cys

[103][36]

[104]Physical Instability

Dimerization and further aggregation

Lower concentration, minimal mechanical stress, organic solvents,alkyl saccharide, alkyl polyglycoside

Cys-CysTyr-Tyr

[103][36]

[104]

Adsorption surfactans and polymers HisArg [33]

25

2

adjusting the primary and secondary packaging to protect from light, and the use of antioxidants in the formulation. Watermann and co-workers have made a guideline on the use of excipients to optimize oxidative stability including their recommended concentrations [101].

3.2.1 BuffersThe side chains of tryptophan, methionine and cysteine and, to a lesser extent, tyrosine and histidine are potential oxidation sites. The modification of indole group of Trp, thiol group of Cys, imidazole group of His, and phenolic side chain of Tyr by the reactive oxygen species is most significant at neutral and alkaline condition. Therefore, a lower pH reduces oxidation of peptides containing these amino acids [114]. In contrast, the thioether group of methionine can be readily oxidized by certain reagents under acidic conditions [58].

3.2.2 Air exclusion Removal of oxygen by bubbling another gas, for example nitrogen, through the solution may be effective to exclude oxygen. For some oxidation-sensitive peptide drugs, processing and filling steps should be carried out in the presence of an inert gas such as nitrogen, argon or helium. To further minimize air oxidation violent stirring should be avoided [115].

3.2.3 AntioxidantsAntioxidants are commonly used to protect peptides from oxidation during processing and formulation. Some antioxidants, however, can be problematic in peptide formulations [37,114]. For example, bisulfite is not a suitable agent because it is a strong nucleophile, which may interact with peptides. Furthermore, addition of antioxidants to trace metal ions contaminated peptide solutions will not protect the peptide from oxidative modification. In contrast, it can even accelerate the oxidation process, as demonstrated by the use of ascorbic acid which promotes rather than inhibits oxidation of the Met residue of small model peptides in the presence of metal ions [116,117]. Methionine, a sulfur containing amino acid, is readily oxidized to methionine sulfoxide by many reactive oxygen species and oxidizes more easily than the peptide, and therefore can act as a sacrificial antioxidant [118].

3.2.4 Chelating agentsAnother method to reduce free radical oxidation is to include chelating agents. Chelating agents are used to inhibit oxidation by complexation of trace metal ions. The most commonly used chelating agents in pharmaceutical formulations are ethylenediaminetetraacetic acid (EDTA), desferal, diethylenetriaminepentaacetic acid (DTPA), inositol hexaphosphate, ethylenediamine bis(o-hydroxyphenylacetic acid), tris(hydroxymethyl)aminomethane (TRIS), citric acid, and tartaric acid. EDTA has been recommended as a chelating agent for copper ions, whereas desferal, DTPA, inositol hexaphosphate and ethylenediamine bis(o-hydroxy-phenylacetic acid) were recommended for iron ions [113]. It is well known that chelating agents are generally effective in stabilizing peptides against oxidation. However, it cannot be assumed that addition of a certain chelating agents will be able to interact with all trace metal ions and completely eliminate oxidation. Under certain circumstances, chelating agents may even accelerate the oxidation process [37,119].

3.2.5 PolyolsPolyols such as mannitol, trehalose, sucrose, maltose, and raffinose can prevent oxidation of therapeutic peptides. For example, mannitol was shown to inhibit the iron-catalyzed oxidation of Met-containing peptides [102]. Sucrose has been shown to decrease the

26

2


rate of oxidation of parathyroid hormone hPTH (1–34) and brain natriuretic hormones hBNP (1–32) in liquid formulations [51].

3.3 Protection against disulfide exchange reactionFormulating octreotide in glycine with pharmaceutically acceptable salts and HCl was found to be effective in protecting its disulfide-bridge from cleavage, and was also reported to be better tolerated than the formulation in acetate, lactate or bicarbonate [120]. Divalent metal ions in combination with specific buffers may protect peptide drugs against disulfide exchange reaction. Recently, it was shown that Zn2+, Ca2+ and Mg2+ ions in combination with dicarboxylic and tricarboxylic acids improved the stability of oxytocin [44,104]. The addition of at least 2 mM CaCl2, MgCl2, or ZnCl2 to 5 and 10 mM citrate-buffered solutions at pH 4.5 increased oxytocin stability in aqueous formulations and the stability is further increased with increasing concentration of the divalent metals ions up to 50 mM. The improved stability is due to the reactivity of carboxyl ions in citrate buffer which can attack the N-terminal group of Cys to form an adduct and together with divalent metal ions suppress the disulfide exchange reaction [44,104].

3.4 Inhibition of dimerization, aggregation, and precipitation Dimerization and aggregation may involve the interchange of covalent bonds, such as disulfide bridges or non-covalent forces such as hydrophobic interactions. Both soluble and insoluble aggregates may occur. A peptide in aqueous solution can be stabilized against dimerization and aggregation by optimizing the pH and ionic strength of the solution. Furthermore, dimerization can be prevented by preferential exclusion using sugars, amino acids, and/or polyols, and by using surfactants [36]. The use of a combination of divalent metal ions and citrate buffer has been found to inhibit the cysteine-mediated dimerization of oxytocin [44]. PEG has been shown to reduce the aggregation of several peptides [9,12,111]. The stabilizing effects increased with increasing concentration and increasing molecular weight [99,100]. Dicarboxylic amino acids such as aspartic and glutamic acid have been used to reduce aggregation [121] and glycine, arginine and lysine have also been reported to prevent aggregation [121-123]. Polysorbate 20 and 80 [124,125] can reduce agitation-induced aggregation of peptides. However, surfactants are reported to be less effective in reducing thermally-induced aggregation. Therefore, the optimum concentration required to protect a specific peptide should be evaluated for each different type of stress that may be encountered.

3.5 Preferential exclusion Another strategy to minimize peptide degradation is the use of extremolytes. Extremolytes are small organic molecules produced by extremophilic microorganisms, to protect their biological macromolecules from damage by external stresses [126]. Examples of extremolytes are the polyol derivatives ectoin and hydroxyectoin [127], betain [128], trehalose [129], amino acids (e.g. proline), and the mannose derivative mannosylglycerate [130]. Mannosylglycerate was reported to have the ability to inhibit β-amyloid peptide aggregation [131]. Extremolytes are known to stabilize peptides by forming solute hydrate clusters that are preferentially excluded from the hydrate shell of the peptide as a result of the repulsive interactions between extremolytes and the backbone of the peptides. Accumulation of water near peptide domains assembles the peptide into a more compact structure with a reduced surface area [132-134]. However, there are no

27

2

extremolytes that can be used as a universal stabilizer for all peptides in aqueous solution. In addition, it should be realized that extremolytes may also act as destabilizer for certain peptides under specific conditions [135].

Surfactants have also been described as stabilizers for peptides. Pluronic® F68 has been used to enhance stability of peptide drugs ceftazidime in parenteral solutions. In this approach, the peptides are kept in a microenvironment that shields them from exterior conditions, limiting the effect of moisture and pH [40].

5. conclusIonIn aqueous solutions peptides are often unstable. Peptides have unique structures that differ from proteins in that they are smaller and rarely have a tendency to physical degradation e.g. unfolding. Peptides often do not possess a higher order structure that can sequester reactive groups, the side chain of nearly all of the amino acid residues are fully solvent exposed, allowing maximal contact with solvents and the degradation rates appear to correlate with the degree of solvent exposure. Based on knowledge of the peptide’s structure and an understanding of the predominant degradation pathways, the strategies may be developed to achieve adequate stability of the formulation. The degradation pathways of peptides are mainly dependent on the amino acid sequence. The most prominent degradation pathways for peptides are hydrolysis, oxidation, and dimerization. Formulating peptides in a specific pH with a specific buffer, avoiding oxygen reactive species, and minimizing solvent exposure eliminate chemical degradation. Increasing solution viscosity by using sugars or polymers reduces peptide mobility and further decelerates physical degradation.

references

1. V. du Vigneaud, C. Ressler, S. Trippett, The sequence of amino acids in oxytocin, with a proposal for the structure of oxytocin, J. Biol. Chem. 205 (1953) 949-57.

2. J. Sterling, GEN’s 30th Anniversary: Peptide Therapeutics, Editor’s Note, Genetic Engineering and Biotechnology News 31 (2011).

3. L. Malavolta, F.R. Cabral, Peptides: Important tools for the treatment of central nervous system disorders, Neuropeptides 45 (2011) 309-16.

4. V.H.L. Lee (Ed.), Peptide and Protein Drug Delivery, Marcel Dekker Inc., New York, 1991.

5. S.Y. Lin, S.L. Wang, Advances in simultaneous DSC-FTIR microspectroscopy for rapid solid-state chemical stability studies: Some dipeptide drugs as examples, Adv. Drug Deliv. Rev. 64 (2012) 461-78.

6. T. Wieland, M. Bodanszky (Eds.), The Worlds of Peptides: A Brief History of Peptide Chemistry, Springer-Verlag, Berlin, 1991.

7. J. Pless, The history of somatostatin analogs, J. Endocrinol. Invest. 28 (2005) 1-4.

8. V. Windisch, F. DeLuccia, L. Duhau, F. Herman, J.J. Mencel, S.Y. Tang, M. Vuilhorgne, Degradation pathways of salmon calcitonin in aqueous solution, J. Pharm. Sci. 86 (1997) 359-64.

9. S. Frokjaer, L. Hovgaard (Eds.), Pharmaceutical Formulation Development of Peptides and Proteins, CRC Press, United States America, 2000.

10. P.A. Zhou XH, Peptide and protein drugs. Therapeutic applications, absorption and parenteral administration, Int J Pharm 75 (1991) 97-115.

11. A.M. Papini, Peptide Chemistry Revolution, Chemistry Today 29 (2011).

12. M.F. Powell, L.M. Sanders, A. Rogerson, V. Si, Parenteral peptide formulations: chemical and physical properties of native luteinizing hormone-releasing hormone (LHRH) and hydrophobic analogues in aqueous solution, Pharm. Res. 8 (1991) 1258-63.

13. F. Veuillez, Y.N. Kalia, Y. Jacques, J. Deshusses, P. Buri, Factors and strategies for improving buccal absorption of peptides, Eur. J. Pharm. Biopharm. 51 (2001) 93-109.

28

2


14. D.H. Oh, K.H. Chun, S.O. Jeon, J.W. Kang, S. Lee, Enhanced transbuccal salmon calcitonin (sCT) delivery: Effect of chemical enhancers and electrical assistance on in vitro sCT buccal permeation, Eur. J. Pharm. Biopharm. 79 (2011) 357-63.

15. K.V. Patel, N.D. Patel, H.D. Dodiya, P.K. Shelat, Buccal Bioadhesive Drug Delivery System: An Overview, Int. J. Pharm. Biol. Arch. 2 (2011) 600-9.

16. A. Nasr, A.Y. Shahin, A.M. Elsamman, M.S. Zakherah, O.M. Shaaban, Rectal misoprostol versus intravenous oxytocin for prevention of postpartum hemorrhage, Int. J. Gynaecol. Obstet. 105 (2009) 244-7.

17. K. Morimoto, H. Akatsuchi, R. Aikawa, M. Morishita, K. Morisaka, Enhanced rectal absorption of [Asu’.‘]-eel calcitonin in rats using polyacrylic acid aqueous gel base, J. Pharm. Sci. 73 (1984) 1366-8.

18. A.P. Sayani, Y.W. Chien, Systemic delivery of peptides and proteins across absorptive mucosae, Rev. Ther. Drug Carrier Syst. 13 (1996) 85-184.

19. J.L. Richardson, L. Illum, (D) Routes of delivery: Case studies: (8) The vaginal route of peptide and protein drug delivery, Adv. Drug Deliv. Rev. 8 (1992) 341-66.

20. N. Seo, M. Takigawa, The current status and future direction of percutaneous peptide immunization against melanoma, J. Dermatol. Sci. 48 (2007) 77-85.

21. S.P. Read, S.M. Cashman, R. Kumar-Singh, A poly(ethylene) glycolylated peptide for ocular delivery compacts DNA into nanoparticles for gene delivery to post-mitotic tissues in vivo, J. Gene Med. 12 (2010) 86-96.

22. D. Harris, J.H. Liaw, J.R. Robinson, (D) Routes of Delivery: Case Studies (7) Ocular delivery of peptide and protein drugs, Adv. Drug Deliv. Rev. 8 (1992) 331-9.

23. H.E. Bodde, J.C. Verhoef, M. Ponec, Transdermal peptide delivery, Biochem. Soc. Trans. 17 (1989) 943-5.

24. M.R. Prausnitz, A peptide chaperone for transdermal drug delivery, Nature Biotechnology 24 (2006) 416-7.

25. A.E. Pontiroli, Peptide hormones: Review of current and emerging uses by nasal delivery, Adv Drug Deliv Rev 29 (1998) 81-7.

26. S. Chhajed, S. Sangale, S.D. Barhate, Advantageous Nasal Drug Delivery System: A Review, IJPSR 2 (2011) 1322-36.

27. S. Turker, E. Onur, Y. Ozer, Nasal route and drug delivery systems, Pharm. World Sci. 26 (2004) 137-42.

28. S.A. Shoyele, A. Slowey, Prospects of formulating proteins/peptides as aerosols

for pulmonary drug delivery, Int. J. Pharm. 314 (2006) 1-8.

29. R.I. Henkin, Inhaled insulin-intrapulmonary, intranasal, and other routes of administration: mechanisms of action, Nutrition 26 (2010) 33-9.

30. X.H. Zhou, A. Li Wan Po, Peptide and protein drugs: II. Non-parenteral routes of delivery, Int. J. Pharm. 75 (1991) 117-30.


32. International Conference on Harmonisation; Stability Data Package for Registration Applications in Climatic Zones III and IV; Stability Testing of New Drug Substances and Products; availability. Notice, Fed. Regist. 68 (2003) 65717-8.

33. S.T. Anik, J.Y. Hwang, Adsorption of D-Nal(2)6LHRH, a decapeptide, onto glass and other surface., Int. J. Pharm. 16 (1983) 181-190.

34. V.G. Badelin, O.V. Kulikov, V.S. Vatagin, Physico-chemical properties of peptides and their solutions, Thermochimica Acta 169 (1990) 81-93.

35. M. Zahn, P.W. Kallberg, G.M. Slappendel, H.M. Smeenge, A risk-based approach to establish stability testing conditions for tropical countries, J. Pharm. Sci. 95 (2006) 946-65.

36. M.C. Manning, D.K. Chou, B.M. Murphy, R.W. Payne, D.S. Katayama, Stability of protein pharmaceuticals: an update, Pharm. Res. 27 (2010) 544-75.

37. S. Li, C. Schoneich, R.T. Borchardt, Chemical instability of protein pharmaceuticals: Mechanisms of oxidation and strategies for stabilization, Biotechnol. Bioeng. 48 (1995) 490-500.

38. T. Hashimoto, K. Ohki, N. Sakura, Hydrolytic cleavage of pyroglutamyl-peptide bond. I. The susceptibility of pyroglutamyl-peptide bond to dilute hydrochloric acid, Chem. Pharm. Bull. (Tokyo) 43 (1995) 2068-74.

39. J.M. Hwang, T.E. Piccinini, C.J. Lammel, W.K. Hadley, G.F. Brooks, Effect of storage temperature and pH on the stability of antimicrobial agents in MIC trays, J. Clin. Microbiol. 23 (1986) 959-61.

40. C.A.D. Santos, G.B. Ribeiro, M.C. Knirsch, A.P. Junior, T.C.V. Penna, Influence of Pluronic® F68 on ceftazidime biological activity in parenteral solutions, J. Pharm. Sci. 100 (2011) 715-20.

29

2

41. L. Zhao, S.H. Yalkowsky, Stabilization of eptifibatide by cosolvents, Int. J. Pharm. 218 (2001) 43-56.

42. R. Krishnamoorthy, A.K. Mitra, Kinetics and mechanism of degradation of a cyclic hexapeptide (somatostatin analogue) in aqueous solution, Pharm. Res. 10 (1992) 1314-20.

43. S.W. Jang, B.H. Woo, J.T. Lee, S.C. Moon, K.C. Lee, P.P. DeLuca, Stability of Octastatin, a somatostatin analog cyclic octapeptide, in aqueous solution, Pharm. Dev. Technol. 2 (1997) 409-14.

44. C. Avanti, H.P. Permentier, A.V. Dam, R. Poole, W. Jiskoot, H.W. Frijlink, W.L.J. Hinrichs, A new strategy to stabilize oxytocin in aqueous solutions: II. Suppression of cysteine-mediated intermolecular reactions by a combination of divalent metal ions and citrate, Mol. Pharmaceutics 9 (2012) 554-62.

45. S.L. Law, K.J. Huang, V.H. Chou, Stability of desmopressin loaded in liposomes, J. Liposome Res. 13 (2003) 269-77.

46. S.C. Hall, M.M. Tan, J.J. Leonard, C.L. Stevenson, Characterization and comparison of leuprolide degradation profiles in water and dimethyl sulfoxide, J. Pept. Res. 53 (1999) 432-41.

47. M.A. Hoitink, J.H. Beijnen, M.U.S. Boschma, A. Bult, E. Hop, J. Nijholt, C. Versluis, Wiese,G.,Underberg,W.J.M., Identification of the degradation products of gonadorelin and three analogues in aqueous solution, Anal. Chem. 69 (1997) 4972-8.

48. V.J. Helm, B.W. Muller, Stability of gonadorelin and triptorelin in aqueous solution, Pharm. Res. 7 (1990) 1253-6.

49. K.C. Lee, Y.J. Lee, H.M. Song, C.J. Chun, P.P. DeLuca, Degradation of synthetic salmon calcitonin in aqueous solution, Pharm. Res. 9 (1992) 1521-3.

50. S. Seyferth, G. Lee, Structural studies of EDTA-induced fibrillation of salmon calcitonin, Pharm. Res. 20 (2003) 73-80.

51. M. Kamberi, Y.J. Kim, B. Jun, C.M. Riley, The effects of sucrose on stability of human brain natriuretic peptide [hBNP (1-32)] and human parathyroid hormone [hPTH (1-34)], J. Pept. Res. 66 (2005) 348-56.

52. H. Stocker, C. Kloft, N. Plock, A. Breske, G. Kruse, C. Herzmann, H. Schulbin, P. Kreckel, C. Weber, F. Goebel, J. Roeling, S. Staszewski, A. Plettenberg, C. Moecklinghoff, K. Arasteh, M. Kurowski, Pharmacokinetics of enfuvirtide in patients treated in typical routine clinical settings, Antimicrob. Agents Chemother. 50 (2006) 667-73.

53. R.W. Payne, R. Nayar, R. Tarantino, S. Del Terzo, J. Moschera, J. Di, D. Heilman, B. Bray,

M.C. Manning, C.S. Henry, Second virial coefficient determination of a therapeutic peptide by self-interaction chromatography, Biopolymers 84 (2006) 527-33.

54. N.P. Bhatt, K. Patel, R.T. Borchardt, Chemical pathways of peptide degradation. I. Deamidation of adrenocorticotropic hormone, Pharm. Res. 7 (1990) 593-9.

55. I.Q. Assil, M.E. Shomali, A.B. Abou-Samra, An oxidation resistant radioligand for corticotropin-releasing factor receptors, Peptides 22 (2001) 1055-61.

56. K. Inoue, A. Nakagawa, T. Hino, H. Oka, Screening assay for metal-catalyzed oxidation inhibitors using liquid chromatography-mass spectrometry with an N-terminal beta-amyloid peptide, Anal. Chem. 81 (2009) 1819-25.

57. D.P. Smith, G.D. Ciccotosto, D.J. Tew, M.T. Fodero-Tavoletti, T. Johanssen, C.L. Masters, K.J. Barnham, R. Cappai, Concentration dependent Cu2+ induced aggregation and dityrosine formation of the Alzheimer’s disease amyloid-beta peptide, Biochemistry 46 (2007) 2881-91.

58. M.C. Manning, K. Patel, R.T. Borchardt, Stability of protein pharmaceuticals, Pharm Res 6 (1989) 903-18.

59. R.G. Strickley, M. Brandl, K.W. Chan, K. Straub, L. Gu, High-performance liquid chromatographic (HPLC) and HPLC-mass spectrometric (MS) analysis of the degradation of the luteinizing hormone-releasing hormone (LH-RH) antagonist RS-26306 in aqueous solution., Pharm. Res. 7 (1990) 530-6.

60. M.A. Hoitink, J.H. Beijnen, A. Bult, O.A.G.J. Van der Houwen, J. Nijholt, W.J.M. Underberg, Degradation kinetics of gonadorelin in aqueous solution, J. Pharm. Sci. 85 (1996) 1053-59.

61. T. Geiger, S. Clarke, Deamidation, isomerization, and racemization at asparaginyl and aspartyl residues in peptides: Succinimide-linked reactions that contribute to protein degradation, J. Biol. Chem. 262 (1987) 785-94.

62. K. Patel, R.T. Borchardt, Chemical pathways of peptide degradation. III. Effect of primary sequence on the pathways of deamidation of asparaginyl residues in hexapeptides, Pharm. Res. 7 (1990) 787-93.

63. R.W. Payne, M.C. Manning, Peptide formulation: challenges and strategies, Innov. Pharm. Technol. 28 (2009) 64-8.

64. S. Catak, G. Monard, V. Aviyente, M.F. Ruiz-Lopez, Deamidation of asparagine residues: direct hydrolysis versus succinimide-

30

2


mediated deamidation mechanisms, J Phys Chem A 113 (2009) 1111-20.

65. K. Patel, R.T. Borchardt, Chemical pathways of peptide degradation. II. Kinetics of deamidation of an asparaginyl residue in a model hexapeptide, Pharm. Res. 7 (1990) 703-11.

66. H. Yang, R.A. Zubarev, Mass spectrometric analysis of asparagine deamidation and aspartate isomerization in polypeptides, Electrophoresis 31 (2010) 1764-72.

67. N.E. Robinson, Protein deamidation, Procl. Natl. Acad. Sci. 99 (2002) 5283-8.

68. N.P. Sargaeva, A.A. Goloborodko, P.B. O’Connor, E. Moskovets, M.V. Gorshkov, Sequence-specific predictive chromatography to assist mass spectrometric analysis of asparagine deamidation and aspartate isomerization in peptides, Electrophoresis 32 (2011) 1962-9.

69. A.A. Wakankar, R.T. Borchardt, Formulation considerations for proteins susceptible to asparagine deamidation and aspartate isomerization, J. Pharm. Sci. 95 (2006) 2321-36.

70. Y. Nabuchi, E. Fujiwara, H. Kuboniwa, Y. Asoh, H. Ushio, The stability and degradation pathway of recombinant human parathyroid hormone: deamidation of asparaginyl residue and peptide bond cleavage at aspartyl and asparaginyl residues, Pharm. Res. 14 (1997) 1685-90.

71. T.W.G. Solomons (Ed.), Organic Chemistry, 5th ed., John Wiley and Sons, New York, 1992.

72. [72] J. March (Ed.), Advanced Organic Chemistry, 4th ed., John Wiley and Sons, New York, 1992.

73. S. Li, C. Schoneich, R.T. Borchardt, Chemical pathways of peptide degradation. VIII. Oxidation of methionine in small model peptides by prooxidant/transition metal ion systems: influence of selective scavengers for reactive oxygen intermediates, Pharm. Res. 3 (1995) 348-55.

74. Z. Guan, N.A. Yates, R. Bakhtiar, Detection and characterization of methionine oxidation in peptides by collision-induced dissociation and electron capture dissociation, J. Am. Soc. Mass Spectrom. 14 (2003) 605-13.

75. J. Patel, R. Kothari, R. Tunga, N.M. Ritter, B.S. Tunga, Stability considerations for biopharmaceuticals. Part 1 Overview of protein and peptide degradation pathways , Bioprocess International (2011) 20-31.

76. M. Khossravi, R.T. Borchardt, Chemical pathways of peptide degradation. X: Effect

of metal-catalyzed oxidation on the solution structure of a histidine-containing peptide fragment of human relaxin, Pharm. Res. 17 (2000) 851-8.

77. M. Khossravi, R.T. Borchardt, Chemical pathways of peptide degradation: IX. Metal-catalyzed oxidation of histidine in model peptides, Pharm. Res. 15 (1998) 1096-1102.

78. B.A. Kerwin, R.L. Remmele Jr, Protect from light: photodegradation and protein biologics, J. Pharm. Sci. 96 (2007) 1468-1479.

79. A.J. Grosvenor, J.D.D. Morton J.M., Profiling of residue-level photo-oxidative damage in peptides , Amino Acids 39 (2010) 285-96.

80. J. Zhang, D.S. Kalonia, The effect of neighboring amino acid residues and solution environment on the oxidative stability of tyrosine in small peptides, AAPS PharmSciTech 8 (2007) 176-83.

81. N. Rathore, R.S. Rajan, Current perspectives on stability of protein drug products during formulation, fill and finish operations, Biotechnol. Prog. 24 (2008) 504-14.

82. A.K. Galande, J.O. Trent, A.F. Spatola, Understanding base-assisted desulfurization using a variety of disulfide-bridged peptides, Peptide Science 71 (2003) 534-51.

83. S.L. Cohen, C. Price, J. Vlasak, Beta-elimination and peptide bond hydrolysis: two distinct mechanisms of human IgG1 hinge fragmentation upon storage, J. Am. Chem. Soc. 129 (2007) 6976-7.

84. A.P. Ryle, F. Sanger, Disulphide interchange reactions, Biochem. J. 60 (1955) 535.

85. D.B. Volkin, A.M. Klibanov, Thermal Destruction Processes in Proteins Involving Cystine Residues, J. Biol. Chem. 262 (1987) 2945-50.

86. R.E. Benesch, R. Benesch, The mechanism of disulfide interchange in acid solution; role of sulfenium ions, J. Am. Chem. Soc. 80 (1958) 1666-9.

87. K. Wakabayashi, H. Nakagawa, A. Tamura, S. Koshiba, K. Hoshijima, M. Komada, T. Ishikawa, Intramolecular disulfide bond is a critical check point determining degradative fates of ATP-binding cassette (ABC) transporter ABCG2 protein, J. Biol. Chem. 282 (2007) 27841-6.

88. D.A. Malencik, S.R. Anderson, Dityrosine formation in calmodulin: conditions for intermolecular cross-linking, Biochemistry 33 (1994) 13363-72.

89. I.C. Smith, R. Deslauriers, H. Saito, R. Walter, C. Garrigou-Lagrange, H. McGregor, D. Sarantakis, Carbon-13 NMR studies of

31

2

peptide hormones and their components, Ann. N Y Acad. Sci. 222 (1973) 597-627.

90. M. Kamberi, P. Chung, R. DeVas, L. Li, Z. Li, X.S. Ma, S. Fields, C.M. Riley, Analysis of non-covalent aggregation of synthetic hPTH (1-34) by size-exclusion chromatography and the importance of suppression of non-specific interactions for a precise quantitation, J. Chromatogr. B. Analyt Technol. Biomed. Life. Sci. 810 (2004) 151-5.

91. W. Wang, Protein aggregation and its inhibition in biopharmaceutics, Int. J. Pharm. 289 (2005) 1-30.

92. M.M. Tan, C.A. Corley, C.L. Stevenson, Effect of gelation on the chemical stability and conformation of leuprolide, Pharm. Res. 15 (1998) 1442-8.

93. G.A. Brazeau, B. Cooper, K.A. Svetic, C.L. Smith, P. Gupta, Current perspectives on pain upon injection of drugs, J. Pharm. Sci. 87 (1998) 667-77.

94. Extemp.ie, Section 2: Sterile Pharmaceuticals - Parenterals, accessed April, 01 2012.

95. L. Jorgensen, S. Hostrup, E.H. Moeller, H. Grohganz, Recent trends in stabilising peptides and proteins in pharmaceutical formulation - considerations in the choice of excipients, Expert Opin. Drug Deliv. 6 (2009) 1219-30.

96. J.L. Cleland, M.F. Powell, S.J. Shire, The development of stable protein formulations: a close look at protein aggregation, deamidation, and oxidation, Crit. Rev. Ther. Drug Carrier Syst. 10 (1993) 307-77.

97. M. Cholewinski, B. Lückel, H. Horn, Degradation pathways, analytical characterization, and formulation strategies of a peptide and a protein: calcitonin and human growth hormone in comparison, Pharm. Acta Helv. 71 (1996) 405-19.

98. N.E. Good, G.D. Winget, W. Winter, T.N. Connolly, S. Izawa, R.M. Singh, Hydrogen ion buffers for biological research, Biochemistry 5 (1966) 467-77.

99. R. Li, M.J. Hageman, E.M. Topp, Effect of viscosity on the deamidation rate of a model Asn-hexapeptide, J. Pept. Res. 59 (2002) 211-20.

100. R. Li, A.J. D’Souza, B.B. Laird, R.L. Schowen, R.T. Borchardt, E.M. Topp, Effects of solution polarity and viscosity on peptide deamidation. 2000 Nov;56(5):326-334., J Pept Res 56 (2000) 326.

101. K.C. Waterman, R.C. Adami, K.M. Alsante, J. Hong, M.S. Landis, F. Lombardo, C.J. Roberts, Stabilization of pharmaceuticals to oxidative degradation, Pharm. Dev. Technol. 7 (2002) 1-32.

102. S. Li, T.W. Patapoff, T.H. Nguyen, R.T. Borchardt, Inhibitory effect of sugars and polyols on the metal-catalyzed oxidation of human relaxin, J. Pharm. Sci. 85 (1996) 868-72.

103. L.P. Stratton, R.M. Kelly, J. Rowe, J.E. Shively, D.D. Smith, J.F. Carpenter, M.C. Manning, Controlling deamidation rates in a model peptide: Effects of temperature, peptide concentration, and additives, J. Pharm. Sci. 90 (2001) 2141-8.

104. C. Avanti, J.P. Amorij, D. Setyaningsih, A. Hawe, W. Jiskoot, J. Visser, A. Kedrov, A.J. Driessen, W.L. Hinrichs, H.W. Frijlink, A new strategy to stabilize oxytocin in aqueous solutions: I. The effects of divalent metal ions and citrate buffer, AAPS J. 13 (2011) 284-90.

105. W. Wang, S. Martin-Moe, C. Pan, L. Musza, Y.J. Wang, Stabilization of a polypeptide in non-aqueous solvents, Int. J. Pharm. 351 (2008) 1-7.

106. Y.C.J. Wang, M.A. Hanson, Parenteral formulations of proteins and peptides: stability and stabilizers, J. Parenter. Sci. Technol. 42 (1987) S2-S26.

107. K. Zheng, C.R. Middaugh, T.J. Siahaan, Evaluation of the physical stability of the EC5 domain of E-cadherin: effects of pH, temperature, ionic strength, and disulfide bonds, J. Pharm. Sci. 98 (2009) 63-73.

108. S.A. Bursakov, C. Carneiro, M.J. Almendra, R.O. Duarte, J. Caldeira, I. Moura, J.J. Moura, Enzymatic properties and effect of ionic strength on periplasmic nitrate reductase (NAP) from Desulfovibrio desulfuricans ATCC 27774, Biochem. Biophys. Res. Commun. 239 (1997) 816-22.

109. P. O’Neill, S. Davies, E.M. Fielden, L. Calabrese, C. Capo, F. Marmocchi, G. Natoli, G. Rotilio, The effects of pH and various salts upon the activity of a series of superoxide dismutases, Biochem. J. 251 (1988) 41-6.

110. R. Tyler-Cross, V. Schirch, Effects of amino acid sequence, buffers, and ionic strength on the rate and mechanism of deamidation of asparagine residues in small peptides, J. Biol. Chem. 266 (1991) 22549-56.

111. K. Neelon, H.J. Schreier, H. Meekins, P.M. Robinson, M.F. Roberts, Compatible solute effects on thermostability of glutamine synthetase and aspartate transcarbamoylase from Methanococcus jannaschii, Biochim. Biophys. Acta 1753 (2005) 164-73.

112. T.V. Brennan, S. Clarke, Spontaneous degradation of polypeptides at aspartyl and asparaginyl residues: effects of the solvent dielectric, Protein Sci. 2 (1993) 331-8.

32

2


113. D.A. Parkins, U.T. Lashmar, The Formulation of Biopharmaceutical Products, Pharm. Sci. Technol. Today 3 (2000) 129-37.

114. S. Li, C. Schoneich, R.T. Borchardt, Chemical instability of protein pharmaceuticals: Mechanisms of oxidation and strategies for stabilization, Biotechnol. Bioeng. 48 (1995) 490-500.

115. S. Landi, H.R. Held, Effect of oxidation on the stability of tuberculin purified protein derivative (PPD), Dev. Biol. Stand. 58 (1986) 545-52.

116. C. Schoneich, F. Zhao, G.S. Wilson, R.T. Borchardt, Iron-thiolate-induced oxidation of methionine to methionine sulfoxide in small model peptides. Catalysis by histidine., Biochim. Biophys. Acta 1158 (1993) 307-22.

117. S. Li, C. Schoneich, G.S. Wilson, R.T. Borchardt, Chemical pathways of peptide degradation. V. Ascorbic acid promotes rather than inhibits the oxidation of methionine to methionine sulfoxide in small model peptides, Pharm. Res. 10 (1993) 1572-9.

118. R.L. Levine, J. Moskovitz, E.R. Stadtman, Oxidation of methionine in proteins: roles in antioxidant defense and cellular regulation, IUBMB Life 50 (2000) 301-7.

119. P.K. Tsai, D.B. Volkin, J.M. Dabora, K.C. Thompson, M.W. Bruner, J.O. Gress, B. Matuszewska, M. Keogan, J.V. Bondi, C.R. Middaugh, Formulation design of acidic fibroblast growth factor, Pharm. Res. 10 (1993) 649-59.

120. B.P. Obiols, G.J. Farres, F.J.C. Rodriguez, S.P. Fernandez, J.B. Cabado, Stable pharmaceutical formulation for intravenous or intramuscular administration of active peptide compound (2003) March, 20, 2012.

121. T. Arakawa, S.J. Prestrelski, W.C. Kenney, J.F. Carpenter, Factors affecting short-term and long-term stabilities of proteins, Adv. Drug Deliv. Rev. 10 (1993) 1-28.

122. T. Matsuoka, S. Tomita, H. Hamada, K. Shiraki, Amidated amino acids are prominent additives for preventing heat-induced aggregation of lysozyme, J. Biosci. Bioeng. 103 (2007) 440-3.

123. R. Quinn, J.D. Andrade, Minimizing the aggregation of neutral insulin solutions, J. Pharm. Sci. 72 (1983) 1472-3.

124. D.K. Chou, R. Krishnamurthy, T.W. Randolph, J.F. Carpenter, M.C. Manning, Effects of Tween 20 and Tween 80 on the

stability of Albutropin during agitation, J. Pharm. Sci. 94 (2005) 1368-81.

125. A. Lahlou, B. Blanchet, M. Carvalho, P. P M., A. Astier, Mechanically-induced aggregation of the monoclonal antibody cetuximab, Ann. Pharm. Fr. 67 (2009) 340-52.

126. G. Lentzen, T. Schwarz, Extremolytes: Natural compounds from extremophiles for versatile applications, Appl. Microbiol. Biotechnol. 72 (2006) 623-34.

127. I. Yu, Y. Jindo, M. Nagaoka, Microscopic understanding of preferential exclusion of compatible solute ectoine: direct interaction and hydration alteration, J. Phys. Chem. B 111 (2007) 10231-8.

128. S. Knapp, R. Ladenstein, E.A. Galinski, Extrinsic protein stabilization by the naturally occurring osmolytes beta-hydroxyectoine and betaine, Extremophiles 3 (1999) 191-8.

129. A. Hedoux, J.F. Willart, L. Paccou, Y. Guinet, F. Affouard, A. Lerbret, M. Descamps, Thermostabilization mechanism of bovine serum albumin by trehalose, J. Phys. Chem. B 113 (2009) 6119-26.

130. T.Q. Faria, S. Knapp, R. Ladenstein, A.L. Macanita, H. Santos, Protein stabilisation by compatible solutes: effect of mannosylglycerate on unfolding thermodynamics and activity of ribonuclease A, ChemBioChem 4 (2003) 734-41.

131. J. Ryu, M. Kanapathipillai, G. Lentzen, C.B. Park, Inhibition of β-amyloid peptide aggregation and neurotoxicity by α-d-mannosylglycerate, a natural extremolyte, Peptides 29 (2008) 578-84.

132. E.A. Galinski, M. Stein, B. Amendt, M. Kinder, The kosmotropic (structure-forming) effect of compensatory solutes, Comp. Biochem. Physiol. 117A (1997) 357-65.

133. T. Arakawa, S.N. Timasheff, The stabilization of proteins by osmolytes, Biophys. J. 47 (1985) 411-4.

134. D.W. Bolen, G.D. Rose, Structure and Energetics of the Hydrogen-Bonded Backbone in Protein Folding, Annu. Rev. Biochem. 77 (2008) 339-62.

135. N. Borges, A. Ramos, N.D. Raven, R.J. Sharp, H. Santos, Comparative study of the thermostabilizing properties of mannosylglycerate and other compatible solutes on model enzymes, Extremophiles 6 (2002) 209-16.

33

Christina Avanti1, Jean-Pierre Amorij1, Dewi Setyaningsih1, Andrea Hawe4, Wim Jiskoot4, Jan Visser2, Alexej Kedrov3, Arnold J. M. Driessen3,

Wouter L. J. Hinrichs1, and Henderik W. Frijlink1

1 Department of Pharmaceutical Technology and Biopharmacy, 2 Department of Pharmacokinetics, Toxicology, and Targeting,

3 Department of Molecular Microbiology, University of Groningen, Groningen, The Netherlands.

4 Division of Drug Delivery Technology, Leiden/Amsterdam Center for Drug Research, Leiden University, Leiden, The Netherlands.

3

a new strategy to stabIlIze oxytocIn In aqueous solutIons: I. the effects of dIvalent metal Ions and cItrate buffer

AAP S Journal 2011; 13(2): 284–290

3

abstract In the current study, the effect of metal ions in combination with buffers (citrate, acetate, pH 4.5) on the stability of aqueous solutions of oxytocin was investigated. Both monovalent metal ions (Na+ and K+) and divalent metal ions (Ca2+, Mg2+, and Zn2+) were tested all as chloride salts. The effect of combinations of buffers and metal ions on the stability of aqueous oxytocin solutions was determined by RP-HPLC and HP-SEC after 4 weeks of storage at either 4°C or 55°C. Addition of sodium or potassium ions to acetate- or citrate-buffered solutions did not increase stability, nor did the addition of divalent metal ions to acetate buffer. However, the stability of aqueous oxytocin in aqueous formulations was improved in the presence of 5 and 10 mM citrate buffer in combination with at least 2 mM CaCl2, MgCl2, or ZnCl2 and depended on the divalent metal ion concentration. Isothermal titration calorimetric measurements were predictive for the stabilization effects observed during the stability study. Formulations in citrate buffer that had an improved stability displayed a strong interaction between oxytocin and Ca2+, Mg2+, or Zn2+, while formulations in acetate buffer did not. In conclusion, our study shows that divalent metal ions in combination with citrate buffer strongly improved the stability of oxytocin in aqueous solutions.

3

StabIlIty of the dIvalent Metal-cItrate-oxytocIn coMplex

1. IntroductIonAccording to the World Health Organization, half a million of women in Africa, Asia, and Latin America die each year due to problems during pregnancy and childbirth. At least 25% of those deaths can be attributed to bleeding after child birth (post-partum hemorrhage), mainly caused by failure of the uterus to contract adequately after child birth (atonicity) [1].

The preferred drug to prevent post-partum hemorrhage is oxytocin. Oxytocin is a cyclic nonapeptide hormone [sequence: cyclo (Cys1-Tyr2-Ile3-Gln4-Asn5-Cys6), -Pro7-Leu8-Gly9-NH2],which is naturally produced in the hypothalamus. It is involved primarily in uterine contraction and stimulation of milk release from the mammary tissue [2]. Oxytocin, which is currently available in synthetic form [3], has been widely used for indications such as induction of labor, augmentation of labor, post-partum hemorrhage, or uterine atony, and also for other indications such as diabetes insipidus and vasodilatory shock. Reported additional functions for oxytocin include an antiduretic effect and blood vessel contraction [2,4,5].

Unfortunately, oxytocin preparations are highly unstable at elevated temperatures, which is an issue particularly in tropical countries [6]. Stability studies conducted by Groot et al. [6] have shown that injectable oxytocin formulations are rapidly degraded as the storage temperature rises to 30°C or higher. Oxytocic tablets for oral administration (ergometrine, ethylergometrine, oxytocin, and desamino-oxytocin) are also not stable under simulated tropical conditions. Because of its poor stability at elevated temperatures, the use of oxytocin in many developing countries is limited. Thus, there is a clear need for a heat-stable oxytocin formulation, preferably an aqueous injectable solution, with improved thermal stability.

One way to effectively improve stability of several peptides in aqueous solution is using metal salts in combination with a suitable buffer [7]. To investigate the effect of metal ions in buffered solutions on the stability of oxytocin, we screened various combinations of unbuffered and buffered solutions with monovalent or divalent metal ions. Hawe et al. [8] observed that the degradation of oxytocin strongly depends on the pH of the formulation, with the highest stability at pH 4.5. Therefore, all formulations will be set to pH 4.5. The purpose of this study is to investigate whether specific combinations of buffer and metal ions can stabilize oxytocin.

2. materIals and method2.1 MaterialsThe following materials were used in this study: oxytocin monoacetate powder (Diosynth, Oss, The Netherlands), citric acid, calcium chloride (Riedel-de Haen, Seelze, Germany), acetic acid, magnesium chloride, zinc chloride (Fluka, Steinheim, Germany), sodium hydroxide, sodium chloride, potassium chloride, sodium dihydrogen phosphate dihydrate, acetonitrile, formic acid (Merck, Darmstadt, Germany) and Baxter Viavlo Ringer’s lactate solution for intravenous infusion (Baxter, Utrecht, The Netherlands).

2.2 Formulation and Stability StudyOxytocin was formulated at a concentration of 0.1 mg/ml in citrate (5 or 10 mM) or acetate (10 mM) buffer at pH 4.5 (pH adjusted with sodium hydroxide) with different additions of metal ions. pH samples were controlled and remained within ±0.1 pH units

37

3

during the stability study. The initial concentration of oxytocin was determined using UV spectrophotometry [9] at 280 nm with an extinction coefficient of 1.52 ml mg−1 cm−1. All metal ion solutions were prepared using their chloride salts at concentrations of 2, 5, 10, and 50 mM. Control solutions were formulated in water (pH 6.9±0.2) and Ringer’s lactate (6.4±0.2). Ringer’s lactate solution consists of 131 mM sodium, 5 mM potassium, 2 mM calcium, 111 mM chloride, and 29 mM bicarbonate (as lactate). In this report, the following codes were used: First character(s) refer to the type of buffer or water; CB (citrate), AC (acetate), RL (Ringer’s lactate), and W (water). Following digit(s) refer to buffer concentration in mM, following character(s) to the type of metal ion, and last digit(s) to metal ion concentration in mM. Thus, e.g., CB10Mg10 means 10 mM citrate buffer (pH 4.5) and 10 mM MgCl2. After preparation, the solutions were stored in 6R glass type 1 vials for 4 weeks at either 4°C or 55°C, and protected from light.

Based on the results of the screening study, oxytocin formulations in 10 mM citrate buffer pH 4.5 with 10 or 50 mM divalent metal salts were selected for a longer period of stability study for 6 months at 40°C according to ICH guidelines for long-term and accelerated stability study for climatic zone III and IV [10].

2.3 Reversed-Phase High-Performance Liquid ChromatographyThe recovery of oxytocin (remaining oxytocin as percentage of initial amount) was determined by RP-HPLC. RPHPLC was performed according to the procedure described by Hawe et al. [8] An Alltima C-18 RP column with 5 μm particle size, inner diameter of 4.6 mm, and length of 150 mm (Alltech, Ridderkerk, Netherlands), a Waters (Millipore) 680 Automated Gradient Controller, two Waters 510 HPLC pumps, a Waters 717 Plus Autosampler, and a Waters 486 Tunable Absorbance UV Detector were used. Samples of 20 μl were injected and the separation was carried out at a flow rate of 1.0 ml/min and UV detection at 220 nm. Samples were eluted using 15% (v/v) acetonitrile in 65 mM phosphate buffer pH 5.0 as solvent A and 60% (v/v) acetonitrile in 65 mM phosphate buffer pH 5.0 as solvent B. The acetonitrile concentration was linearly increased from 15% at the beginning, to 20% at 10 min, to 30% at 20 min, and finally to 60% at 25 min.

2.4 Size Exclusion HPLCThe fraction of monomeric oxytocin (percentage of total remaining oxytocin) was assessed by Size Exclusion HPLC (HP-SEC). HP-SEC was carried out using a Superdex peptide 10/300 GL column (GE Healthcare Inc., Brussels, Belgium) on an isocratic HPLC system, according to the method previously reported by Hawe et al. [8]. A Waters 510 pump, a Waters 717 plus auto sampler, a Waters 474 Scanning Fluorescence Detector and Waters 484 Tunable Absorbance Detector (Waters, Milford Massachusetts, USA) were used. Samples of 50 μl were injected, and separation was performed at a flow rate of 1 ml/min. Peaks were detected by UV absorption at 274 nm, as well as fluorescence detection at excitation wavelength of 274 nm and emission wavelength of 310 nm. The mobile phase consisted of 30% acetonitrile and 70% 0.04 M formic acid.

2.5 Isothermal Titration CalorimetryIsothermal titration calorimetry (ITC) was used to investigate the interaction between oxytocin and divalent metal ions in the presence of citrate buffer and acetate buffer. Microcalorimetric titrations of divalent metal ions to oxytocin were conducted by using a

38

3


MicroCal ITC 200 Microcalorimeter (Northampton, MA 01060 USA). A solution of 300 μL of 5 mM oxytocin in 10 mM of either citrate or acetate pH 4.5 was placed in the sample cell, while 30 μL of 125 mM divalent metal chloride either calcium, magnesium, or zinc in 10 mM citrate or acetate buffer pH 4.5 was placed in the syringe. The reference cell contained 300 μL of the corresponding buffer. Experiments were performed at 55°C. Automated titrations were conducted up to a divalent metal ion/oxytocin molar ratio of 5:1. The effective heat of the peptide-metal ion interaction upon each titration step was corrected for dilution and mixing effects, as measured by titrating the divalent metal ion solution into buffer and by titrating buffer into oxytocin solution. To investigate the possibility of oxytocin or metal ion binding to the buffer components, control experiments were performed in water. The heats of bimolecular interactions were obtained by integrating the peak following each injection. All measurements were performed in triplicate.

ITC data were analyzed by using the ITC non-linear curve fitting functions for one or two binding sites from MicroCal Origin 7.0 software (MicroCal Software, Inc.). The calculated curve was determined by the best-fit parameter, which was used to determine the molar enthalpy change for binding and the corresponding association constant, Ka. The molar free energy of binding ΔG° and the molar entropy change ΔS° were derived from the fundamental equations of thermodynamics ΔG° = -RTlnKa and ΔG° = ΔH°- TΔS°.

3. results3.1 Influence of Divalent Metal Ions on Oxytocin Stability in Unbuffered SolutionsFirst, the effect of divalent metal ions on the stability of oxytocin in water without any buffer salt was investigated. RP-HPLC results (Fig. 1a) showed that after 4 weeks of storage at 4°C, oxytocin recovery was almost 100% in the presence of 2–50 mM zinc or 50 mM calcium ions. No stabilizing effect was observed from the presence of magnesium (2–50 mM) and calcium (2–10 mM), where the recovery was reduced to about 65%, similar to levels found for oxytocin solutions in water. HPSEC results (Fig. 1b) showed a similar trend in the recovery of monomeric oxytocin. However, when the solutions were stored at 55°C, both RP-HPLC and HP-SEC measurements showed substantial degradation of oxytocin after 4 weeks. These results demonstrate that divalent metal ions in non-buffered aqueous oxytocin formulations have only a limited stabilizing effect at elevated temperature.

3.2 Oxytocin Stability in Buffered SolutionsTo determine the effect of buffer on stability, oxytocin was formulated in 5 and 10 mM citrate buffer and 10 mM acetate buffer.As a reference, the stability of oxytocin in pure water and in Ringer’s lactate buffer was investigated. Figure 2a shows the oxytocin recovery in RP-HPLC after 4 weeks of storage either at 4°C or 55°C in the buffered solutions. Compared to pure water, the stability of oxytocin was substantially increased at 4°C in the presence of the buffer salts. After storage at 55°C, the recovery of oxytocin after 4 weeks in citrate and acetate buffer was much higher as compared to water or Ringer’s lactate solution. However, the recovery of oxytocin was still poor. In addition, only about 20% of oxytocin remained in its monomeric form after storage (Fig. 2b).

39

3

3.3 Oxytocin Stability in Buffered Solutions Containing Monovalent Metal IonsTo investigate the stability of oxytocin in the presence of a combination of buffer and monovalent metal ions, citrate and acetate buffer were used at a concentration of 10 mM, in combination with the monovalent metal ions, sodium and potassium, added at a concentration of 10 and 20 mM (excluding sodium from the buffer component). The

Figure 1. Recovery of oxytocin in the presence of divalent metal ions in non-buffered pure water, stored for 4 weeks at pH 4.5 and a temperature of 4°C (light gray bars) or 55°C (dark gray bars). The divalent metal ions (Ca2+, Mg2+, and Zn2+) were used in concentrations of 2, 5, 10, and 50 mM. a recovery determined by RP-HPLC. b oxytocin monomer recovery determined by HP-SEC. The results are depicted as averages of three independent measurements±SD

40

3


presence of monovalent metal ions had only a minor effect on the stability of oxytocin (stored at 55°C). Although the oxytocin recovery was slightly improved compared to oxytocin in the presence of buffer alone, the maximum recovery of oxytocin was only 35% in 10 mM acetate buffer with 10 mM sodium chloride. In addition, only about 30% of oxytocin in this formulation remained monomeric (data not shown). These results clearly indicate that the presence of a combination of buffer and monovalent metal ions is not sufficient to substantially stabilize oxytocin in aqueous solution.

Figure 2. Recovery of oxytocin in pure water, with or without a buffer. Citrate buffer at a concentration of 5, 10, or 50 mM, acetate buffer at a concentration of 10 mM, or Ringer’s lactate solution were used. The formulations contained no metal ions and were stored for 4 weeks at pH 4.5 or 6.4 for Ringer’s lactate solution at a temperature of 4°C (light gray bars) or 55°C (dark gray bars). a recovery determined by RP-HPLC. b oxytocin monomer recovery determined by HP-SEC. The results are depicted as averages of three independent measurements±SD

41

3

3.4 Oxytocin Stability in Citrate-Buffered Solutions Containing Divalent Metal Ions To study the stability of oxytocin in the presence of citrate buffers and divalent metal ions, formulations containing 5 and 10 mM citrate buffer in combination with divalent metal ions (calcium, magnesium, and zinc) added at concentrations of 2, 5, 10, or 50 mM were used. The results of RP-HPLC and HP-SEC of formulations in 10 mM citrate buffer are presented in Fig. 3a and b. The stability of oxytocin solutions was clearly improved when formulating them with citrate buffer in combination with calcium ions. The oxytocin stability increased

Figure 3. Effect of Ca2+ (squares), Mg2+ (circles), and Zn2+ (triangles) concentration on the recovery of oxytocin in citrate buffer at the concentration of 10 mM after 4 weeks of storage at either 4°C or 55°C and pH 4.5. Solid symbols denoted 4°C storage, while open symbols correspond to 55°C. a recovery determined by RP-HPLC. b oxytocin monomer recovery determined by HP-SEC. The results are depicted as averages of three independent measurements±SD

42

3


with increasing calcium ion concentrations. The recovery of oxytocin and the remaining percentage of oxytocin monomers after 4 weeks of storage at 55°C were increased up to almost 80% in the presence of 50 mM calcium.

Similar results were obtained for the combination of citrate with magnesium. The degradation of oxytocin in citrate buffer at 5 and 10 mM decreased with an increasing concentration of magnesium ions. Formulations with zinc ions in citrate buffer also preserved oxytocin during storage. These combinations exert even a stronger effect on oxytocin stability than combinations of citrate buffer and calcium or magnesium ions. Stability was strongly improved at zinc concentrations as low as 2 or 5 mM. Both oxytocin recovery (RP-HPLC) and the monomeric oxytocin fraction (HP-SEC) were substantially higher (up to 90%) in the presence of 10 mM zinc ions (CB5Zn10) after storage for 4 weeks at 55°C. When citrate buffer was used at a concentration of 5 mM, similar results were found (data not shown).

Beside citrate, we also carried out several further experiments to investigate whether divalent metal ions affect the stability in the presence of acetate buffer. Acetate buffer at a concentration of 10 mM was used with and without calcium, magnesium, and zinc ions at various concentrations (2, 5, 10, 50 mM). The combination of these ions with acetate buffer was found to be less efficient in stabilizing oxytocin (data not shown).

3.5 Long-term Stability of Oxytocin in Selected Formulations Containing Citrate and Divalent Metal IonsFor the combination of citrate with Ca2+, Zn2+, and Mg2+, a long-term stability study for 6 months at 40°C was conducted. A temperature of 40°C was chosen to simulate tropical conditions [10,11]. The long-term stability study at 40°C clearly demonstrates the synergistic stabilizing effect of citrate buffer and divalent metal ions. Even though the oxytocin recovery decreased gradually with time, the recovery of oxytocin (Fig. 4a) and the remaining percentage of oxytocin monomers (Fig. 4b) after 6 months storage at 40°C were increased up to 80%in the presence of 50 mM calcium, and even higher (up to 90%) in the presence of 50 mM magnesium.

Formulations with 10 mM zinc ions in citrate buffer exerted the same effect on oxytocin stability as combinations of citrate buffer and 50 mM magnesium ions. This shows that zinc ions at lower concentrations have already a higher impact on increasing oxytocin stability compared with calcium or magnesium ions.

This result also confirms the short-term stability study that showed up to 90%remaining oxytocin in the presence of 10mM zinc ions after storage for 4 weeks at 55°C.

3.6 ITC to Study the Interaction between Oxytocin and Divalent Metal IonsTo examine the interaction between oxytocin and the metal ions, ITC experiments were carried out that are summarized in Table 1. The titration of calcium ions into an oxytocin solution in citrate buffer resulted in an exothermic reaction (Fig. 5a) with a Ka value of about 400 M−1 and an apparent ion/oxytocin stoichiometry close to 4:1 (Table 1).

Remarkably, when titrating magnesium into oxytocin we observed heat absorption (Fig. 5a) and this endothermic reaction occurred with an identical apparent stochiometry (4:1) and a similar Ka value of about 200 M−1 as with calcium ions. However, with magnesium ions the ITC trace was complex and showed an exothermic phase at magnesium concentrations below 5 mM. Due to

43

3

the rapid saturation of this early phase—typically, within three injections—and the low enthalpy of the ion/oxytocin interaction, it was not possible to quantitatively analyze the exothermic stage. To analyze the following endothermic phase we omitted the first 4-titration steps and fitted the remaining data using a single site model. A similar dual-phase response was observed for the zinc: oxytocin interaction (Fig. 5a, open triangles), but the exothermic and endothermic stages were well-resolved and suitable for the analysis using a two sites model (Table 1). In contrast, in the presence of acetate buffer there was no measurable interaction between oxytocin and calcium, magnesium, or zinc ion (Fig. 5b).

Figure 4. Oxytocin recovery over time storage at 40°C and pH 4.5 in the presence of 10 mM citrate buffer, without (star) and with divalent metal ions. Ca2+ (square), Mg2+ (triangle), and Zn2+ (circle) were used in concentrations of 10 mM (open symbols), and 50 mM (solid symbols). a recovery determined by RP-HPLC. b oxytocin monomer recovery determined by HP-SEC. The results are depicted as averages of three independent measurements±SD

44

3


4. dIscussIonOur study clearly demonstrates that the stability of oxytocin in aqueous formulations is greatly increased in citrate buffer in combination with divalent metal ions. The improved stability was found to be dependent on the divalent metal ions concentration.

The WHO reported that there is no loss of potency of oxytocin in injection preparation after 12 months refrigerated storage (2–8°C). However, oxytocin lost 14% of its potency after 1 year at 30°C [12]. In another study, the oxytocin concentration in Ringer’s lactate solution was reduced by about 10% after 35 days storage at room temperature (near 23°C) [13]. From our observation Ringer’s lactate solution was able to stabilize oxytocin in aqueous solution at low temperature (4°C). However at a higher temperature (55°C), the stability of Ringer’s lactate solution was poorer than in the presence of citrate or acetate buffers. This result can be attributed to the pH and/or the amount of metal ions in solution. Oxytocin in Ringer’s lactate solution has a pH of 6.4 and contains less than 2 mM of divalent metal ions. Hawe et al. [8] observed that the degradation of oxytocin strongly depends on the pH of the formulation, with the highest stability at pH 4.5. Therefore, a pH of 6.4 might have caused an increased rate of decomposition. However, formulating oxytocin in acetate buffer at pH 4.5 only maintain approximately 30% oxytocin recovered after 1 month storage at 55°C.

The decomposition of oxytocin is mainly caused by deamidation, oxidation, hydrolysis, and dimerization [7,8]. Deamidation is likely to occur in Gln4 [14], Asn5 [15], and Gly9 [7,16]. Under acidic condition (pH below 3), deamidation of Asn5 and Gln4 occur by direct hydrolysis [7]. Oxidation might occur at Tyr2 [17] and Cys1,6 [18], whereas dimerization might occur due to thiol exchange in Cys1,6 [19].

Although the specific mechanism has not been elucidate yet, the presence of sufficient amounts of divalent metal ions at pH 4.5 in citrate buffer, however, greatly improved the stability of oxytocin in aqueous solution. In previous studies, the interaction of oxytocin with calcium [20] and zinc [21] ions was investigated using NMR and nanoelectrospray mass spectrometry (MS). Those studies which were also supported with molecular modeling, showed that Ca2+ is coordinated by seven carbonyl oxygen atoms (O-Tyr2, O-Ile3, O-Gln4, O-Asp5, O-Cys6, O-Leu8, and O-Gly9) which formed a more compact structure for the oxytocin–Ca2+ complex compared to free oxytocin [20]. Whereas zinc ions formed an octahedral complex with six of the backbone carbonyl oxygen atoms (O-Tyr2,O-Ile3, O-Gln4, O-Cys6, O-Leu8, and O-Gly9) [21]. It was suggested that in the presence of such divalent ions, the hydrophobic groups are situated inside the peptide keeping them away

Table 1. Thermodynamics of Divalent Metal binding to Oxytocin as Determined by Isothermal Titration Calorimetry in 10 mM Citrate Buffer

Metal Phase N (sites) Ka (M−1) ΔH°(cal mol−1) ΔS° (cal/mol/deg)

Ca2+ 1 0.26 400 −2,100 5.6Mg2+ 1 n.d. n.d. <0 n.d.

2 0.23 220 4,400 24Zn2+ 1 0.78 2,700 −800 13

2 1.48 800 610 15

The results are depicted as averages of three independent measurements with relative standard deviations below 10% n.d. not determined.

45

3

from water molecules. These conformational changes can increase the stability of oxytocin in aqueous medium as they will prevent dimerization and further aggregation by hydrophobic interactions among oxytocin molecules. Metal salts are often used to stabilize peptides or proteins by chelation or ionic interactions [22]. Wang et al. examined the peptide (P66) stability in the presence of ZnCl2, MgCl2, and CaCl2 in non aqueous solution, and found that in the presence of 1 mM ZnCl2, P66 was significantly stabilized. However in the aqueous solution (pure water), these ions did not show any stabilizing effect [22]. In our experiments, the addition of calcium, magnesium, and zinc ions in combination with citrate buffer had a large impact on oxytocin stability in contrast to similar experiments in pure water or acetate buffer. This study suggests that there is a synergistic effect between citrate buffer and the divalent metal ions, possibly due to the protection of the disulfide bridge by complex formation of divalent metal ion and citrate with oxytocin which suppressed intermolecular reaction leading to tri/tetrasulfide formation as well as dimerization (unpublished data).

ITC is a sensitive method for studying the thermodynamics of binding events and quantifying binding reactions. When divalent metal ion are added to oxytocin, the ITC data indicate an interaction between oxytocin and Ca2+, Mg2+, or Zn2+ ions in the presence of citrate buffer. Both Mg2+ and Zn2+ ions demonstrated complex, dual-phase interaction profile, while a single phase was observed for Ca2+. Each interaction was entropy driven, while both exothermic and endothermic reactions were observed. It may be speculated that the solvation effect, i.e., release of structured water molecules plays a key role in binding, while the specific ion–oxytocin interaction further contributes to the complex stability. The latter is also predicted by molecular dynamic simulations [20,21]. Remarkably, no

Figure 5. Least squares fit of the data from calorimetric titration profiles of aliquots of 125 mM divalent metal ions: Ca2+ (solid square), Mg2+ (open square), and Zn2+ (open triangle) into 5 mM oxytocin in 10 mM a citrate buffer and b acetate buffer pH 4.5. The heat absorbed per mol of titrant is plotted versus the ratio of the total concentration of divalent metal ions to the total concentration of oxytocin

a b

46

3


interaction between oxytocin and either of the tested ions was detected in the acetate buffer or deionized water. These observations underscore the role of a particular environment in the ion–oxytocin interaction and agree well with our findings on the peptide stability. Isothermal titration calorimetric measurements were predictive for the effects observed during the stability study.

In conclusion, this study shows that with a combination of divalent metal salts and citrate buffer, the stability of oxytocin in aqueous solution can be strongly improved. The increased stability of oxytocin aqueous formulations was achieved in the presence of citrate acid buffer and 2 mM or more of the salts CaCl2, MgCl2, or ZnCl2. The oxytocin stability is further increased with increasing concentration of the divalent metals ions up to 50 mM. In combination with citrate buffer, Zn2+ has a superior stabilizing effect as compared with Ca2+ or Mg2+.

acknowledgmentsThe authors want to thank MSD Oss for providing oxytocin for the study. This study was performed within the framework of the Dutch Top Institute Pharma project: number D6–202. Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

references1. K.S. Khan, D. Wojdyla, L. Say, A.M.

Gulmezoglu, P.F. Van Look, WHO analysis of causes of maternal death: a systematic review, Lancet 367 (2006) 1066-74.

2. V.S. Ananthanarayanan, K.S. Brimble, Interaction of oxytocin with Ca2+: I. CD and fluorescence spectral characterization and comparison with vasopressin, Biopolymers 40 (1996) 433-43.

3. E.H. Bishop, Synthetic oxytocin; a clinical evaluation, Obstet Gynecol 11 (1958) 290-4.

4. K.P. Conrad, M. Gellai, W.G. North, H. Valtin, Influence of oxytocin on renal hemodynamics and electrolyte and water excretion, Am. J. Physiol. 251 (1986) F290-6.

5. A.V. Somlyo, C.Y. Woo, A.P. Somlyo, Responses of Nerve-Free Vessels to Vasoactive Amines and Polypeptides, Am J Physiol 208 (1965) 748-53.

6. A.N.J.A.D. Groot, T.B. Vree, H.V. Hogerzeil, G.J.A. Walker, WHO Action Programme on Essential Drugs: Stability of Oral Oxytocics in Tropical Climates : Results of Simulation Studies on Oral Ergometrine, Oral Methylergometrine, Buccal Oxytocin and Buccal Desamino-Oxytocin, World Health Organization, Geneva, 1994.



9. S.C. Gill, P.H. von Hippel, Calculation of protein extinction coefficients from amino acid sequence data, Anal. Biochem. 182 (1989) 319-26.

10. International Conference on Harmonisation; Stability Data Package for Registration Applications in Climatic Zones III and IV; Stability Testing of New Drug Substances and Products; availability. Notice, Fed. Regist. 68 (2003) 65717-8.

11. W. Grimm, Extension of the International Conference on Harmonization Tripartite Guideline for Stability Testing of New Drug Substances and Products to countries of climatic zones III and IV, Drug Dev. Ind. Pharm. 24 (1998) 313-25.


13. L.A. Trissel, Y. Zhang, K. Douglass, E. Kastango, Extended Stability of Oxytocin in common infusion solution, International Journal of Pharmaceutical Compounding 10 (2006) 156-8.

47

3

14. A.B. Joshi, M. Sawai, W.R. Kearney, L.E. Kirsch, Studies on the mechanism of aspartic acid cleavage and glutamine deamidation in the acidic degradation of glucagon, J. Pharm. Sci. 94 (2005) 1912-27.

15. H. Yang, R.A. Zubarev, Mass spectrometric analysis of asparagine deamidation and aspartate isomerization in polypeptides, Electrophoresis 31 (2010) 1764-72.

16. N.E. Robinson, Protein deamidation, Procl. Natl. Acad. Sci. 99 (2002) 5283-8.

17. C. Leeuwenburgh, J.E. Rasmussen, F.F. Hsu, D.M. Mueller, S. Pennathur, J.W. Heinecke, Mass spectrometric quantification of markers for protein oxidation by tyrosyl radical, copper, and hydroxyl radical in low density lipoprotein isolated from human atherosclerotic plaques, J. Biol. Chem. 272 (1997) 3520-6.

18. A. Fiser, I. Simon, Predicting the oxidation state of cysteines by multiple sequence alignment, Bioinformatics 16 (2000) 251-6.

19. M. Klingenberg, M. Appel, The uncoupling protein dimer can form a disulfide cross-link between the mobile C-terminal SH groups, Eur J Biochem 180 (1989) 123-31.

20. V.S. Ananthanarayanan, M.P. Belciug, B.S. Zhorov, Interaction of oxytocin with Ca2+: II. Proton magnetic resonance and molecular modeling studies of conformations of the hormone and its Ca2+ complex, Biopolymers 40 (1996) 445-64.

21. D. Liu, A.B. Seuthe, O.T. Ehrler, X. Zhang, T. Wyttenbach, J.F. Hsu, M.T. Bowers, Oxytocin-receptor binding: why divalent metals are essential, J. Am. Chem. Soc. 127 (2005) 2024-5.

22. W. Wang, S. Martin-Moe, C. Pan, L. Musza, Y.J. Wang, Stabilization of a polypeptide in non-aqueous solvents, Int. J. Pharm. 351 (2008) 1-7.

48

Christina Avanti1, Hjalmar P. Permentier2, Annie van Dam2, Robert Poole3, Wim Jiskoot3, Henderik W. Frijlink1, and Wouter L. J. Hinrichs1

1Department of Pharmaceutical Technology & Biopharmacy and 2Mass Spectrometry Core Facility,

University of Groningen, Groningen, The Netherlands 3Division of Drug Delivery Technology, Leiden/Amsterdam Center for Drug Research,

Leiden University, Leiden, The Netherlands

a new strategy to stabIlIze oxytocIn In aqueous solutIons: II. suppressIon of cysteIne-medIated Intermolecular

reactIons by a combInatIon of dIvalent metal Ions and cItrate

Mol. Pharmaceutics. 2012 ; 9(3): 554-62

4

4

abstract A series of studies have been conducted to develop a heat-stable liquid oxytocin formulation. Oxytocin degradation products have been identified including citrate adducts formed in a formulation with citrate buffer. In a more recent study we have found that divalent metal salts in combination with citrate buffer strongly stabilize oxytocin in aqueous solutions (Avanti, C.; et al. AAPS J. 2011, 13, 284−290). The aim of the present investigation was to identify various degradation products of oxytocin in citrate-buffered solution after thermal stress at a temperature of 70°C for 5 days and the changes in degradation pattern in the presence of divalent metal ions. Degradation products of oxytocin in the citrate buffer formulation with and without divalent metal ions were analyzed using liquid chromatography−mass spectrometry/ mass spectrometry (LC−MS/MS). In the presence of divalent metal ions, almost all degradation products, in particular citrate adduct, tri- and tetrasulfides, and dimers, were greatly reduced in intensity. No significant difference in the stabilizing effect was found among the divalent metal ions Ca2+, Mg2+, and Zn2+. The suppressed degradation products all involve the cysteine residues. We therefore postulate that cysteine-mediated intermolecular reactions are suppressed by complex formation of the divalent metal ion and citrate with oxytocin, thereby inhibiting the formation of citrate adducts and reactions of the cysteine thiol group in oxytocin.

4

SuppreSSIon of cySteIne-MedIated InterMolecular reactIonS

1. IntroductIon Oxytocin is a neurohypophyseal hormone, which was first discovered by H. H. Dale in 1909 [1,2]. Oxytocin is produced by neurons of the posterior lobe of the hypophysis and pulsatively released into the periphery. In clinical practice, oxytocin has been prescribed primarily for labor induction and augmentation, control of postpartum hemorrhage and uterine hypotonicity in the third stage of labor [3,4]. Oxytocin is commonly administered by intravenous infusion [5].

The oxytocin structure was elucidated in 1951 [6-8] and the characterization and biosynthesis of oxytocin were reported in 1953 by du Vigneaud [9]. Oxytocin consists of nine amino acids: cyclo-(Cys1-Tyr2-Ile3-Gln4-Asn5-Cys6)-Pro7-Leu8-Gly9-NH2 with a disulfide bridge between Cys residues 1 and [6,10,11]. The primary structure of oxytocin is shown in Figure 1. A major problem of the compound is its intrinsic instability in aqueous formulations [12]. Recently significant attention was focused on efforts to overcome the instability of oxytocin [13,14]. We have conducted several studies with the aim to develop a heat-stable oxytocin formulation.

We identified the main degradation products of oxytocin stressed at a temperature of 70°C in various buffers, pH values and storage time [13] as well as citryl oxytocin in citrate-buffered formulations [14]. The degradation reactions and target residues of oxytocin are indicated in Figure 1.

In a recent publication we described that formulations containing divalent metal salts in combination with citrate buffer strongly stabilize oxytocin in aqueous solutions [15]. However, the role by which divalent metal ions stabilize oxytocin in citrate buffer and their influence on (inhibition of) formation of degradation products were not elucidated. Therefore, the aim of the present study was to identify by LC−MS/MS the degradation products of oxytocin solutions after thermal stress and to investigate which degradation pathways are suppressed by the formulations containing combinations of divalent metal ions and citrate buffer.

Figure 1. Molecular structure of oxytocin with its ring and tail fragment ions formed upon MS/MS fragmentation.

53

4

2. materIals and methods2.1 MaterialsThe following materials were used in this study: oxytocin monoacetate powder (Diosynth, Oss, The Netherlands), citric acid, calcium chloride (Riedel-de Haen, Seelze, Germany), magnesium chloride, zinc chloride (Fluka, Steinheim, Germany), sodium hydroxide, acetonitrile, ammonium acetate (Merck, Darmstadt, Germany), DL-dithiothreitol and ammonium bicarbonate (Sigma-Aldrich Chemie Gmbh, Steinheim, Germany).

2.2 Formulation and Stability StudyThree independent batches of each of the four formulations given in Table 1 were prepared. The concentration of the citrate buffer was 10 mM, and the pH was adjusted to 4.5. All metal ion (M2+) solutions were prepared using their respective chloride salts at concentrations of 10 mM (MCl2). The solutions were stored for 5 days protected from light at either 4 or 70 °C. Prior to analysis, the samples were diluted 10-fold with water.

2.3 Reversed-Phase High Performance Liquid Chromatography HPLC was performed using a Shimadzu LC system, consisting of LC-20AD gradient pumps and a SIL-20AC autosampler. Chromatographic separation was achieved on an Alltima C18 column (2.1 × 150 mm 5 μm, Grace Davison Discovery Sciences). The injection volume was 50 μL. Elution was performed by a linear gradient from 5% to 60% B in 30 min, followed by an increase to 90% B in 1 min, where it was kept 4 min, after which it returned to the starting conditions. Eluent A was 95% water/5% acetonitrile (AcN), and eluent B was 5% water/95% AcN, both containing 0.05% v/v acetic acid and 10 mM ammonium acetate. The flow rate was 0.2 mL/min. The UV signal was recorded at 220 nm.

2.4 Liquid Chromatography−Mass Spectrometry The HPLC system was coupled to an API 3000 triple-quadrupole mass spectrometer (Applied Biosystems/MDS Sciex) via a turbo ion spray source. The ionization was performed by electrospray in the positive mode. Full scan spectra were recorded at a scan rate of 2 s from m/z 500 to 1300 and a step size of 0.2 amu with a declustering potential (DP) of 40 V and a focusing potential (FP) of 250 V. Product ion scans were acquired in specified time windows with a DP of 60 V, a FP of 300 V, a collision energy of 40 V and a collision cell exit

Table 1. The Composition of Oxytocin Liquid Formulation

formulationa oxytocin metal ion buffer

OCB 0.1 mM citrate 10 mMOCBCa 0.1 mM Ca2+ 10 mM citrate 10 mMOCBMg 0.1 mM Mg2+ 10 mM citrate 10 mMOCBZn 0.1 mM Zn2+ 10 mM citrate 10 mM

aOCB = oxytocin in the absence of divalent metal ions in 10 mM citrate-buffered solution at pH 4.5. OCBCa = oxytocin in the presence of Ca2+ in 10 mM citrate-buffered solution at pH 4.5. OCBMg = oxytocin in the presence of Mg2+ in 10 mM citrate-buffered solution at pH 4.5. OCBZn = oxytocin in the presence of Zn2+ in 10 mM citratebuffered solution at pH 4.5.

54

4


potential of 20 V. Data acquisition and processing was performed using Analyst version 1.4.2 and 1.5 software (Applied Biosystems/MDS Sciex).

LC−MS/MS were used to identify the oxytocin degradation products observed by RP-HPLC. The analyses were carried out on purified fractions corresponding to each of the major degradation peaks for oxytocin in citrate buffer formulation without divalent metal salts as well as by LC−MS for the formulation with and without divalent metal salts.

Reduction of disulfide bonds was performed by adding 10 mM DL-dithiothreitol (DTT) in 200 mM ammonium carbonate. The samples were analyzed after an incubation time of at least 10 min at room temperature.

3. results3.1 Degradation Products of Oxytocin in Citrate Buffer with and without Divalent Metal IonsAssignment of degradation products was done based on the m/z value of each compound from the LC−MS data, using known assignments and retention time order from previous studies [13,14]. MS/MS data were used for the more intense peaks to confirm the identification and to determine in which part of the molecule the modification had occurred.

The highest degradation product intensities were found in the formulation without divalent metal salts stressed at 70°C for 5 days. The HPLC profiles based on ultraviolet (UV) absorbance at 220 nm and the total ion current (TIC) of mass spectra with m/z between 900 and 1200 are shown in Figure 2a and Figure 2b, respectively. The UV profile shows 11 main peaks and the TIC 12 main peaks (see labels in Figure 2b). The LC−MS contour plot (Figure 2c) reveals that some peaks contain several molecular species, which cannot be distinguished in the TIC, notably peaks 2 and 3 reveal two compounds at 12.0 and 12.1 min, peaks 10 and 11 reveal two compounds at 16.7 and 16.8 min, and peaks 14 and 15 reveal two compounds at 18.7 and 18.9 min. All 15 assigned molecular species are presented in Table 2, and MS/MS product ion spectra are shown in Figures S1−S10 in the Supporting Information.

The protonated molecular ion of unmodified oxytocin was found at a retention time of 12.6 min. Figure 2d shows its extracted ion chromatogram (XIC) at m/z 1007.6. MS/MSfragmentation of unmodified oxytocin results in fragments of m/z 723.4 and 285.2, from the disulfide-linked ring of residues 1−6 and the C-terminal residues 7−9, respectively (Figure 1 and Figure S4 in the Supporting Information).

The degradation peaks eluting before the oxytocin peak (labeled 1 and 2 in Figure 2b) were identified as amide- and imide-linked N-citryl oxytocin with m/z of 1181.8 and 1163.8, respectively.14 The MS/MS spectrum of m/z 1181.6 showed that the ring fragment was shifted from m/z 723.4 to 897.4, but the tail fragment was still observed at m/z 285.2. It shows that the citrate modification is in the ring fragment and the most likely location is on the N-terminal amine (see Figure 1 and Figure S2 in the Supporting Information) [14].

At a retention time of 12.0 and 12.1 min (Figure 3b and 3c), there is overlap of two different compounds (peaks 2 and 3), namely a monodeamidated species of oxytocin with a protonated molecular ion at m/z 1008.6 in a very low intensity and dehydrated (imide-linked) N-citryl oxytocin with the protonated molecular ion at m/z 1163.6 [14]. The MS/MS spectrum of m/z 1008.6 showed a shift of the ring fragment from m/z 723.4 to 724.4, but no shift of the tail fragment mass (Figure S3 in the Supporting Information), which means that the deamidation occurred at Asn5 or Gln4 [13]. The 13C denoted as peak in Figure 3c is

55

4

Figure 2. LC−UV trace (a), total ion chromatogram (m/z 900−1200) (b), and LC−MS contour plot (c) of stressed oxytocin in citrate buffered solution with the extracted ion current of oxytocin (d) and acetylated oxytocin (e).

56

4


the second isotope peak of oxytocin (m/z 1007.6), which is also found at m/z 1008.6. The three other peaks in Figure 3b, at retention times of 16.8, 17.8, and 18.9 min, are the dimer degradation products described in Figure 4c.

A compound at m/z 1049.6 was found at a retention time of 14.4 min (Figure 2e) in every sample including the original oxytocin preparation. The difference of +42 Da with respect to oxytocin suggests acetylation, presumably at the N-terminal amine (see Figure S6 in the Supporting Information). This acetylation reaction may have occurred during synthesis since oxytocin is supplied as the monoacetate salt.

At a retention time of 13.5 min a trisulfide degradation product was identified from the m/z 1039.6 peak. Upon MS/ MS fragmentation, the ring fragment shifted to m/z 755.4, while the tail fragment mass did not change (see Figure S5 in the Supporting Information). The mass difference of +32 Da can be assigned to a trisulfide modification in the oxytocin ring [13]. At a retention time of 13.9 min, a product of m/z 1195.6 was found, which is tentatively identified as the N-citryl oxytocin with the trisulfide modification. The formation of a tetrasulfide product (m/z 1071.6) was found at the retention time of 14.8 min. Molecular structure N-citryl oxytocin, trisulfide and tetrasulfide modification produced from the degradation of stressed oxytocin in citrate buffered solution are shown in Figure 5.

Seven peaks of dimers were evident as doubly charged ions in MS eluting between 15.5 and 18.9 min (Table 2 and Figure 4). Dimer 1 and 2 had identical masses (m/z 975.6) and were also observed in a previous study of stressed oxytocin solutions [13] where they were assigned to the doubly protonated form of a sulfur-linked dimeric oxytocin species that has been doubly deamidated and which has lost two sulfur atoms through β-elimination [13].

Identification of dimers is difficult due to the overlapping LC peaks and multiple ion forms (protonated and ammoniated, and doubly charged, Figure 2c). In addition, their MS/MS

Table 2. Summary of the Observed Degradation Products of Oxytocin after Stressing at 70°C for 5 Days in 10 mM Citrate Buffer

no. tR (min) m/z charge Assignment

1 11.8 1181.6 1+ oxytocin N-citryl amide2 12.0 1008.6 1+ monodeamidation at Asn5 or Gln4

3 12.1 1163.6 1+ oxytocin N-citryl imide4 12.6 1007.6 1+ oxytocina

5 13.5 1039.6 1+ oxytocin trisulfide6 13.9 1195.6 1+ oxytocin trisulfide N-citryl imide7 14.4 1049.6 1+ N-acetyl oxytocina

8 14.8 1071.6 1+ oxytocin tetrasulfide9 15.5 975.6 2+ dimer 110 16.7 975.6 2+ dimer 211 16.8 1008.6 2+ dimer 312 17.4 1024.6b 2+ dimer 413 17.8 1008.6 2+ dimer 514 18.7 1024.6b 2+ dimer 615 18.9 1008.6 2+ dimer 7

aNo degradation products; compounds were also found in the original oxytocin preparation. bAmmoniated.

57

4

Figure 3. LC−MS extracted ion currents for oxytocin N-citryl amide (a) and N-citryl imide (b), monodeamidation (c), trisulfide (d), trisulfide N-citryl imide (e), and tetrasulfide products from the degradation of stressed oxytocin in citrate buffered solution.

58

4


fragmentation spectra only show the tail fragment of m/z 285.2 (Figures S7−10 in the Supporting Information) and consequently they are not informative.

In order to distinguish disulfide linked dimers from other types of dimers, we have done the LC−MS analysis on the most degraded solution after disulfide bond reduction with DTT. All monomeric products observed were reduced, as summarized in Table 3, producing oxytocin, oxytocin acetate, and N-citryl amide/imide in the reduced thiol form (Figures S11−S17 in the Supporting Information). Tri- and tetrasulfide forms were also absent after reduction and are expected to have been converted to reduced oxytocin. The product ion spectrum after DTT reduction of m/z 993.6 at a retention time of 16.6 min (Figure S18 in the Supporting Information) showed that it is a monomer which has undergone β-elimination at Cys1 and deamidation at Asn5 or Gln4. This is as evidenced by the presence of an unmodified y4 ion and a b6 ion (Figure S1 in the Supporting Information) mass of −33 Da with respect to oxytocin. β-Elimination leads to a loss of 34 Da (-H2S), and

Figure 4. LC−MS extracted ion currents for dimer degradation products.

59

4

Figure 5. Molecular structures of oxytocin N-citryl amide (a), N-citryl imide (b), trisulfide (c) and tetrasulfide (d) produced from the degradation of stressed oxytocin in citrate buffered solution.

deamidation increases the mass by 1 Da. The presence of this reduced monomer at high intensity confirms the assignment of dimers 1 and 2, which proves that these dimers are disulfide linked at the unmodified Cys6 residues.

In the LC−MS profile (Figure 6) it was also found that the LC−MS peaks of all dimers had disappeared after reduction. However there were 2 new peaks present at retention times of 17.9 and 18.1 min with m/z 994.0 and 993.6 respectively, which are doubly charged ion (Figures S19−S20 in the Supporting Information). The presence of dimers after DTT reduction, combined with the observation that all original 7 dimer peaks were absent after reduction, shows that all dimers had at least one disulfide bond and some had an additional thio-ether bond (which may form after β-elimination and reaction with another Cys residue). Tyrosine-linked dimers were unlikely to be present, since no monomeric tyrosine oxidation products were observed.

3.2 Effects of Divalent Metal Ions on Oxytocin’s Degradation ProfileAddition of divalent metal salts in the formulation did not result in any additional degradation products compared to samples without metal salts. As expected from our previous study [14], it resulted in a significant reduction of most degradation peaks. Figure 7 shows that, of all formulations tested, oxytocin formulated in 10 mM citrate buffer (pH 4.5) without divalent metal ions (OCB) was most degraded with only approximately 35% oxytocin recovered after incubation at 70°C for 5 days. However, when divalent metal ions were added, the degradation was decelerated and a recovery of about 70% oxytocin was found, irrespective of the nature of the divalent metal ion, which is in line with our previous study [15].

60

4


From each stressed formulation of oxytocin in citrate buffer in the presence of calcium, magnesium, or zinc ions, we recorded the intensity of each degradation product from their respective extracted MS ion currents. The relative intensities in percentage of the peak area of each degradation product in the presence of divalent metal ions with respect to that of the same product in the absence of divalent metal ions are listed in Table 4.

Addition of divalent metal ions reduced the peak intensity of almost all degradation products. The formation of the N-citryl oxytocin and the tri- and tetrasulfide was reduced by 20−70%. The formation of the dehydrated N-citryl oxytocin and dimmers 1, 2, 4, 5 was even more strongly reduced, by 90%. No clear reduction of the intensity of the deamidated species (m/z 1008.6, retention time 12.1 min) and dimer 3 (m/z 1024.8, retention time 17.4 min) was observed, possibly because of their low signal intensity which made peak integration difficult (Figures 3b and 4b).

The acetylated oxytocin species (peak number 7) appeared to be very stable: no significant difference in its relative intensity was observed either in the presence or in the absence of divalent metal ions. In addition, there is no significant difference of the intensity of this product in unstressed or stressed condition. We conclude that this molecule is not a degradation product formed under stress conditions but, as mentioned above, an impurity of oxytocin formed during synthesis.

4. dIscussIonSeveral biological and physicochemical methods have been described to monitor the stability of oxytocin. HPLC with UV/vis detection is the most frequently used physicochemical method to monitor oxytocin stability. Especially when combined with mass spectrometry (MS) detection, most degradation products can be identified and quantified [16,17]. Hyphenation of the LC unit to MS via an electrospray ionization (ESI) interface allows sensitive and selective confirmation of degradation products by extracting corresponding ion chromatograms from the recorded total ion current (TIC)[14] and by using MS/MS for the identification of degradation products. Furthermore, previous studies have demonstrated that LC−MS can be used to monitor the stability of oxytocin in pharmaceutical dosage forms [13,18].

Table 3. Summary of the Observed Degradation Products of Oxytocin after Stressing at 70 °C for 5 Days in 10 mM Citrate Buffer Followed by Reduction by DTTa

no. tR (min) m/z charge Assignment

a 12.8 1007.4 1+ oxytocin (disulfide)b 13.1 1200.6b 1+ oxytocin N-citryl amidec 13.5 1009.6 1+ oxytocind 14.6 1182.6b 1+ oxytocin N-citryl imidee 15.5 1068.8b 1+ N-acetyl oxytocinf 16.6 993.6b 1+ β-elimination at Cys1, monodeamidation at Asn5 or Gln4

g 17.9 994.0 2+ dimerh 18.1 993.6 2+ dimer

aAssigned products are in reduced form. bAmmoniated.

61

4

The degradation of oxytocin in various formulations was analyzed after 5 days of incubation at 70°C. A temperature of 70°C was chosen to ensure the formation of qualitatively similar degradation products as found by Poole et al., who incubated oxytocin in citrate buffered solutions at 70°C for 2 days [14]. The aim of the present study was to investigate the effect of stabilizing divalent metal ions on the degradation profile.

In order to observe sufficient amounts for characterization of all degradation products in the stabilized formulations, it was necessary to prolong the thermal stress from 2 to 5 days.

The strongly decreased intensity of the dimeric products suggests that formulation with divalent metal ions protects against thiol exchange, hence avoiding dimerization. At low pH, dimerization occurs via thiol exchange in the disulfide bridge, which progresses via

Figure 7. Recovery of oxytocin in the absence (OCB) and presence of 10 mM Ca2+ (OCBCa), Mg2+ (OCBMg) and Zn2+ (OCBZn) in 10 mM citrate-buffered solution at pH 4.5 under stressed condition at a temperature of 70 °C for 5 days. Oxytocin recovery determined by LC−MS. The results are depicted as averages of three independent measurements±SD.

Figure 6. Total ion chromatogram (m/z 900−1200) of stressed oxytocin in citrate buffered solution followed by reduction using DTT.

62

4


Table 4. The Effect of Divalent Metal Ions on the Intensity of Degradation Products of Oxytocin Found after Stressing at 70°C for 5 Days in 10 mM Citrate Buffera

Assignment

relative peak intensity (%)

OCBCa OCBMg OCBZn

oxytocin N-citryl amide 71 ± 19 67 ± 20 52 ± 12monodeamidation at Asn5 or Gln4 534 ± 86b 436 ± 65b 906 ± 50b

oxytocin N-citryl imide 58 ± 19 73 ±9 50± 8oxytocin trisulfide 40 ± 39 37 ±5 30± 28oxytocin trisulfide N-citryl imide 3 ±4 4±5 1± 1N-acetyl oxytocinc 107 ±7 89± 13 114 ± 13oxytocin tetrasulfide 53 ± 64 31 ±5 33± 31dimer 1 5 ±1 5±4 2± 1dimer 2 7 ±1 10± 7 3± 2dimer 3 15 ± 11 14 ±7 6± 4dimer 4 172 ± 42b 150 ± 54b 90 ± 7b

dimer 5 15 ± 11 15 ±5 6± 5dimer 6 27 ± 21 17 ±9 17± 16dimer 7 43 ± 42b 42 ± 11b 31 ± 30b

aThe level of degradation products is expressed relative to the levels found in the solution without metal ions. bWeak MS signal. cNo degradation product; compound was also found in the original oxytocin preparation.

a sulfonium cation, which is formed following protonation of the disulfide bridge [19,20]. This could also explain the ability of these formulations to inhibit the formation of tri/tetrasulfide oxytocin species.

The combination of citrate buffer and divalent metal ions greatly reduces the formation of most dimers after thermal stress. The absence of a significant decrease in the intensity of dimer 4 hardly contributes to the total amount of dimers because its intensity in the formulation without divalent metal ions after thermal stress was already low.

In this study, the buffer used was citrate, which is considered to be safe and occurs in many foods and is also a normal metabolite in the body. Its calcium, potassium and sodium salts do not constitute a significant hazard to humans [21]. After thermal stress, covalent citrate-oxytocin adducts were formed through a mechanism involving the intermediate production of citrate anhydride, which reacts with the N-terminal amino group from the cysteine residue [14]. Inhibition of the formation of N-citryl oxytocin by divalent metal ions as found in the present study might be due to a differential interaction of oxytocin and divalent metal ions for citrate. Wyttenbach et al [22] found that under acidic conditions (pH 3.0) divalent metal ions bind to the carbonyl groups in the ring structure of oxytocin. The presence of doubly charged cations, such as Zn2+, Mg2+, Ni2+, Mn2+ and Co2+, has been found to be essential in increasing the potency of the specific binding of oxytocin to its receptor. Therefore the complex formed might increase the biological activity [23]. Further, isothermal titration calorimetry data showed that citrate interacts with divalent metal ions, and this interaction is stronger than that of citrate with oxytocin (Tables SI and SII and Figures S21−S22 in the Supporting Information). Free citrate ions are probably more reactive toward the N-terminal amino group from the cysteine residue than divalent metal−citrate

63

4

adducts. In the presence of divalent metal ions, the formation of metal−citrate adducts reduces the concentration of free citrate, which reduces the driving force for citrate adduct formation with oxytocin [15]. A second possibility is that binding of a divalent metal ion to oxytocin may change the position of the N-terminal amino group from the cysteine residue rendering it less accessible for binding with citrate.

It can be concluded that citrate has two opposite effects on oxytocin stability. First, as also shown by Poole et al.,[14] it is reactive itself and can attack the N-terminal amino group from the cysteine residue to form an adduct. Second, as shown in our previous study, [15] it protects oxytocin from degradation in the presence of divalent metal ions. In this study, we clearly show that the stabilization is due to the suppression of N-citryl oxytocin, tri/tetrasulfide and dimer formation. Furthermore, all reactions that were suppressed occurred on Cys1, and possibly Cys6. Cysteine is susceptible to oxidation and β-elimination, and degradation of oxytocin involving cysteine leads to dimerization, formation of tri/tetrasulfide and β-elimination followed by thio-ether formation.

Divalent metal ions in combination with citrate buffer suppress intermolecular reactions in the ring structure of oxytocin presumably by forming a complex in the region where the degradation reaction occurs. No significant difference was observed among the three tested divalent metal ions, Ca2+, Mg2+, and Zn2+, suggesting that divalency is the most important property of the metal contributing to stabilization of the oxytocin-metal-citrate cluster.

assocIated content *Supporting InformationFigures depicting the MS/MS series of reduced oxytocin, numerous product ion spectra, and calorimetric titration profiles and tables of thermodynamics data.

acknowledgmentsThe authors want to thank MSD Oss for providing oxytocin for the study. This study was performed within the framework of the Dutch Top Institute Pharma project: number D6−202.

references1. H.H. Dale, The Action of Extracts of the

Pituitary Body, Biochem. J. 4 (1909) 427-47. 2. C.M. Karbiwnyk, K.C. Faul, S.B. Turnipseed,

W.C. Andersen, K.E. Miller, Determination of oxytocin in a dilute IV solution by LC-MS(n), J harm Biomed Anal 48 (2008) 672-7.

3. J. Owen, J.C. Hauth, Oxytocin for the induction or augmentation of labor, Clin Obstet Gynecol 35 (1992) 464-75.

4. F.A. Chaibva, R.B. Walker, Development and validation of a stability-indicating analytical method for the quantitation of oxytocin in pharmaceutical dosage forms, J Pharm Biomed Anal 43 (2007) 179-85.

5. J.W. Gard, J.M. Alexander, R.E. Bawdon, J.T. Albrecht, Oxytocin preparation stability

in several common obstetric intravenous solutions, Am. J. Obstet. Gynecol. 186 (2002) 496-8.

6. R.A. Turner, J.G. Pierce, V. du Vigneaud, The purification and the amino acid content of vasopressin preparations, J. Biol. Chem. 191 (1951) 21-8.

7. H. Davoll, R.A. Turner, J.G. Pierce, V. du Vigneaud, An investigation of the free amino groups on oxytocin and desulfurized oxytocin preparations, J. Biol. Chem. 193 (1951) 363-70.

8. M. Winkler, W. Rath, A risk-benefit assessment of oxytocics in obstetric practice, Drug. Saf. 20 (1999) 323-45.

64

4


9. V. du Vigneaud, C. Ressler, S. Trippett, The sequence of amino acids in oxytocin, with a proposal for the structure of oxytocin, J. Biol. Chem. 205 (1953) 949-57.


11. G. Gimpl, F. Fahrenholz, The oxytocin receptor system: structure, function, and regulation, Physiol. Rev. 81 (2001) 629-83.



14. R.A. Poole, P.T. Kasper, W. Jiskoot, Formation of amide- and imide-linked degradation products between the peptide drug oxytocin and citrate in citrate-buffered formulations, J. Pharm. Sci. 100 (2011) 3018-22.

15. C. Avanti, J.P. Amorij, D. Setyaningsih, A. Hawe, W. Jiskoot, J. Visser, A. Kedrov, A.J. Driessen, W.L. Hinrichs, H.W. Frijlink, A new strategy to stabilize oxytocin in aqueous

solutions: I. The effects of divalent metal ions and citrate buffer, AAPS J. 13 (2011) 284-90.

16. The United States Pharmacopeial Convention,Rockville, USP-29 NF-24, MD, 2005. (2006).

17. G.S. Shaw, Synthetic calcium-binding peptides, Methods Mol Biol 173 (2002) 175-82.

18. C.W. Huck, V. Pezzei, T. Schmitz, G.K. Bonn, A. Bernkop-Schnurch, Oral peptide delivery: are there remarkable effects on drugs through sulfhydryl conjugation?, J Drug Target 14 (2006) 117-25.


20. G. Bulaj, Formation of disulfide bonds in proteins and peptides, Biotechnology Advances 23 (2005) 87-92.

21. Joint FAO/WHO Expert Committee on Food Additives, Citric acid and its calcium, potassium and sodium salts 539 (1974).

22. T. Wyttenbach, D. Liu, M.T. Bowers, Interactions of the hormone oxytocin with divalent metal ions, J. Am. Chem. Soc. 130 (2008) 5993-6000.

23. A.F. Pearlmutter, M.S. Soloff, Characterization of the metal ion requirement for oxytocin-receptor interaction in rat mammary gland membranes, J Biol Chem 254 (1979) 3899-906.

65

4

Figure S1. Oxytocin MS/MS series

gggFigure S2. Product ion spectrum of m/z 1198.8 at a retention time of 11.8 min of oxytocin N-citryl amide

gggggg

supportIng InformatIon

66

4


ggggggFigure S3. Product ion spectrum of m/z 1007.6 at a retention time of 12.0 min of monodeamidated species of oxytocin

Figure S4. Product ion spectrum of m/z 1007.6 at a retention time of 12.6 min of unmodified oxytocin

67

4

ggggggFigure S5. Product ion spectrum of m/z 1039.6 at a retention time of 13.5 min of oxytocin trisulfide

Figure S6. Product ion spectrum of m/z 1066.4 at a retention time of 14.4 min of oxytocin acetate (ammoniated)

68

4


ggggggFigure S7. Product ion spectrum of m/z 993.2 at a retention time of 15.5 min of oxytocin dimer 1(ammoniated)

Figure S8. Product ion spectrum of m/z 1000.4 at a retention time of 16.8 min of oxytocin dimer 3

69

4

ggggggFigure S9. Product ion spectrum of m/z 1000.4 at a retention time of 17.8 min of oxytocin dimer 5

Figure S10. Product ion spectrum of m/z 1000.4 at a retention time of 18.7 min of oxytocin dimer 6

70

4


gggTIC of +Q1: from Sample 14 (july 3 DTT) of Q1.wiff (Turbo Spray) Max. 1.9e7 cps.

6 8 10 12 14 16 18 20 22

Time, min

0.0

2.0e6

4.0e6

6.0e6

8.0e6

1.0e7

1.2e7

1.4e7

1.6e7

1.8e7

2.0e7 no DTTwith DTT

Figure S11. Overlay of UV-RP-HPLC traces of oxytocin in citrate buffer after storage at 70°C for 5 days detected at a wavelength of 220 nm before DTT reduction (light gray) and after DTT reduction (dark gray).

Figure S12. Overlay of LC-MS TIC traces of oxytocin in citrate buffer after storage at 70°C for 5 days before and after DTT reduction

71

4

ggggggFigure S13. Product ion spectrum after DTT reduction of m/z 1007.4 at a retention time of 12.9 min of oxytocin disulfide

Figure S14. Product ion spectrum after DTT reduction of m/z 1200.6 at a retention time of 13.1 min of N-citryl oxytocin reduced

72

4


ggggggFigure S15. Product ion spectrum after DTT reduction of m/z 1009.4 at a retention time of 13.5 min of oxytocin reduced

Figure S16. Product ion spectrum after DTT reduction of m/z 1182.6 at a retention time of 14.7 min of N-citryl oxytocin dehydrated (ammoniated)

73

4

ggggggFigure S17. Product ion spectrum after DTT reduction of m/z 1068.8 at a retention time of 15.6 min of Oxytocin acetate ammoniated

Figure S18. Product ion spectrum after DTT reduction of m/z 993.6 at a retention time of 16.6 min of β-elimination and deamidation at Asn5 or Gln4

74

4


Figure S19. Product ion spectrum after DTT reduction of m/z 994.0 at a retention time of 18.0 min of dimer

Figure S20. Product ion spectrum after DTT reduction of m/z 993.6 at a retention time of 18.1 min of dimer

gggggg75

4

(a) Ca-citrate-water (b) Mg- citrate-water (c) Zn-citrate-water

(a) Ca-OT-water (b) Mg-OT-water (c) Zn-OT-water

Figure S21. Calorimetric titration profiles of aliquots of 125 mM divalent metal ions: (a) Ca2+, (b) Mg2+, and (c) Zn2+ into 5 mM oxytocin in water pH 4.5. The heat absorbed per mol of titrant is plotted versus the ratio of the total concentration of divalent metal ions to the total concentration of oxytocin

Figure S22. Calorimetric titration profiles of aliquots of 125 mM divalent metal ions: (a) Ca2+, (b) Mg2+, and (c) Zn2+ into 10 mM citrate buffer pH 4.5. The heat absorbed per mol of titrant is plotted versus the ratio of the total concentration of divalent metal ions to the total concentration of oxytocin

76

4


Table SI. Thermodynamics of divalent metal binding to oxytocin as determined by isothermal titration calorimetry in the absence of buffer.

Metal Phase N (sites) Ka (M-1) ΔH° (cal mol-1) ΔS° (cal/mol/deg)

Ca2+ 12

10.27

30300

-4000-3516

-81

Mg2+ No binding observedZn2+ No binding observed

The results are depicted as averages of three independent measurements with relative standard deviations below 10%

Table SII. Thermodynamics of divalent metal binding to citrate as determined by isothermal titration calorimetry in the absence of oxytocin.

Metal Phase N (sites) Ka (M-1) ΔH° (cal mol-1) ΔS° (cal/mol/deg)

Ca2+ 12

0.31.1

80038

5700-1800

151.5

Mg2+ 1 0.2 300 5800 29Zn2+ 1 0.42 540 1600 18

The results are depicted as averages of three independent measurements with relative standard deviations below 10%

77

Christina Avanti1, Wouter L.J. Hinrichs1, Angela Casini2, Anko C. Eissens1, Annie van Dam3, Alexej Kedrov4 Arnold J. M. Driessen4, Henderik W. Frijlink1,

and Hjalmar P. Permentier3

1Department of Pharmaceutical Technology & Biopharmacy, 2Pharmacokinetics, Toxicology and Targeting, Research Institute of Pharmacy,

3Mass Spectrometry Core Facility and 4Department of Molecular Microbiology,

University of Groningen, The Netherlands

InsIght Into the stabIlIty of the zInc-aspartate-oxytocIn formulatIon

5

5

abstractThe stability of the peptide hormone oxytocin in pharmaceutical formulations during prolonged storage at high temperatures is strongly dependent on the solution composition. The aim of this study was to investigate the effect of divalent metal ions (Ca2+, Mg2+ and Zn2+) on the stability of oxytocin in aspartate buffer (pH 4.5) and determine their interaction with the peptide in aqueous solution. Reversed phase and size exclusion-high performance liquid chromatography (RP-HPLC and HP-SEC) measurements indicated that after 4 weeks of storage at 55°C all tested divalent metal ions improved the stability of oxytocin in aspartate buffered solutions (pH 4.5). However, the stabilizing effects of Zn2+ were by far superior compared to Ca2+ and Mg2+. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) showed that the combination of aspartate and Zn2+ in particular suppressed the formation of peptide dimers. As shown by isothermal titration calorimetry, Zn2+ interacted with oxytocin in the presence of aspartate buffer while Ca2+ or Mg2+ did not. In conclusion, the stability of oxytocin in the aspartate buffered-solution is strongly improved in the presence of zinc ions, and the stabilization effect is correlated with the ability of the divalent metal ions in aspartate buffer to interact with oxytocin. The reported results are discussed in relation to the possible mode of interactions between the peptide, zinc and buffer components leading to the observed stabilization effects.

5

StabIlIty of the dIvalent Metal -aSpartate-oxytocIn coMplex

1. IntroductIonPregnant women may face life-threatening blood loss at the time of delivery. As stated in the ICM-FIGO joint statement, the drug of choice to prevent bleeding after child-birth (post-partum hemorrhage) is oxytocin [1]. Oxytocin (Figure 1) is a nonapeptide hormone that is composed of a cyclic sequence of Cys1-Tyr2-Ile3-Gln4- Asn5-Cys6 with an N-terminal amino group, and of a linear Pro7-Leu8-Gly9 with a C-terminal amide [2].

Unfortunately, a major problem in practice is that injectable oxytocin formulations are highly unstable as the storage temperature rises to 30°C or higher [3]. Therefore, oxytocin should be stored and transported refrigerated, the so-called cold chain, which is not always guaranteed especially in rural and tropical areas [4]. At a molecular level oxytocin can undergo degradation via deamidation, oxidation or thiol exchange. In particular, the stability of the Cys1-Cys6 disulfide bridge with respect to thiol exchange, or to oxidation due to the presence of oxygen, light and/or metal ions, is crucial to avoid oxytocin degradation and progressive aggregation [5].

Several studies have been conducted to improve the stability of oxytocin formulations [6-8]. Within this frame, we have previously demonstrated that the use of combinations of divalent metal ions with citrate buffer greatly improve the stability of oxytocin in aqueous solution, while divalent metal ions added to non-buffered aqueous oxytocin formulation have only limited stabilizing effects at 40 or 55°C [7]. In a consecutive study, we have shown that formation of a complex of divalent metal ions and citrate with oxytocin leads to the suppression of cysteine-mediated inter- or intramolecular reactions, thus suppressing tri/tetrasulfide and dimer formation [8].

We have also observed that different type of buffers may lead to a different interaction between divalent metal ions and oxytocin, therefore having a different impact on oxytocin stability. For example, we demonstrated that addition of divalent metal ions to oxytocin solutions in either acetic or citric acid buffers results in different stabilization effects of the peptide in aqueous solution [7]. While Ca2+, Mg2+ and Zn2+ ions in combination with citrate buffer were successful in

Figure 1. Oxytocin structure.

81

5

stabilizing oxytocin, no stabilizing effect was observed for the combination of the same divalent metal ions and acetate buffer [7]. Acetic acid and citric acid are carboxylic acids with one and three carboxylate groups, respectively. Thus, the above mentioned results suggested that one carboxylate group is not sufficient for stabilizing the oxytocin-metal-buffer salt cluster.

The aim of the present study is to investigate the effects of divalent metal ions (Ca2+, Mg2+ and Zn2+) on the stability of oxytocin in aspartate buffer. Aspartate is a commonly used buffer in parenteral products approved by the FDA for formulation purposes [9]. It has one amine and two carboxylate groups, which may establish different interactions and affect differently the peptide stability with respect to citric acid.

It should be noted that both the pH and concentration of metal ions in the formulation were found to be crucial in the effectiveness of oxytocin stabilization in citrate buffer [7]. Thus, since the optimum stability of oxytocin was reported to be at pH 4.5 [6], the peptide formulations were maintained at that pH, while various divalent metal ion concentrations were tested. Afterwards, we have also investigated the effect of divalent metal ions on the degradation profile of oxytocin in aspartate buffer by liquid chromatography-tandem mass spectrometry (LC-MS/MS), as well as the thermodynamics of the oxytocin-metal-buffer system by isothermal titration calorimetry (ITC).

2. materIals and methods 2.1. MaterialsOxytocin monoacetate powder (Diosynth. Oss, The Netherlands) was kindly provided by MSD, Oss, The Netherlands. L-aspartic acid, magnesium chloride, and zinc chloride were purchased from Fluka, Steinheim, Germany. Calcium chloride was from Riedel-de Haen, Seelze, Germany, and sodium hydroxide, sodium dihydrogen phosphate dihydrate, acetonitrile (supergradient grade), as well as formic acid were purchased from Merck, Darmstadt, Germany.

2.2. Methods2.2.1. Formulation of oxytocin solution and stability studyOxytocin solution was formulated at a concentration of 0.1 mg/mL (0.094 mM) in 10 mM aspartate buffer at pH 4.5 (pH adjusted with sodium hydroxide) with different concentrations of divalent metal ions Ca2+, Mg2+ and Zn2+. All divalent metal ion solutions were prepared using their chloride salts at concentrations of 2, 5, 10 and 50 mM. The concentration of oxytocin was determined using a UV spectrometer as described previously [7,10]. After preparation, the solutions were stored in 6R glass type 1 vials for 4 weeks at either 4 or 55°C, and protected from light. Some selected formulations were also stored for 5 days at 70°C and the samples were diluted 10-fold with water for LC-MS/MS analysis. During the stability study controlled pH levels in the samples were within 0.1 pH units.

It must be noted that oxytocin is commonly formulated at very low concentration, which is about 1/10 of the concentration within this study (RP-HPLC and HP-SEC). The higher concentration was chosen to have better intensity of signal for the degradation products produced by the heat stress.

2.2.2. Reversed-Phase High-Performance Liquid Chromatography (RP-HPLC)RP-HPLC was carried out according to the procedure described earlier [6,7]. An Alltima C-18 RP column with 5 μm particle size, inner diameter of 4.6 mm, and length of 150 mm

82

5


(Alltech, Ridderkerk, Netherlands), a Waters (Millipore) 680 Automated Gradient Controller, two Waters 510 HPLC pumps, a Waters 717 Plus Autosampler, and a Waters 486 Tunable Absorbance UV Detector were used. Samples of 20 μL were injected and the separation was carried out at a flow rate of 1.0 mL/min and UV detection at 220 nm. Samples were eluted using 15% (v/v) acetonitrile in 65 mM phosphate buffer, pH 5.0 as solvent A and 60% (v/v) acetonitrile in 65 mM phosphate buffer, pH 5.0 as solvent B. The acetonitrile concentration was linearly increased from 15% at the beginning, to 20% at 10 min, to 30% at 20 min and finally to 60% at 25 min. The recovery of oxytocin is expressed as the percentage of initial amount.

2.2.3. Size Exclusion HPLC (HP-SEC)HP-SEC was performed according to the method previously reported [6,7]. A Superdex Peptide 10/300 GL column (GE Healthcare Inc., Brussels, Belgium) was used on an isocratic HPLC system with a Waters 510 pump, a Waters 717 plus auto sampler, a Waters 474 Scanning Fluorescence Detector and Waters 484 Tunable Absorbance Detector (Waters, Milford Massachusetts, USA). Samples of 50 μL were injected, and separation was performed at a flow rate of 1 mL/min. Chromatograms were obtained using fluorescence detection at excitation wavelength of 274 nm and emission wavelength of 310 nm. The mobile phase consisted of 30% acetonitrile and 70% 0.04 M formic acid. The recovery of monomeric oxytocin is expressed as the percentage of initial amount.

2.2.4. Liquid Chromatography-Mass Spectrometry/Mass Spectrometry (LC-MS/ MS)The LC-MS/MS system was set up according to the method described earlier [6,8]. A Shimadzu LC system equipped with LC-20AD gradient pumps and a SIL-20AC autosampler was used. The chromatographic separation was carried out on an Alltima C18 column (internal diameter 2.1 mm, length 150 mm, particle size 5 µm, Grace Davison Discovery Sciences). The gradient mobile phase composition was a mixture of solvent A consisting of 95% water / 5% acetonitrile and solvent B consisting of 5% water / 95% acetonitrile, both containing 0.05% (v/v) acetic acid and 10 mM ammonium acetate. Elution was performed by a linear gradient from 5 to 60% of the solvent B in 30 min, followed by an increase to 90% solvent B in 1 min, where it was kept 4 min, after which it returned to the starting conditions. The flow rate was 0.2 mL/min. The UV signal was recorded at 220 nm. The injection volume was 50 μL.

The HPLC system was coupled to an API 3000 triple-quadrupole mass spectrometer (Applied Biosystems/MDS Sciex) via a Turbo Ion Spray source. The ionization was performed by electrospray in the positive mode. Full scan spectra were recorded at a scan rate of 2 s from m/z 500 to 1300 and a step size of 0.2 amu with a Declustering Potential (DP) of 40 V and a Focusing Potential (FP) of 250 V. Product ion scans were acquired in specified time windows with a DP of 60 V, a FP of 300 V, a Collision Energy of 40 V and a Collision Cell Exit Potential of 20 V. Data acquisition and processing was performed using Analyst version 1.5 software (Applied Biosystems/MDS Sciex).

LC-MS/MS was used to identify the oxytocin degradation products observed by RP-HPLC with UV detection. The analyses were carried out on each of the major degradation peaks for oxytocin in aspartate buffer formulation without divalent metal salts as well as by LC-MS for the formulation with divalent metal salts.

83

5

2.2.5. Isothermal Titration Calorimetry (ITC)Microcalorimetric titrations of divalent metal ions to oxytocin in aspartate buffer were performed by using a MicroCal ITC200 microcalorimeter (Northampton, MA 01060 USA) as described previously [7]. A solution of 30 µL of 125 mM divalent metal chloride (calcium, magnesium, or zinc chloride) in 10 mM aspartate buffer, pH 4.5 was placed in the syringe, while 300 µL of 5 mM oxytocin in 10 mM of aspartate buffer, pH 4.5 was placed in the sample cell. The reference cell contained 300 µL of aspartate buffer. Experiments were performed at 55°C. The effective heat of the peptide-metal ion interaction upon each titration step was corrected for dilution and mixing effects, as measured by titrating the divalent metal ion solution into the buffer and by titrating the buffer into the oxytocin solution (reference measurement). To investigate the possibility of oxytocin or metal ion binding to the buffer components, control experiments were performed in water. The heats of bimolecular interactions were obtained by integrating the peak following each injection. All measurements were performed in triplicate. ITC data were analyzed by using the ITC non-linear curve fitting functions for one or two binding sites from Origin 7.0 software (MicroCal Software, Inc.).

3. results 3.1. The effect of divalent metal ions on oxytocin stability in aspartate buffer solutionTo investigate the stability of oxytocin in the aspartate buffered solutions in the presences of divalent metal ions, oxytocin formulation in 10 mM aspartate buffer (pH 4.5) with various concentrations of divalent metal ions were prepared. Oxytocin is commonly formulated at a concentration of about 0.01 mg/mL, which is 10 times lower than the concentration used in this study. The higher concentration was chosen to have better intensity of degradation product produced by the heat stress in RP-HPLC and HP-SEC analysis. Oxytocin recovery and the presence of oxytocin monomer were determined for all the samples after 4 weeks at different storage temperatures by RP-HPLC and HP-SEC, respectively, as described in the experimental section. The obtained results are shown in Figure 2.

After 4 weeks of storage at 4°C, oxytocin remained stable in all formulations. Instead, after 4 weeks of storage at 55°C oxytocin stability increased with increasing divalent metal ions concentration. Ca2+ and Mg2+ had similar effects on improving the oxytocin stability: the recoveries of oxytocin, as well as the remaining percentage of oxytocin monomer were increased up to 45% in the presence of 50 mM Mg2+, and up to 35% in the presence of 50 mM Ca2+. Notably, Zn2+ was much more effective in stabilizing oxytocin: at a concentration of 2 mM Zn2+ increased oxytocin recovery up to 35% (i.e. to a comparable extent as 50 mM Ca2+ or Mg2+), and almost ca. 75% of oxytocin monomer was recovered. Overall, both RP-HPLC and HP-SEC showed that Zn2+ has superior stabilizing effects in aspartate buffer compared to Ca2+ and Mg2+

3.2. The effect of divalent metal ions on the degradation profile of oxytocin in aspartate bufferTo assign the degradation products of oxytocin in the formulations, oxytocin in10 mM aspartate buffer (pH 4.5) in the absence and presence of 10 mM divalent metal ions was analyzed by LC-MS/MS after incubation of the samples at 70°C for 5 days to ensure the

84

5


formation of sufficient degradation products intensity. MS/MS analysis was applied to confirm the identification of the most intense peaks, and degradation products were assigned following our previous reports [6,8,11]. Figure 3A shows the total ion-current (TIC) of mass spectra in the m/z range between 900 and 1200 for formulation of oxytocin without divalent metal ions (solid line), and with 10 mM Zn2+ (dashed line). Spectra of formulations with calcium and magnesium ions are reported in the supplementary material available.

Nine peaks are observed in the MS profile, which revealed 10 molecular species in the LC-MS contour plot (Figure 3B). The protonated molecular ion of unmodified oxytocin was found at a retention time of 12.8 min with m/z of 1007.4. A peak at a retention time of 14.5 min with m/z 1066.6 corresponded to oxytocin acetate (ammoniated), but also appeared in the LC-MS of the unstressed oxytocin preparation. Most likely, oxytocin acetate was formed during synthesis as the oxytocin is supplied as a monoacetate salt. Complete assignment of the degradation products is listed in Table 1. At retention time of 13.7 and 15.0 min tri- and tetrasulfide form degradation products were identified from the m/z 1039.6 and 1071.6 peaks, respectively (see supplementary material for the structures of tri- and tetrasulfide oxytocin species). Other peaks (Dimers 1 to 6 in Table 1) are interpreted as various forms of disulfide or thioether-linked oxytocin dimers as described previously [8].

80

.

Figure 2 Effect of Ca2+ (squares), Mg2+ (circles), and Zn2+ (triangles) concentration on the

recovery of oxytocin in 10 mM aspartate buffer, pH 4.5, after 4 weeks of storage at either 4°C

(solid symbols) or 55°C (open symbols). A: Oxytocin recovery as determined by RP-HPLC.

B: Oxytocin monomer remaining as determined by HP-SEC. The results are depicted as

averages of three independent measurements ± SD

After 4 weeks of storage at 4°C, oxytocin remained stable in all formulations. Instead,

after 4 weeks of storage at 55°C oxytocin stability increased with increasing divalent

metal ions concentration. Ca2+ and Mg2+ had similar effects on improving the

oxytocin stability: the recoveries of oxytocin, as well as the remaining percentage of

oxytocin monomer were increased up to 45% in the presence of 50 mM Mg2+

, and up

to 35% in the presence of 50 mM Ca2+

. Notably, Zn2+

was much more effective in

stabilizing oxytocin: at a concentration of 2 mM Zn2+

increased oxytocin recovery

up to 35% (i.e. to a comparable extent as 50 mM Ca2+

or Mg2+

), and almost ca. 75%

of oxytocin monomer was recovered. Overall, both RP-HPLC and HP-SEC showed

Figure 2. Effect of Ca2+ (squares), Mg2+ (circles), and Zn2+ (triangles) concentration on the recovery of oxytocin in 10 mM aspartate buffer, pH 4.5, after 4 weeks of storage at either 4°C (solid symbols) or 55°C (open symbols). A: Oxytocin recovery as determined by RP-HPLC. B: Oxytocin monomer remaining as determined by HP-SEC. The results are depicted as averages of three independent measurements ± SD

85

5

Addition of divalent metal salts in the formulation with the aspartate buffer did not result in any additional degradation products compared to samples without metal salts (see dashed line in Figure 3A for zinc salt addition). Conversely, as expected from the RP-HPLC and HP-SEC data, it resulted in a significant reduction of the major degradation peaks. The intensity of each degradation product from each stressed formulation of oxytocin in aspartate buffer in combination with Ca2+, Mg2+, and Zn2+ was also recorded. Table 1 reports the relative intensities in percentage of the peak area of each degradation product in the presence of divalent metal ions with respect to the same formulations without divalent metal ions.Addition of Ca2+ and Mg2+ reduced the formation of dimer 1 and 2 of about 30%, an effect that was even more marked in the presence of Zn2+: i.e., 53 and 60% reduction for dimer 1 and 2, respectively.

Zinc is the most efficient element in suppressing the total amount of the degradation products of oxytocin in aspartate buffer. In fact, at variance with Ca2+ and Mg2+, Zn2+ also reduced the formation of other dimers: dimer 3 was reduced by 14% and dimer 5 was reduced by 30%. However, an increase in the formation of tri and tetrasulfide species was

Figure 3. A: Total ion chromatogram (m/z 900-1200) and B: Contour plot of oxytocin and its degradation products in 10 mM aspartate buffer after 5 days of storage at 70°C and pH 4.5 (solid-line) and in the presence of 10 mM Zn2+ (dashed-line).

86

5


Table 1. Effect of divalent metal ions on the intensity of oxytocin in aspartate buffered solution. Relative peak intensity is expressed with respect to that of the same product in the absence of divalent metal ions.

Assignmentretentiontime (min) m/z

Relative peak intensity of oxytocina (%)

Ca2+ Mg2+ Zn2+

oxytocin trisulfide 13.7 1039.4 111 ± 4 101 ± 8 146 ± 5Oxytocin tetrasulfide 15.0 1071.6 135 ± 13 114 ± 10 187 ± 12dimer 1 15.7 993.2b 68 ± 3 68 ± 2 47 ± 0dimer 2 16.7 993.2b 69 ± 3 76 ± 2 39 ± 2dimer 3 16.9 1008.6 115 ± 4 111 ± 2 86 ± 2dimer 4 17.6 1024.6b 147 ± 19 135 ± 3 127 ± 5dimer 5 18.0 1008.6 111 ± 2 111 ± 1 71 ± 3dimer 6 18.9 1024.6b 152 ± 8 144 ± 3 115 ± 18

a Aspartate buffer solution with 10 mM of the indicated divalent metal ionb ammoniated.

Figure 4. Recovery of oxytocin in the absence (OAP) and presence of 10 mM Ca2+ (OAPCa), Mg2+ (OAPMg) and Zn2+ (OAPZn) in 10 mM aspartate-buffered solution at pH 4.5 under stressed condition at a temperature of 70 °C for 5 days. Oxytocin recovery was determined by LC-MS. The results are depicted as averages of three independent measurements ± SD

also observed. The signal intensity of the tetrasulfide form (m/z 1071.6, retention time 15.0 min) is very low compared to that of the dimers, therefore peak integration is less accurate. We have no explanation for the increased intensity of the trisulfide form (m/z 1039.4 retention time 13.7 min) in the presence of Zn2+.

Oxytocin formulated in aspartate buffer after 5 days at 70°C without divalent metal ions was the most degraded of all formulations tested, and only approximately 20% oxytocin could be recovered (Figure 4). Interestingly, addition of 10 mM Ca2+ and Mg2+ did not markedly improve the peptide recovery. However, when Zn2+ was added, the degradation was reduced and a recovery of about 45% oxytocin was found.

87

5

3.3. Interaction of oxytocin and divalent metal ions in aspartate bufferThe interaction of oxytocin with divalent metal ions was investigated by ITC at a temperature of 55°C. Figure 5 shows the exothermic (DHobs) events when Ca2+ or Mg2+ was titrated into a solution of oxytocin in 10 mM aspartate buffer. However, the effects were very small (not more than 0.05 kcal/mole of injectant) and only slightly different from the corresponding reference measurements. When Zn2+ was titrated into oxytocin in aspartate buffer, a strong endothermic binding reaction was observed and the heat effects upon the titration reached more than 0.5 kcal/mole for the first injection. The shape of the titration curve for the Zn2+-oxytocin in aspartate is indicative for a binding reaction. Using the analysis model with two distinct types of binding sites, the binding constants Ka at 55°C for Zn2+-oxytocin interaction are 6.2 ± 1.0 × 103 M-1 and 72 ± 17 M-1, respectively. Thus, the only system that shows thermodynamics interactions has the most pronounced effect on oxytocin stability in formulations.

4. dIscussIonIn a previous study, we have found that oxytocin can be stabilized by a combination of divalent metal ions and citrate buffer at pH 4.5. Divalent metal ions in combination with citrate buffer suppressed intermolecular reactions in the ring structure of oxytocin presumably by

85

Mg2+

was titrated into a solution of oxytocin in 10 mM aspartate buffer. However, the

effects were very small (not more than 0.05 kcal/mole of injectant) and only slightly

different from the corresponding reference measurements. When Zn2+

was titrated

into oxytocin in aspartate buffer, a strong endothermic binding reaction was observed

and the heat effects upon the titration reached more than 0.5 kcal/mole for the first

injection. The shape of the titration curve for the Zn2+-oxytocin in aspartate is

indicative for a binding reaction. Using the analysis model with two distinct types of

binding sites, the binding constants Ka at 55°C for Zn2+

-oxytocin interaction are 6.2 ±

1.0 × 103 M

-1 and 72 ± 17 M

-1, respectively. Thus, the only system that shows

thermodynamics interactions has the most pronounced effect on oxytocin stability in

formulations.

Figure 5 Least squares fit of the data from calorimetric titration profiles of aliquots of 125

mM divalent metal ions: Ca2+

(solid square), Mg2+

(open square), and Zn2+

(open triangle)

into 5 mM oxytocin in 10 mM aspartate buffer pH 4.5. The heat absorbed per mole of titrant

is plotted versus the ratio of the total concentration of divalent metal ions to the total

concentration of oxytocin.

4 Discussion

In a previous study, we have found that oxytocin can be stabilized by a combination

of divalent metal ions and citrate buffer at pH 4.5. Divalent metal ions in combination

with citrate buffer suppressed intermolecular reactions in the ring structure of

oxytocin presumably by forming a complex in the disulfide bridge region where the

degradation reaction occurred. There were no significant differences observed among

Figure 5. Least squares fit of the data from calorimetric titration profiles of aliquots of 125 mM divalent metal ions: Ca2+ (solid square), Mg2+ (open square), and Zn2+ (open triangle) into 5 mM oxytocin in 10 mM aspartate buffer pH 4.5. The heat absorbed per mole of titrant is plotted versus the ratio of the total concentration of divalent metal ions to the total concentration of oxytocin.

88

5


forming a complex in the disulfide bridge region where the degradation reaction occurred. There were no significant differences observed among the three tested divalent metal ions, Ca2+, Mg2+, and Zn2+ [7]. Our present study indicates that formulations with a combination of divalent metal ions and aspartate buffer behave differently in improving oxytocin stability in aqueous solution. In fact, we have shown that in aspartate buffer, only addition of Zn2+ results in a comparable stabilizing effect on oxytocin as citrate with Ca2+, Mg2+ and Zn2+ [7]. The LC-MS/MS results show that formation of oxytocin dimers is hampered by the presence of Zn2+ in aspartate, while it is not so affected by Ca2+and Mg2+. A previous study by us [8] showed that all the dimers are produced from the thiol exchange in the disulfide bridge, and there are no additional dimers produced from this formulation. Thus, it appears that the combination of Zn2+ and aspartate was able to protect the disulfide Cys1,6 bridge on oxytocin. In line with these results, ITC data demonstrate that only zinc, among the tested divalent metal ions, is able to strongly interact with oxytocin in the formulation conditions.

The observed stabilization effects might be ascribed to the formation of divalent metal ion adducts to oxytocin. In fact, several studies reported on the ability of Ca2+, Mg2+ and Zn2+ to form complexes with oxygen atoms from the carbonyl backbone of the peptide, but the strength of the interactions is different [12-14]. The fact that the three metals have a smaller ionic radius than the oxytocin ring [15,16] suggested that Ca2+, Mg2+ and Zn2+ are located within the ring of oxytocin, and adduct formation, arising from coordination of the metal ions via the backbone carbonyl oxygens of oxytocin, induces a compact structure of an oxytocin-metal octahedral complex [12,17,18].

Most importantly, the binding affinity of oxytocin for Mg2+ has been reported to be lower than in the case of Ca2+ and far below Zn2+ [12,17,18]. Indeed, Zn2+ ions very effectively coordinate with oxytocin and strongly affect its conformation in physiological environment [18].

The oxytocin stabilizing effect of Zn2+ with respect to Ca2+ and Mg2+ might be due to the higher stability of the oxytocin-Zn2+ adduct in the reported formulation conditions.

In addition, our results indicate that also the type of buffer used in the formulations plays an important role in the peptide stabilization effects. At a pH of 4.5, citrate has one carboxylate ion (–COO-) in α position and other two carboxylates that can contribute to the coordination of a metal or to binding to oxytocin itself. At the same pH of 4.5, aspartate has only two carboxylate ions that could act as electron donors towards a metal ion or oxytocin. This difference of available electron donor groups between the two buffers might influence the oxytocin-metal adduct formation, as well as its stabilization. The lack of an additional electron donor oxygen moiety in aspartate is likely to discriminate the binding of Ca2+ or Mg2+ with respect to Zn2+, as suggested by the ITC results, which did not show any significant interactions between oxytocin and Ca2+ or Mg2+ in aspartate buffer.

5. conclusIonOur study clearly shows that the stability of oxytocin in the aspartate buffered-solution is strongly improved in the presence of zinc ions, and the stabilization effect is correlated with the strength of interaction between oxytocin and divalent metal ions. Further studies using NMR and molecular modeling are initiated to characterize the mechanisms of oxytocin stabilization at a molecular level. We can conclude that Zn2+ binding to peptide in aspartate solution suppress intermolecular degradation reactions near the Cys1,6 disulfide bridge that is responsible for oxytocin degradation.

89

acknowledgementThe authors want to thank MSD, Oss, The Netherlands for providing oxytocin for the study. This research was performed within the framework of the Dutch Top Institute Pharma (project number D6–202) and Rosalind Franklin fellowship funding for Angela Casini.

references1. International Confederation of Midwives,

International Federation of Gynaecology and Obstetrics. Joint statement management of the third stage of labour to prevent post-partum haemorrhage, Http://www.Internationalmidwives. org/Whatwedo/ P ro g r a m m e s / P O P P H I / Po s t Pa r t u m -Haemorrhage/tabid/339/Default. Aspx 2012 (2012).

2. A. Ohno, N. Kawasaki, K. Fukuhara, H. Okuda, T. Yamaguchi, Complete NMR analysis of oxytocin in phosphate buffer, Magn. Reson. Chem. 48 (2010) 168-72.

3. J.W. Gard, J.M. Alexander, R.E. Bawdon, J.T. Albrecht, Oxytocin preparation stability in several common obstetric intravenous solutions, Am. J. Obstet. Gynecol. 186 (2002) 496-8.

4. H.V. Hogerzeil, G.J. Walker, Instability of (methyl)ergometrine in tropical climates: an overview, Eur. J. Obstet. Gynecol. Reprod. Biol. 69 (1996) 25-9.




8. C. Avanti, H.P. Permentier, A.V. Dam, R. Poole, W. Jiskoot, H.W. Frijlink, W.L.J. Hinrichs, A New Strategy To Stabilize Oxytocin in Aqueous Solutions: II. Suppression of Cysteine-Mediated Intermolecular Reactions by a Combination of Divalent Metal Ions and Citrate, Mol. Pharmaceutics 9 (3) (2012) 554-62.

9. P. Jurgens, C. Panteliadis, G. Fondalinski, Total parenteral nutrition of premature infants: metabolic effects of an exogenous

supply of L-aspartic acid and L-glutamic acid, Z. Ernahrungswiss. 21 (1982) 225-45.

10. [10] S.C. Gill, P.H. von Hippel, Calculation of protein extinction coefficients from amino acid sequence data, Anal. Biochem. 182 (1989) 319-26.

11. R.A. Poole, P.T. Kasper, W. Jiskoot, Formation of amide- and imide-linked degradation products between the peptide drug oxytocin and citrate in citrate-buffered formulations, J. Pharm. Sci. 100 (2011) 3018-22.


13. J.P. Glusker, A.K. Katz, C.W. Bock, Metal ions in biological systems, The Rigaku Journal 16 (1999) 8-19.

14. H. Einspahr, C.E. Bugg, The Geometry of Calcium-Carboxylate Interactions in Crystalline Complexes, Acta Cryst. B37 (1981) 1044-52.

15. A.K. Katz, J.P. Glusker, S.A. Beebe, C.W. Bock, Calcium Ion Coordination: A Comparison with That of Beryllium, Magnesium, and Zinc, J. Am. Chem. Soc. 118 (1996) 5752-63.

16. R.H. Holm, P. Kennepohl, E.I. Solomon, Structural and Functional Aspects of Metal Sites in Biology, Chem. Rev. 96 (1996) 2239-314.


18. D. Liu, A.B. Seuthe, O.T. Ehrler, X. Zhang, T. Wyttenbach, J.F. Hsu, M.T. Bowers, Oxytocin-receptor binding: why divalent metals are essential, J. Am. Chem. Soc. 127 (2005) 2024-5.

5


Figure S1. Total ion chromatogram (m/z 900-1200) of oxytocin and its degradation products in 10 mM aspartate buffer after 5 days of storage at 70°C and pH 4. in the presence of 10 mM Ca2+.

Figure S2. Total ion chromatogram (m/z 900-1200) of oxytocin and its degradation products in 10 mM aspartate buffer after 5 days of storage at 70°C and pH 4. in the presence of 10 mM Ca2+.

92

5


Figure S3. Structure of oxytocin trisulfide

Figure S4. Structure of oxytocin tetrasulfide

93

Christina Avanti1,*, Nur Alia Oktaviani2,*, Wouter L.J. Hinrichs1, Henderik W. Frijlink1, and Frans A.A. Mulder2,3

#Equally contributed first author

1Department of Pharmaceutical Technology and Biopharmacy, 2Groningen Biomolecular Sciences and Biotechnology Institute,

University of Groningen, Groningen, The Netherlands 3Department of Chemistry and Interdisciplinary Nanoscience Center iNANO,

University of Aarhus, Aarhus C, Denmark

aspartate buffer and dIvalent metal Ions affect the oxytocIn

conformatIon In aqueous solutIon and protect It from degradatIon

6

6

abstractOxytocin is a peptide drug used to induce labor and prevent bleeding after child birth. Due to its instability, transport and storage of oxytocin formulations under tropical condition is problematic. In a previous study, we described that the stability of oxytocin in aspartate buffered formulation is improved by the addition of divalent metal ions. The stabilizing effect of Zn2+ was by far superior compared to that of Mg2+. In addition, it was found that stabilization correlated well with the ability of the divalent metal ions to interact with oxytocin in aspartate buffer. Furthermore, LC-MS (MS) measurements indicated that the combination of aspartate buffer and Zn2+ in particular suppressed intermolecular degradation reactions near the Cys1,6 disulfide bridge. These results lead to the hypothesis that in aspartate buffer, Zn2+ changes the conformation of oxytocin in such a way that the Cys1,6 disulfide bridge is shielded from its environment thereby suppressing intermolecular reactions involving this region of the molecule.

To verify this hypothesis, in this study the conformation of oxytocin in aspartate buffer in the presence of Mg2+ or Zn2+, was investigated by 2D NMR spectroscopy, i.e. NOESY, TOCSY, 1H-13C HSQC and 1H-15N HSQC. Almost all 1H, 13C and 15N resonances could be assigned using HSQC spectroscopy of oxytocin without 13C or 15N enrichment. 1H-13C and 1H-15N HSQC spectra showed that aspartate buffer alone induces a minor change in oxytocin in D2O with the largest chemical shift changes are observed in Cys1. Zn2+ causes more extensive changes in oxytocin in aqueous solution than Mg2+. Our findings suggest that the carboxylate group of aspartate neutralizes the positive charge of the N terminus of Cys1, allowing the interactions with Zn2+ to become more favorable. These interactions may explain the protection of the disulfide bridge against intermolecular reactions that lead to dimerization and inactivation.

6

conforMatIon of dIvalent Metal-aSpartate-oxytocIn coMplex

1. IntroductIonOxytocin is a nonapeptide hormone secreted by the posterior lobe of the pituitary gland which is involved in the control of labor and bleeding cessation after child birth [1]. The peptide consists of nine amino acids (Cys-Tyr-Ile-Gln-Asn-Cys-Pro-Leu- Gly) and an amidated C-terminus [2]. Oxytocin is the preferred drug to prevent postpartum hemorrhage and commonly formulated in aqueous solution for parenteral administration [3]. Instability of oxytocin in aqueous solution under severe conditions, particularly in tropical conditions, presents a significant challenge for pharmaceutical scientists [4]. Oxytocin instability in aqueous solution has been reported in several studies [5,6] and the degradation strongly depends on the pH of the formulation, with the highest stability reported at pH 4.5 [4]. Several studies have been aimed at the improvement of the stability of oxytocin in aqueous solution [4,7]. The most recent finding is the use of divalent metal ions in combination with certain buffers that strongly increases the stability of oxytocin in aqueous solution [8,9].

In a previous study [9] we found that Zn2+ in combination with aspartate buffer strongly stabilizes oxytocin in aqueous solutions, while Ca2+ and Mg2+ only have minor effects. The stabilization occurred as a result of complex formation of Zn2+ ions and aspartate with oxytocin, which, by protecting the Cys1,6 disulfide bridge, suppressed dimerization. In line with those results, isothermal titration calorimetry data demonstrate that only Zn2+ ions, among the tested divalent metal ions (Zn2+, Ca2+, and Mg2+) is able to strongly interact with oxytocin in the formulation conditions [9]. Those results lead to the hypothesis that Zn2+

induces a conformational change, thereby stabilizing oxytocin in aspartate buffer. Aspartic acid is one of the non-essential amino acids which normally synthesized in the body. It consists of two carboxylate groups with pKa1 and pKa2 of 2.1 and 3.9, and one amine group (pKa3 of 9.8). Aspartate is a commonly used buffer in parenteral products approved by the FDA for formulation purposes [10]. To investigate the conformation of oxytocin in aspartate buffer in the presence of divalent metal ions (Zn2+ and Mg2+), two-dimensional Nuclear Magnetic Resonance (NMR) spectroscopy was used.

Nuclear Magnetic Resonance spectroscopy is a suitable technique to study the conformational details of proteins or peptides in solutions. Several one-dimensional NMR studies of oxytocin and vasopressin analogs have previously been performed in various solvents such as dimethyl sulfoxide [13], deuterated dimethylsulfoxide [14,15], deuterated trifuoroethanol [16,17], and aqueous solutions [14,18,19]. Since resonance overlap is much reduced in two-dimensional NMR spectra in comparison with one-dimensional NMR, we used 2D NMR spectroscopy.

Our most recent study [9] showed that Zn2+ in combination with aspartate buffer strongly stabilizes oxytocin in aqueous solutions, while Ca2+ and Mg2+ only have minor effects. The stabilization occurred as a result of complex formation of Zn2+ ions and aspartate with oxytocin, which, by protecting the Cys1,6 disulfide bridge, suppressed dimerization. In line with those results, ITC data demonstrate that only Zn2+ ions, among the tested divalent metal ions (Zn2+, Ca2+, and Mg2+) is able to strongly interact with oxytocin in the formulation conditions [9]. Those results lead to the hypothesis that Zn2+

ions induce a different conformation, thereby stabilizing oxytocin in aspartate buffer. To investigate the conformation of oxytocin in aspartate buffer in the presence of divalent metal ions (Zn2+ and Mg2+), two-dimensional Nuclear Magnetic Resonance (NMR) spectroscopy has been used.

97

6

Nuclear Magnetic Resonance spectroscopy is a suitable technique to study conformational detail of proteins or peptides in solution. Several one-dimensional NMR studies of oxytocin and vasopressin analogs have been previously performed in various solvents such as dimethyl sulfoxide [13], deuterated dimethylsulfoxide [14,15], deuterated trifuoroethanol [16,17], and water [14,18,19]. Since resonance overlap is much reduced in two-dimensional NMR spectra in comparison with one-dimensional NMR, we used 2D NMR spectroscopy.

Oxytocin solutions are commonly formulated in the concentration of 5 IE/mL or approximately 0.01 mM, while the stability studies previously have been done at a concentration of 0.1 mM. The concentration used in this study (10 mM) was higher, since it enables NMR measurements without 13C or 15N enrichment, relying only on the low natural abundance of these isotopes.

A complete NMR analysis of oxytocin in phosphate buffer has been reported [20]. However, no reports are available on the 1H, 13C, and 15N resonance assignments of oxytocin in aspartate buffer in the absence and presence of divalent metal ions. In this study, we used 2D NMR spectroscopy to investigate the conformation of oxytocin in the presence of Zn2+

or Mg2+ in aspartate buffer at pH 4.5.

2. materIals and methods 2.1 MaterialsOxytocin monoacetate powder (Diosynth. Oss, The Netherlands) was kindly provided by MSD, Oss, The Netherlands. Deuterium oxide (D2O, isotopic purity 99.9 atom % D) containing 0.75% TSP(3-(trimethylsilyl) propionic-2,2,3,3,-d4 acid, sodium salt) was purchased from Aldrich, Steinheim, Germany and deuterated L-aspartic acid-2,3,3-d3 was purchased from Medical Isotope, Inc, NH. TSP was used as an internal standard having a chemical shift (d) of 0.0 ppm. Zinc chloride was purchased from Fluka, Steinheim, Germany. All reagents used for the NMR experiments were of analytical grade (purity > 99%), and were used without further purification.

2.2 Sample preparationTwo different types of NMR samples were prepared with the following compositions:1. 2D 13C-1H HSQC NMR samples 10 mM oxytocin (natural abundance) in 10 mM deuterated aspartate buffer (pH 4.5 = pD 5.1)

in D2O containing 0.75% TSP in the absence (OT-AP) and presence of 100 mM ZnCl2 (OT-AP-Zn) or 100 mM MgCl2 (OT-AP-Mg). Reference solution was oxytocin in D2O.

2. 2D 15N-1H HSQC, 1H-1H TOCSY, and 1H-1H NOESY samples. 10 mM oxytocin (natural abundance) in 10 mM aspartate buffer (pH 4.5) in H2O

containing 0.75% TSP in the absence and presence of 100 mM ZnCl2 or 100 mM MgCl2. Reference solution was oxytocin in water.

2.3 Sample preparationSpectra of these samples were recorded using a Varian Unity INOVA 600 MHz NMR spectrometer equipped with pulsed field-gradient probes. The spectra were recorded at 278 K, processed using NMR Pipe [21] and analyzed using Sparky [22]. All information about the NMR measurements is summarized in Table 1.

98

6


3. results 3.1 Assignment and data deposition of backbone and aliphatic side chain resonances of oxytocin in aspartate buffer.The assignment of proton, carbon and nitrogen resonances was accomplished from 2D 13C-1H HSQC, 15N-1H HSQC and 1H-1H TOCSY NMR data. Since the peptide consists of only 9 amino acids, the assignment of 13C-1H HSQC spectra can be easily achieved based on the uniqueness of carbon and proton resonance of each amino acid [28], To avoid disturbance of Ha signals by the strong signal of H2O protons in 13C-1H HSQC experiments, D2O was chosen as solvent.

Amide proton and amide nitrogen resonances in the 15N-1H HSQC were assigned by means of 1H-1H correlations visible in a 1H-1H TOCSY spectrum. Figure 1 summarizes the assignment of most of the oxytocin resonances. All peaks observed in 1H-15N (A) and 1H-13C HSQC spectra (B) are annotated with the one letter amino acid symbol and its position in the sequence.

3.2 NOE analysisTo obtain information about pairs of protons which are close in space, a 2D 1H-1H NOESY spectrum from oxytocin in aspartate buffer in the absence or presence of Zn2+ was recorded. From the NOESY spectrum (Figure 2), it was found that in aspartate buffer residue Ile3 is close to Cys6 and to Asn5. A similar NOESY spectrum was observed in the presence of Zn2+ (see Supporting Information). These interactions schematically presented in Figure 3.

Table 1. Information about the NMR experiments

Experiment Correlation

Number of

scans Nucleus

SpectralWidth (Hz)

Carrier (ppm)

Maximum evolution time (ms) Ref.

15N-1H HSQC

15N and 1H separated by one bond (NH-HN Nε-Hε for Gln, Nd-Hd for Asn, N-H for amide of Gly9)

51215N1H

19448000

120.395.03

20.585

[23][24]

13C-1H HSQC

13C and 1H separated by one bond (Ha-Ca, Hβ-Cβ, Hγ-Cγ, etc.)

4013C1H

80008000

48.191 5.04

16100

[23][25]

1H-1H TOCSYCorrelates all protons in a J-coupled spin system

201H1H

80008000

5.045.04

1685 [26]

1H-1H NOESYCorrelates all protons which are close in space (<0.5 nm)

161H1H

80008000

5.045.04

25160 [27]

99

6

95

Figure 1 2D NMR spectra of oxytocin in

aspartate buffer: 15

N-1H HSQC spectrum which

shows the correlation between amide protons

and amide nitrogens in the backbone of the

peptide. The correlation between amide protons

and amide nitrogens in the side chain of Gln4,

Asn5 and the C-terminal Gly

9 (which has a

carboxy-amide group instead of a regular

carboxylate, indicated by N1-H11 and

N1-H12)are also indicated (A); 13

C-1H HSQC

spectrum shows the correlation between all side

chain carbon signals with signals due to the

attached protons. The Tyr2 Cδ and Cε signals

are aliased in this spectrum. The real Tyr2 Cδ

and Cε chemical shifts are 132.69 and 117. 95

ppm, respectively (B)

Figure 1. 2D NMR spectra of oxytocin in aspartate buffer: 15N-1H HSQC spectrum which shows the correlation between amide protons and amide nitrogens in the backbone of the peptide. The correlation between amide protons and amide nitrogens in the side chain of Gln4, Asn5 and the C-terminal Gly9 (which has a carboxy-amide group instead of a regular carboxylate, indicated by N1-H11 and N1-H12)are also indicated (A); 13C-1H HSQC spectrum shows the correlation between all side chain carbon signals with signals due to the attached protons. The Tyr2 Cd and Cε signals are aliased in this spectrum. The real Tyr2 Cd and Cε chemical shifts are 132.69 and 117. 95 ppm, respectively (B)

100

6


Figure 2. 2D 1H-1H NOESY spectrum of oxytocin in aspartate buffer.

101

6

3.3 Chemical shift difference (Dd)The chemical shift differences induced by the divalent metal ion upon complexation with oxytocin in aspartate buffer were analyzed to get information about which residues are involved in binding of the divalent metal ions to oxytocin and to learn about the extent of the perturbation caused by these ions.

3.3.1 Influence of aspartate buffer on the conformation of oxytocin in water.Aspartate buffer is one of the buffers known to stabilize oxytocin if divalent metal ions are present in the liquid formulation. To investigate the influence of aspartate buffer on the conformation of oxytocin, the chemical shift (d) of Ca and Ha backbone resonances of oxytocin in deuterated aspartate buffer was compared with the chemical shift of Ca and Ha resonances of oxytocin in D2O. The difference between those chemical shifts was expressed as chemical shift difference (Dd) in ppm.

Figure 4A shows that aspartate buffer induced a minor change in oxytocin in D2O. The largest chemical shift changes were observed in the Ca and Hα resonances of Cys1. Ca’s of Cys1 and Cys6 were more shielded and Ca’s of Tyr2 and Ile3 were more deshielded in the presence of aspartate buffer. Ca’s of Gln4, Asn5, Pro7 and Leu8 were not affected by the presence of aspartate buffer. As shown by black bars in Figure 4A, the Ha’s of Cys1, Ile3

Figure 3. Short distances observed in the 2D 1H-1H NOESY spectrum of oxytocin in aspartate buffer.

102

6


and Cys6 are also affected. The opposite effects of aspartate on the Ca and Ha chemical shifts of Cys1 may be due to an electrostatic interaction between the positive charge of the N terminus of Cys1 and the negative charge of the carboxylate group of the aspartate at pD 5.1. In order to test this hypothesis, we recorded 2D 13C-1H HSQC of oxytocin in the presence and absence of aspartate buffer at pD 2.0. The same chemical shifts were found in the two spectra (see Supporting information). This result shows that at a very low pH, where the carboxylate groups of aspartate are totally protonated, there is no effect at Cys1, apparently because the electrostatic interaction between the aspartate and the N terminus of Cys1 has disappeared.

3.3.2 Influence of zinc ions on the conformation of oxytocin in water.Figure 4B shows that Ca and Ha of the amino acid residues of oxytocin other than Leu8

are shifted in the presence of Zn2+. The largest chemical shift changes are observed in Ca of Tyr2. In Tyr2, Gln4, Cys6, Pro7, and Gly9 the Ca spins are more deshielded in the presence of Zn2+. In contrast, in Ile3 and Asn5 the Ca spins are more shielded in the presence of Zn2+. The presence of Zn2+ does not cause changes in the Ca and Ha chemical shifts of Leu8. The largest Dd observed for Ha is in Cys1. Interestingly, similar to the effect of aspartate

Figure 4. Chemical shift difference (Dd) of Ca (light grey bars) and Ha (black bars) of A. oxytocin in deuterated aspartate buffer (pD 5.1) and B. oxytocin in D2O in the presence of Zn2+ relative to the chemical shifts of the same spins measured in D2O. Oxytocin in deuterated aspartate buffer (pD 5.1) in the presence of C. Zn2+ and D. Mg2+ relative to the chemical shifts of the same spins measured in deuterated aspartate buffer pD 5, analyzed by 2D 13C-1H HSQC spectroscopy.

103

6

buffer on oxytocin, in many residues the chemical shift change of Ca induced by Zn2+ is of opposite sign compared to that of Ha.

3.3.3 Influence of divalent metal ions on the conformation of oxytocin in aspartate bufferAs depicted in Figure 4C, the effects of Zn2+ on chemical shift of oxytocin in aspartate are very similar to the effects observed in D2O (Figure 4B), suggesting that a similar conformational change is induced in both circumstances. A small chemical shift change was observed only in the Ca of Cys1.

The effects of Mg2+ on the chemical shifts of Ca and Ha resonances of oxytocin in aspartate buffer are much smaller than those of Zn2+ ions. Strikingly, also in the case of Mg2+, the effects on Ca resonances are opposite to the effects on the Ha resonance of the same residue, except for Cys6 and Leu8. The effects of Zn2+ and Mg2+ ions on the Ca and Ha chemical shifts of oxytocin in the presence of aspartate are displayed in Figure 6.

Figure 5. Overlay of Ca and Ha signals from 13C-1H HSQC spectra of oxytocin in deuterated aspartate buffer without (red) and with Mg2+ (green) (A) or Zn2+ (blue) (B). The Pro7 Cd-Hd correlation is also shown in these spectra.

91

Figure 5 Overlay of Cα and Hα signals from 13

C-1H HSQC spectra of oxytocin in

deuterated aspartate buffer without (red) and with Mg2+

(green) (A) or Zn2+

(blue) (B).

The Pro7 Cδ-Hδ correlation is also shown in these spectra.

4 Discussion

The results of this study indicate that Zn2+

and aspartate buffer changes the

conformation of oxytocin. These conformational changes most likely contribute to the

stabilization of oxytocin in aqueous solution. Our study presents nearly complete

NMR assignment of oxytocin in aspartate buffer in the presence and absence of

104

6


4 dIscussIonThe results of this study indicate that Zn2+ and aspartate buffer changes the conformation of oxytocin. These conformational changes most likely contribute to the stabilization of oxytocin in aqueous solution. Our study presents nearly complete NMR assignment of oxytocin in aspartate buffer in the presence and absence of divalent metal ions, Zn2+ and Mg2+. 2D 1H-1H NOESY spectra of oxytocin in aspartate buffer clearly demonstrate that the residue pairs Ile3-Asn5 and Ile3-Cys6 are in near proximity. These NOEs are also present in a 2D 1H-1H NOESY spectrum of oxytocin in water. Molecular dynamic studies by Wyttenbach et al. [11] suggested an interaction between Ha of Tyr2 and the Gly9 amide group of oxytocin in water, but we could not find evidence for this interaction.

Aspartate buffer induces a minor change in the NMR spectrum oxytocin in D2O with the largest chemical shift changes are observed in Cys1. We suggest that there is an electrostatic interaction between the positively charged N terminus of Cys1 and the negatively charged carboxylate groups of aspartate. This hypothesis is supported by the fact that in pD 2, when aspartate is no longer negatively charged, no changes in chemical shifts are induced by aspartate.

Zn2+ has similar effects on oxytocin in aspartate buffer and in D2O, suggesting that a similar conformational change is induced in both circumstances. In our previous study [7], Zn2+ was shown to improve oxytocin stability only slightly in water, but much more in the presence of aspartate. The small effect of Zn2+ on the chemical shift of the Ca of Cys1, which was observed only in the presence of aspartate, may be relevant in this respect. We suggest that the carboxylate group of aspartate neutralizes the positive charge of the N terminus of Cys1, allowing the interactions with Zn2+ to become more favorable. The arrangement of carbonyl/carboxyl groups around the Zn2+ might also play a role and may explain the observed chemical-shift changes and the protection of the disulfide bridge against intermolecular reactions that lead to dimerization and inactivation.

The observation of chemical-shift changes in Ca and Ha resonances suggest that small changes in the backbone dihedral angles occur to accommodate Zn2+. However drastic changes in the overall structure of the molecule are not likely, since no drastic changes were observed in the pattern of NOEs upon addition of Zn2+ (see supporting information). Increased propensities for the formation of helical or extended secondary-structure elements are not likely, because these would be accompanied by correlated changes in Ca and Ha chemical shifts [29], while in our study we found these changes to be mostly anti-correlated. On the other hand, the presence of a positively charged Zn2+ ion per se will cause polarization of nearby chemical bonds and thus may explain the observed anti-correlation in the Ca and Ha chemical-shift changes.

The effects of Mg2+ on the chemical shifts of Ca and Ha resonances of oxytocin in aspartate buffer are much smaller than those of Zn2+. This result is in agreement with our finding from isothermal titration calorimetry that Mg2+ show only very weak heat effects when added to solutions of oxytocin [9]. The small chemical-shift changes in the backbone of residues Cys1, Tyr2 and Ile3 in the presence of Mg2+ in aspartate buffer may be due to a cation-pi interaction between Mg2+ and the aromatic side chain of Tyr2. However, the interactions between Mg2+ and oxytocin in aspartate buffer are too weak to cause a significant shift in the conformational equilibrium that is essential for protection against inactivation.

105

6

5 conclusIonIn conclusion, our NMR studies revealed that Zn2+ cause a shift in the conformational equilibrium of oxytocin in aqueous solution. Zn2+ causes changes in the chemical shifts of almost all residues of oxytocin while Mg2+ only induces very little chemical-shift changes in some residues. In contrast an analysis of the NOESY spectra in the presence and absence of Zn2+ showed that the same NOEs are present under both circumstances, although with slightly different intensities. The exact nature of the changes induced by Zn2+ is not yet clear. Small changes in the backbone dihedral angles to accommodate the Zn2+ may explain the chemical shift changes, but drastic changes in the overall structure of the molecule are not likely, since no drastic changes in the pattern of NOEs could be observed upon addition of Zn2+. Alternatively, the presence of a positively charged zinc ion per se will cause polarization of nearby chemical bonds and thus may explain the observed chemical shift changes. A definitive explanation of these observations must await a more detailed, dynamic description of the peptide in the absence and presence of Zn2+, possibly from MD simulations steered by distance- and angle-restraints derived from the NMR spectra.

acknowledgmentsThis study was performed within the framework of the Dutch Top Institute Pharma project: number D6–202.

The authors want to thank MSD Oss for providing oxytocin for the study and Ruud M. Scheek at the Groningen Biomolecular Sciences and Biotechnology Institute for helpful discussions.

references1. K.L. Maughan, S.W. Heim, S.S. Galazka,

Preventing postpartum hemorrhage: managing the third stage of labor, Am. Fam. Physician 73 (2006) 1025-8.

2. du Vigneaud V, Ressler C,Trippett S., The sequence of amino acids in oxytocin, with a proposal for the structure of oxytocin, J. Biol. Chem 205 (1953) 949-57.

3. J.W. Gard, J.M. Alexander, R.E. Bawdon, J.T. Albrecht, Oxytocin preparation stability in several common obstetric intravenous solutions, Am. J. Obstet. Gynecol. 186 (2002) 496-8.

4. A. Hawe, R. Poole, S. Romeijn, P. Kasper, R. van der Heijden, W. Jiskoot, Towards heat-stable oxytocin formulations: analysis of degradation kinetics and identification of degradation products, Pharm Res 26 (2009) 1679-88.

5. H.V. Hogerzeil, G.J.A. Walker, M.J. De Goeje, Stability of injectable ocytocics in tropical climates, World Health Organization, Geneva WHO/DAP/93.6 (1993).

6. Trissel LA, Zhang Y, Douglass K,Kastango E., Extended Stability of Oxytocin in

common infusion solution, International Journal of Pharmaceutical Compounding 10 (2006) 156-8.


8. C. Avanti, H.P. Permentier, A.V. Dam, R. Poole, W. Jiskoot, H.W. Frijlink, W.L.J. Hinrichs, A New Strategy To Stabilize Oxytocin in Aqueous Solutions: II. Suppression of Cysteine-Mediated Intermolecular Reactions by a Combination of Divalent Metal Ions and Citrate, Mol. Pharmaceutics 9 (3) (2012) 554-62.

9. C. Avanti, W.L.J. Hinrichs, A. Casini, A.C. Eissens, A.V. Dam, A. Kedrov, A.J.M. Driessen, H.W. Frijlink, H.P. Permentier, Insight into the stability of the zinc-aspartate-oxytocin formulation, manuscript in preparation.

10. P. Jurgens, C. Panteliadis, G. Fondalinski, Total parenteral nutrition of premature

106

6


infants: metabolic effects of an exogenous supply of L-aspartic acid and L-glutamic acid, Z. Ernahrungswiss. 21 (1982) 225-45.

11. T. Wyttenbach, D. Liu, M.T. Bowers, Interactions of the hormone oxytocin with divalent metal ions, J. Am. Chem. Soc. 130 (2008) 5993-6000.

12. A.F. Pearlmutter, M.S. Soloff, Characterization of the metal ion requirement for oxytocin-receptor interaction in rat mammary gland membranes, J. Biol. Chem. 254 (1979) 3899-906.

13. E. Sikorska, M.J. Slusarz, B. Lammek, Conformational studies of vasopressin analogues modified with N-methylphenylalanine enantiomers in dimethyl sulfoxide solution, Biopolymers 82 (2006) 603-14.

14. J.D. Glickson, R. Rowan, T.P. Pitner, J. Dadok, A.A. Bothner-By, R. Walter, 1H nuclear magnetic resonance double resonance study of oxytocin in aqueous solution, Biochemistry 15 (1976) 1111-9.

15. L.F. Johnson, I.L. Schwartz, R. Walter, Oxytocin and neurohypophyseal peptides spectral assignment and conformational analysis by 220 M1Hz nuclear magnetic resonance, Proc. Nat. Acad. Sci 64 (1969) 1269-75.

16. M. Rholam, P. Cohen, N. Brakch, L. Paolillo, A. Scatturin, C. Di Bello, Evidence for beta-turn structure in model peptides reproducing pro-ocytocin/neurophysin proteolytic processing site, Biochem, Biophys, Res, Commun, 168 (1990) 1066-73.


18. E. Sikorska, S. Rodziewicz-Motowidlo, Conformational studies of vasopressin and mesotocin using NMR spectroscopy and molecular modelling methods. Part I: Studies in water, J. Pept. Sci. 14 (2008) 76-84.

19. I.C. Smith, R. Deslauriers, H. Saito, R. Walter, C. Garrigou-Lagrange, H. McGregor,

D. Sarantakis, Carbon-13 NMR studies of peptide hormones and their components, Ann. N. Y. Acad. Sci. 222 (1973) 597-627.

20. A. Ohno, N. Kawasaki, K. Fukuhara, H. Okuda, T. Yamaguchi, Complete NMR analysis of oxytocin in phosphate buffer, Magn. Reson. Chem. 48 (2010) 168-172.

21. F. Delaglio, S. Grzesiek, G.W. Vuister, G. Zhu, J. Pfeifer, A. Bax, NMR pipe - a multidimensional spectral processing system based on unix pipes., J Biomol NMR 6 (1995) 277-293.

22. T.D. Goddard, D.G. Kneller, SPARKY 3, University of California, San Francisco, 2003.

23. L.E. Kay, P. Keifer, T. Saarinen, Pure Absorption Gradient Enhanced Heteronuclear Single Quantum Spectroscopy with Improved Sensitivity, J. Am. Chem. Soc. 114 (1992) 10633-5.

24. F.A. Mulder, R. Otten, R.M. Scheek, Origin and removal of mixed-phase artifacts in gradient sensitivity enhanced heteronuclear single quantum correlation spectra, J. Biomol. NMR 51 (2011) 199-207.

25. K.B. John, D. Plant, R.E. Hurd, Improved Proton-Detected Heteronuclear Correlation Using Gradient-Enhanced z and zz filters, J. Magn. Reson. A101 113-7.

26. D.G. Davis, A. Bax, Assignment of Complex 1H NMR spectra via Two-Dimensional Homonuclear Hartman-Hahn Spectroscopy, J. Am. Chem. Soc. 107 (1985) 2820-1.

27. J. Boyd, G.R. Moore, G. Williams, Correlation of proton chemical shifts in proteins using two-dimensional exchange correlated spectroscopy, J. Magn. Reson. 58 (1984) 511-6.

28. C. Cavanagh, W.J. Fairbrother, A.G. Palmer III, N.J. Skelton (Eds.), Protein NMR Spectroscopy, Second Edition: Principles and Practice, 2nd ed., Elsevier Inc., United State of America, 2006.

29. D.S. Wishart, Interpreting protein chemical shift data, Prog Nucl Magn Reson Spectrosc 58 (2011) 62-87.

107

6


Table S1. Resonance list of oxytocin in aspartate buffer

Group Atom Nuc Shift SDev Assignments

C1 Ca 13C 54.963 0 1C1 Cβ 13C 42.593 0.009 2C1 Ha 1H 4.225 0.013 2C1 Hβ2 1H 3.279 0 1C1 Hβ3 1H 3.477 0 1Y2 Ca 13C 58.13 0 1Y2 Cβ 13C 38.843 0.009 2Y2 Cd 13C 132.686 0 1Y2 Cε 13C 117.951 0 1Y2 H 1H 9.118 0.009 10Y2 Ha 1H 4.781 0 1Y2 Hβ2 1H 3.002 0.005 4Y2 Hβ3 1H 3.2 0.007 4Y2 Hd 1H 7.247 0.017 5Y2 Hε 1H 6.875 0.006 6Y2 N 15N 123.84 0 1I3 Ca 13C 62.55 0 1I3 Cβ 13C 38.694 0 1I3 Cd 13C 13.49 0 1I3 Cγ 13C 27.28 0.021 2I3 Cγ1 13C 17.67 0 1I3 H 1H 8.103 0.052 16I3 Ha 1H 4.073 0.011 3I3 Hβ 1H 1.94 0.002 3I3 Hd 1H 0.888 0.009 4I3 Hγ1 1H 0.901 0.004 2I3 Hγ2 1H 1.011 0.004 2I3 Hγ3 1H 1.237 0.003 4I3 N 15N 120.224 0 1Q4 Ca 13C 57.793 0 1Q4 Cβ 13C 28.472 0 1Q4 Cγ 13C 33.688 0 2Q4 H 1H 8.368 0.016 12Q4 Ha 1H 4.116 0.005 3Q4 Hβ 1H 2.079 0.004 3Q4 Hε21 1H 6.989 0.013 6Q4 Hε22 1H 7.705 0.012 8

continued next page

108

6



Q4 HG2 1H 2.41 0.003 2Q4 HG3 1H 2.426 0.003 2Q4 N 15N 120.399 0 1Q4 Nε 15N 112.68 0.001 2N5 Ca 13C 53.044 0 1N5 Cβ 13C 38.413 0 1N5 H 1H 8.455 0.006 13N5 Ha 1H 4.767 0.022 5N5 Hβ 1H 2.878 0.003 2N5 Hd21 1H 7.068 0.01 8N5 Hd22 1H 7.759 0.009 7N5 N 15N 116.465 0 1N5 Nd 15N 112.933 0.004 2C6 Ca 13C 54.019 0 1C6 Cβ 13C 40.798 0.004 2C6 H 1H 8.332 0.006 11C6 Ha 1H 4.914 0.022 2C6 Hβ2 1H 3.005 0.064 3C6 Hb3 1H 3.245 0.003 2C6 N 15N 119.907 0 1P7 Ca 13C 63.203 0 1P7 Cβ 13C 32.035 0.002 2P7 Cd 13C 50.569 0.001 2P7 Cγ 13C 27.467 0 1P7 Ha 1H 4.46 0.004 2P7 Hβ2 1H 1.941 0 1P7 Hβ3 1H 2.314 0 2P7 Hd2 1H 3.744 0.011 2P7 Hd3 1H 3.774 0 1P7 Hγ 1H 2.049 0 1L8 Ca 13C 55.346 0 1L8 Cβ 13C 41.873 0.005 2L8 Cd1 13C 24.903 0 1L8 Cd2 13C 23.345 0 1L8 Cγ 13C 27.046 0 1L8 H 1H 8.705 0.012 11L8 Ha 1H 4.313 0.004 3L8 Hβ2 1H 1.619 0.005 2L8 Hβ3 1H 1.713 0 1L8 Hd1 1H 0.950 0.004 3L8 Hd2 1H 0.911 0.014 2L8 Hγ 1H 1.699 0.004 3L8 N 15N 122.909 0 1

continued next page

109

6


G9 Ca 13C 44.757 0.006 2G9 H 1H 8.589 0.007 7G9 H11 1H 7.51 0.005 6G9 H12 1H 7.215 0.004 10G9 Ha1 1H 3.875 0.004 2G9 Ha2 1H 3.957 0.01 3G9 N 15N 111.181 0 1G9 N1 15N 107.471 0.002 2

Table S2. Resonance list of oxytocin in the presence of Zn2+ in aspartate buffer


C1 Ca 13C 55.064 0 1C1 Cβ 13C 43.085 0.011 2C1 Ha 1H 4.317 0.044 2C1 Hβ2 1H 3.295 0.008 2C1 Hβ3 1H 3.554 0.057 2Y2 Ca 13C 58.649 0 1Y2 Cβ 13C 38.689 0.005 2Y2 Cd 13C 132.905 0 1Y2 Cε 13C 118.028 0 1Y2 Ha 1H 4.748 0.029 2Y2 Hβ2 1H 3.022 0.021 4Y2 Hβ3 1H 3.189 0.003 4Y2 Hd 1H 7.235 0.011 7Y2 Hε 1H 6.872 0.006 6Y2 N 15N 123.932 0 1I3 Ca 13C 61.979 0 1I3 Cβ 13C 38.842 0 1I3 Cd 13C 13.663 0 1I3 Cγ 13C 27.233 0.001 2I3 Cγ1 13C 17.826 0 1I3 H 1H 8.131 0.009 15I3 Ha 1H 4.098 0.025 3I3 Hβ 1H 1.945 0.011 3I3 Hd 1H 0.884 0.011 4I3 Hγ1 1H 0.894 0.002 3I3 Hγ2 1H 1.042 0.021 2I3 Hγ3 1H 1.234 0.006 4I3 N 15N 120.225 0 1Q4 Ca 13C 57.986 0 1

continued next page

110

6



Q4 Cβ 13C 28.54 0 1Q4 Cγ 13C 33.781 0 2Q4 H 1H 8.374 0.027 13Q4 Ha 1H 4.096 0.017 3Q4 Hβ 1H 2.07 0.005 3Q4 Hε21 1H 6.995 0.017 6Q4 Hε22 1H 7.704 0.014 7Q4 Hγ2 1H 2.399 0.009 2Q4 Hγ3 1H 2.423 0.005 2Q4 N 15N 120.391 0 1Q4 Nε 15N 112.749 0.006 2N5 Ca 13C 52.976 0 1N5 Cβ 13C 38.542 0 1N5 H 1H 8.456 0.006 13N5 Ha 1H 4.758 0.012 4N5 Hβ 1H 2.88 0.003 3N5 Hd21 1H 7.068 0.009 6N5 N 15N 116.474 0 1N5 Nd 15N 113.046 0 2C6 Ca 13C 54.282 0 1C6 Cβ 13C 40.909 0.001 2C6 H 1H 8.33 0.005 11C6 Ha 1H 4.878 0.022 2C6 Hβ2 1H 2.97 0.01 2C6 Hβ3 1H 3.264 0.004 2C6 N 15N 119.856 0 1P7 Ca 13C 63.595 0 1P7 Cβ 13C 32.115 0 2P7 Cd 13C 50.779 0.003 2P7 Cγ 13C 27.628 0 1P7 Ha 1H 4.449 0.004 2P7 Hβ2 1H 1.936 0 1P7 Hβ3 1H 2.308 0 2P7 Hd2 1H 3.738 0.017 2P7 Hd3 1H 3.77 0 1P7 Hγ 1H 2.036 0 1L8 Ca 13C 55.351 0 1L8 Cβ 13C 41.604 0.003 2L8 Cd1 13C 25.116 0 1L8 Cd2 13C 23.383 0 1L8 Cγ 13C 27.151 0 1L8 H 1H 8.699 0.005 10L8 Ha 1H 4.308 0.007 3

continued next page

111

6


L8 Hβ2 1H 1.611 0.01 3L8 Hβ3 1H 1.768 0 1L8 Hd1 1H 0.941 0.032 3L8 Hd2 1H 0.879 0.01 2L8 Hγ 1H 1.688 0.022 3L8 N 15N 122.855 0 1G9 Ca 13C 44.726 0.002 2G9 H 1H 8.572 0.006 7G9 H11 1H 7.51 0.004 5G9 H12 1H 7.217 0.003 9G9 Ha1 1H 3.877 0.011 2G9 Ha2 1H 3.94 0 1G9 N 15N 111.157 0 1G9 N1 15N 107.5 0.016 2

Table S3. Resonance list of oxytocin in the presence of Mg2+in aspartate buffer


C1 Ca 13C 54.873 0 1C1 Cβ 13C 42.513 0 2C1 Ha 1H 4.287 0.015 2C1 Hβ2 1H 3.295 0.011 2C1 Hβ3 1H 3.51 0.003 2Y2 Ca 13C 58.233 0 1Y2 Cβ 13C 38.786 0.002 2Y2 Cd 13C 132.706 0 1Y2 Cε 13C 117.98 0 1Y2 H 1H 9.133 0.005 12Y2 Ha 1H 4.771 0.003 2Y2 Hβ2 1H 3.012 0.008 3Y2 Hβ3 1H 3.184 0.004 4Y2 Hd 1H 7.233 0.009 4Y2 Hε 1H 6.872 0.007 7Y2 N 15N 123.963 0 1I3 Ca 13C 62.434 0 1I3 Cβ 13C 38.714 0 1I3 Cd 13C 13.585 0.084 2I3 Cγ 13C 27.312 0 2I3 Cγ1 13C 17.739 0.076 2I3 H 1H 8.1 0.063 13

continued next page

112

6



I3 Ha 1H 4.074 0.008 3I3 Hβ 1H 1.938 0.003 3I3 Hd 1H 0.877 0.007 7I3 Hγ1 1H 0.894 0.005 3I3 Hγ2 1H 1.018 0.008 2I3 Hγ3 1H 1.23 0.006 4I3 N 15N 120.247 0 1Q4 Ca 13C 57.874 0 1Q4 Cβ 13C 28.457 0 1Q4 Cγ 13C 33.668 0 2Q4 H 1H 8.359 0.013 10Q4 Ha 1H 4.108 0.002 2Q4 Hβ 1H 2.069 0.006 3Q4 Hε21 1H 6.974 0 2Q4 Hε22 1H 7.696 0.01 6Q4 Hγ2 1H 2.404 0.003 2Q4 Hγ3 1H 2.435 0 1Q4 N 15N 120.407 0 1N5 Ca 13C 53.012 0 1N5 Cβ 13C 38.448 0 1N5 H 1H 8.452 0.007 10N5 Ha 1H 4.756 0.01 4N5 Hβ 1H 2.881 0.005 3N5 Hd21 1H 7.065 0.009 4N5 Hd22 1H 7.763 0 5N5 N 15N 116.458 0 1N5 Nd 15N 113.06 0.008 2C6 Ca 13C 54.063 0 1C6 Cβ 13C 40.784 0.007 2C6 H 1H 8.325 0.004 7C6 Ha 1H 4.897 0.004 2C6 Hβ2 1H 2.96 0.006 2C6 Hβ3 1H 3.262 0.008 2C6 N 15N 119.877 0 1P7 Ca 13C 63.284 0 1P7 Cβ 13C 32.052 0.002 2P7 Cd 13C 50.61 0 2P7 Cγ 13C 27.497 0 1P7 Ha 1H 4.453 0.006 2P7 Hβ2 1H 1.938 0 1P7 Hβ3 1H 2.312 0 1P7 Hd2 1H 3.745 0.004 2P7 Hd3 1H 3.749 0 1

continued next page

113

6


P7 Hγ 1H 2.044 0 1L8 Ca 13C 55.332 0 1L8 Cβ 13C 41.813 0.003 2L8 Cd1 13C 24.926 0 1L8 Cd2 13C 23.365 0 1L8 Cγ 13C 27.07 0 1L8 H 1H 8.693 0.003 9L8 Ha 1H 4.309 0.006 3L8 Hβ2 1H 1.622 0.007 2L8 Hβ3 1H 1.725 0 1L8 Hd1 1H 0.932 0.029 3L8 Hd2 1H 0.904 0.002 2L8 Hγ 1H 1.695 0.008 2L8 N 15N 122.836 0 1G9 Ca 13C 44.806 0.001 2G9 H 1H 8.564 0.001 5G9 H12 1H 7.214 0.003 9G9 Ha1 1H 3.87 0.006 2G9 Ha2 1H 3.942 0 1G9 N 15N 111.19 0 1G9 N1 15N 107.554 0.009 2

Table S4. Resonance List of oxytocin in water


C1 Ca 13C 55.228 0 1C1 Cβ 13C 42.996 0.004 2C1 Ha 1H 4.199 0.051 3C1 Hβ2 1H 3.22 0 1C1 Hβ3 1H 3.392 0 1Y2 Ca 13C 58.032 0 1Y2 Cβ 13C 38.834 0 2Y2 Cd 13C 132.686 0 1Y2 Cε 13C 117.951 0 1Y2 H 1H 9.125 0.004 11Y2 Ha 1H 4.78 0 1Y2 Hβ2 1H 3.002 0.006 4Y2 Hβ3 1H 3.21 0.039 5Y2 Hd 1H 7.244 0.015 5Y2 Hε 1H 6.876 0.005 6Y2 N 15N 123.788 0 1

continued next page

114

6



I3 Ca 13C 62.415 0 1I3 Cβ 13C 38.757 0 1I3 Cd 13C 13.49 0 2I3 Cγ 13C 27.259 0.002 2I3 Cγ1 13C 17.67 0 2I3 H 1H 8.109 0.055 16I3 Ha 1H 4.072 0.012 3I3 Hβ 1H 1.941 0.006 3I3 Hd 1H 0.885 0.008 5I3 Hγ1 1H 0.902 0.005 3I3 Hγ2 1H 1.01 0.015 2I3 Hγ3 1H 1.228 0.007 4I3 N 15N 120.158 0 1Q4 Ca 13C 57.793 0 1Q4 Cγ 13C 33.683 0.002 2Q4 H 1H 8.371 0.023 12Q4 Ha 1H 4.114 0.006 3Q4 Hβ 1H 2.078 0.005 3Q4 Hε21 1H 6.989 0.014 6Q4 Hε22 1H 7.704 0.009 8Q4 HG2 1H 2.413 0.002 2Q4 HG3 1H 2.423 0.009 2Q4 N 15N 120.41 0 1Q4 Nε 15N 112.664 0.005 2N5 Ca 13C 53.044 0 1N5 Cβ 13C 38.356 0 1N5 H 1H 8.458 0.006 13N5 Ha 1H 4.763 0.016 5N5 Hβ 1H 2.877 0.003 2N5 Hd22 1H 7.756 0.01 7N5 N 15N 116.463 0 1N5 Nd 15N 112.86 0.001 2C6 Ca 13C 54.093 0 1C6 Cβ 13C 40.903 0.007 2C6 H 1H 8.337 0.004 11C6 Ha 1H 4.895 0.015 2C6 Hβ2 1H 3.02 0.088 3C6 Hb3 1H 3.244 0.007 2C6 N 15N 119.936 0 1P7 Ca 13C 63.203 0 1P7 Cβ 13C 32.034 0.002 2P7 Cd 13C 50.57 0.002 2P7 Cγ 13C 27.467 0 1

continued next page

115

6


P7 Ha 1H 4.46 0.003 2P7 Hβ2 1H 1.94 0 1P7 Hβ3 1H 2.312 0.002 2P7 Hd2 1H 3.744 0.011 2P7 Hd3 1H 3.774 0 1P7 Hγ 1H 2.049 0 1L8 Ca 13C 55.346 0 1L8 Cβ 13C 41.873 0.005 2L8 Cd1 13C 24.879 0 1L8 Cd2 13C 23.383 0 1L8 Cγ 13C 27.046 0 1L8 Ha 1H 4.311 0.005 3L8 Hβ2 1H 1.657 0.042 2L8 Hβ3 1H 1.713 0 1L8 Hd1 1H 0.958 0.005 3L8 Hd2 1H 0.911 0 1L8 Hγ 1H 1.696 0.003 3L8 N 15N 122.962 0 1G9 Ca 13C 44.763 0 2G9 H 1H 8.592 0.013 8G9 H11 1H 7.515 0.003 6G9 H12 1H 7.213 0.005 10G9 Ha1 1H 3.872 0.005 2G9 Ha2 1H 3.955 0.009 3G9 N 15N 111.18 0 1G9 N1 15N 107.426 0.005 2

116

6


4.5

4.5

4.0

4.0

3.5

3.5

1H (ppm)

60 60

55 55

50 50

45 45

13C

(pp

m)

N5Cα-Hα

C6Cα-Hα

L8Cα-Hα

C1Cα-Hα

Y2Cα-Hα

I3Cα-HαP7Cα-Hα

G9Cα-Hα

Q4Cα-Hα

P7Cδ-Hδ

5.0

5.0

Figure S1. Overlay of Ca and Ha signals from 13C-1H HSQC spectra of oxytocin in D2O pD 2 (red) and oxytocin in deuterated aspartate buffer pD 2 (green).

117

6

1H (ppm)

1 H (

ppm

)

Y2H-H

Y2Hα-H

C1Hα-Y2H

C1Hβ3-Y2H

C1Hβ2-Y2H

Y2Hβ2-H

Y2Hβ3-H

L8H-H

P7Hα-L8H

L8Hα-H

L8Hδ-H

L8Hγ-H

Q4H-H

N5Hα-Q4H

P7Hδ2-C6H

C6Hβ3-H

C6Hβ2-H

Q4Hβ-N5H

N5Hβ-H

Q4Hα-N5H

I3Hα-Q4HL8Hα-G9H

I3Hγ1-Q4H

I3Hδ-HI3Hγ2-H

I3Hα-H

N5Hα-H

C6Hα-H

I3H-H

Y2H-I3H

N5Hδ22-Hδ21

N5H-H

G9H-H

I3Hδ-N5H

I3Hβ-H

Q4Hβ-H

I3Hβ-Q4HL8Hδ-N5H

Q4Hγ2-H

Y2Hβ3-I3H

G9Hα1-H

I3Hγ3-Q4H

N5Hδ22-Hδ22

Q4Hε22-Hε22

G9H11-H11

N5Hδ21-Hδ22

G9H12-H11 G9H12-H12Y2Hε-Hε

Q4Hε21-Hε22

N5Hβ2-Hδ22

Q4Hε21-Hε22

N5Hδ21-Q4Hε22

Y2Hα-Hδ

Y2Hβ3-Hδ

Y2Hβ2-Hδ

N5Hβ-Hδ21

P7Hβ3-L8H

C6H-H

I3Hγ3-HI3Hγ3-N5H

L8Hβ2-H

N5Hβ3-C6H

I3H-Y2H I3H-Q4HI3H-C6H

Y2Hε-Hδ

G9H11-H12

Y2Hβ2-I3H

Q4Hε21-Hε21

N5Hδ21-Hδ21

Q4Hα-C6H

N5H-C6H

Q4H-N5H

Y2Hδ-Y2Hε

Y2Hδ-Hδ

L8Hβ2-G9H

Q4Hε22-Hε21Q4H-I3H

I3Hδ-Y2Hδ I3Hγ1-Y2Hε

Y2H-Hδ

6.57.07.58.08.59.09.5

0

2

4

6

8

10

Figure S2. 2D 1H-1H NOESY spectrum of oxytocin in the presence of Zn2+ in aspartate buffer.

118

Christina Avanti1,*, Vinay Saluja1,2,*, Erwin L.P. van Streun1, Imma Boyten1, Henderik W. Frijlink1, Wouter L.J. Hinrichs1

*Equally contributed first author

1Department of Pharmaceutical Technology &Biopharmacy, University of Groningen, Groningen, The Netherlands

2Vaccinology, National Institute for Public Health and the Environment, Bilthoven, The Netherlands

extremolytes: are there unIversal stabIlIzers for proteIns In aqueous

solutIon?

7

7

abstractThe purpose of this study was to investigate the ability of extremolytes to stabilize the model proteins lysozyme and insulin in aqueous solutions. The effects of the extremolytes, betaine, hydroxyectoine, trehalose, ectoine, and firoin on the stability of lysozyme were determined by Nile red Fluorescence Spectroscopy and a bioactivity assay. Insulin stability was determined by RP-HPLC and HP-SEC. The effects of extremolytes on the unfolding temperature of the proteins were analyzed using a thermal shift assay for lysozyme and liquid differential scanning microcalorimetry for insulin. The interaction between extremolytes and protein was studied by isothermal titration calorimetry (ITC). During storage at 70ºC for 10 min, firoin protected lysozyme against inactivation better than the other extremolytes. During storage at 55ºC for 4 weeks, firoin also acted as a stabilizer, however, betaine, hydroxyectoine, trehalose and ectoine, destabilized lysozyme. These findings surprisingly indicate that some extremolytes can stabilize proteins under certain stress conditions but destabilize the same proteins under other stress conditions. The increased stability caused by firoin is explained by the observed increased unfolding temperature of lysozyme in the presence of firoin. After storage at 40ºC for 4 weeks, trehalose and ectoine protected insulin against degradation, while betaine, hydroxyectoine and in particular firoin seemed to destabilize insulin. Furthermore, it was shown that firoin sharply decreased the unfolding temperature of insulin. The interaction of firoin with lysozyme and ectoine or trehalose with insulin was negligible as determined by ITC. Thus, firoin appears to be an excellent stabilizer for lysozyme but strongly destabilizes insulin, whereas ectoine showed the opposite behaviour. This study clearly shows that there is not one extremolyte that can acts as a universal stabilizer for proteins in aqueous solution.

7

extreMolyteS aS proteInS StabIlIzerS

1. IntroductIonProtein instability is one of the main issues in the administration of therapeutic protein based medicines, in particular in aqueous formulations. Protein stability can be achieved if there is a balance between intramolecular interaction of protein functional groups and their interaction with solvent environment [1]. A number of experimental studies have been done to overcome protein instability. One of the most promising results is the discovery of extremolytes, small organic osmolytes found in extremophiles.

Extremophiles are microorganisms which are capable of surviving under extreme conditions, such as high or low temperatures, extreme pressure and high salt concentrations. Extremolytes are accumulated in response to these extreme conditions and minimize the denaturation of the biological macromolecules and proteins [2]. The polyols derivatives ectoine and hydroxyectoine are the first extremolytes that are currently produced in a large scale and are already being used as protein stabilizers. Furthermore, also other molecules such as betaine[3] various amino acids, carbohydrates such as trehalose[4]and the mannose derivative firoin[5,6] have been found in extremophyles and have been identified as stabilizers in these species. Figure 1 shows the molecular structures of some of these extremolytes.

Extremolytes stabilize proteins by forming solute hydrate clusters [7] that are preferentially excluded from the hydrate shell of the protein [8]. Exclusion occurs as a result of the repulsive interactions between extremolytes and the backbone of the proteins [9] and generates accumulation of water near protein domains. Thus, proteins turn into a more compact tertiary structure with a reduced surface area.

According to this mechanism, in the presence of extremolytes the native state of proteins is in the lower free energy state than in the unfolded state. Proteins lose their conformational

Figure 1. Molecular structure of the extremolytes (a) mannosylglycerate (firoin), (b) betaine, (c) ectoine, (d) hydroxyectoine, and (e) and trehalose

123

7

entropy with the greater entropic loss in the unfolded state (Su), leading to an overall shift in equilibrium towards the native state [10]. The entropy difference between the two states increases, thus the folded protein (Sf) is less likely to unfold as ilustrated in Figure 3 [11].

The aim of this study was to investigate whether known extremolytes such as betaine, hydroxyectoine, trehalose, ectoine, and firoin are able to stabilize the model proteins lysozyme and insulin at elevated temperatures and whether they can be used as universal stabilizers for proteins in aqueous solutions.

2. materIals and methods2.1 MaterialsThe following materials were used in this study: Hen egg-white lysozyme (Sigma-Aldrich, Steinheim, Germany), Recombinant human insulin, USP (provided by MSD, Oss, The Netherlands), ultrapure trehalose(Cargill, Krefeld, Germany), ultrapure betaine (Fluka Biochemika, Buchs, Switzerland), firoin(Biotop, Berlin-Brandenburg, Germany), ultrapure ectoine and hydroxyectoine(Biomol, Hamburg, Germany), citric acid (Riedel-de Haen, Seelze, Germany), sodium hydroxide, acetonitrile (Merck, Darmstadt, Germany), dimetylsulfoxide(DMSO), Nile red and M. Lysodeickticus(Sigma-Aldrich, Steinheim, Germany), PBS buffer (66 mM phosphate, pH 6.2), SYPRO orange protein gel stain, (invitrogenEugene, Oregon, USA), sodium sulphate, concentrated phosphoric acid, ethanolamine, hydrochloric acid (Merck, Darmstad, Germany), phosphate buffered saline (Fluka Analytical, Steinheim, Germany).

2.2 The effect of extremolytes on the stability of lysozyme

2.2.1 Heat shock stability studyLysozyme solutions at a concentration of 100 μg/ml were prepared in 10 mM citrate buffer (pH 5.0) with and without extremolytes at a concentration of 0.5 M [2,5]. Lysozyme solutions were incubated at 70ºC for 10 minutes. The effect of extremolytes on the stability of lysozyme was determined by Nile red fluorescence spectroscopy and by measuring its bioactivity.

Figure 2. Mode of action extremolytes, adapted from Lentzen G., 2006 (2)

124

7


2.2.2 Accelerated stability studiesLysozyme solutions at a concentration of 100 μg/ml were prepared in 10 mM citrate buffer (pH 5.0) with and without 0.5 M stabilizers. Lysozyme solutions were incubated at 55ºC. The effect of extremolytes on the stability of lysozyme was determined by measuring its bioactivity. Bioassays were performed every week for 4 weeks.

2.2.3 Nile red Fluorescence SpectroscopyFluorescence studies were performed on a SLM-Aminco AB2 Spectrofluorometer at the temperature of 25°C using 5 mm cubical quartz cuvette (Hellma GmbH). Prior to measurement, 5 μl of 20 μg/ml Nile red solution in DMSO was added to a 1 ml of 100 μg/ml lysozyme solution, to obtain a Nile red : lysozyme weight ratio of 1:1000. An excitation wavelength of 550 nm was used, with a band pass of 4.0 nm for the excitation monochromator and an emission wavelength of 610 nm with a band pass of 4.0 nm for the emission monochromator. Data were recorded at 1 nm intervals over the range 560–700 nm with a scanning speed of 100 nm/min. Spectra were corrected for background signal caused by buffer and stabilizers [12].

2.2.4 BioassayThe bioactivity of lysozyme was determined by measuring the rate of lysis of Micrococcus lysodeikticus by using a method as described by Shugar[13] with some modifications. Briefly, 1.3 ml of a 200 µg/ml of Micrococcus lysodeikticus suspension in sterile PBS buffer (66 mM phosphate, pH 6.2) was mixed with 10 µl of the lysozyme solutions in plastic disposable cuvettes. Immediately after mixing, the cuvette was placed in a UV/VIS spectrophotometer and the absorbance was measured at a wavelength of 450 nm. Subsequently, the change of the absorbance was recorded over time. Remaining activity of the stressed samples was obtained by comparing them with the unstressed samples. Statistical analyses were performed using Student’s t test with p < 0.05 as the minimal level of significance. The results are presented as mean ± standard error mean unless indicated otherwise.

2.3 The effect of extremolytes on the stability of insulinInsulin was formulated at a concentration of 20 IE/ml in 2 mM PBS buffer at pH 7.4 with and without extremolytes at a concentration of 0.5 M. After preparation, the solutions were stored at either 4°C or 40°C for 4 weeks, and protected from light. The pH of the samples was measured regularly and was found to be remained at 7.4 ± 0.1 during the stability study.

Figure 3. Proposed mechanism of stabilization of extremolytes, adapted from Arakawa, 2006 (12)

125

7

2.3.1 Reversed-Phase High-Performance Liquid Chromatography (RP-HPLC)The recovery of insulin was determined by RP-HPLC as described in the USP. A Chrompack C18 column with 5 µm particle size, inner diameter of 4.6 mm and length of 250 mm, a Waters 510 HPLC pump, a Waters 717 Autosampler and a Waters 486 tunable absorbance UV detector were used. Samples of 10 µl were injected and the separation was carried out at a flow rate of 1.0ml/min and UV detection at 214 nm. Samples were eluted using 27% acetonitrile in buffer pH 2.3. The buffer for the mobile phase was prepared by dissolving 71.0 g of sodium sulphate in 2100 ml of Milli Q water. After adding 6.75 ml of phosphoric acid, the pH was adjusted to 2.3 with ethanolamine or 0.5 N phosphoric acid. The total volume was completed to 2500 mL and filtered through a 0.45µm filter. Integration of the chromatograms was done with Chromeleon software. The runtime was 30.5 minutes [14].

2.3.2 High-Performance Liquid-Size Exclusion Chromatography (HP-SEC)The monomeric fraction of insulin was assessed by HP-SEC[15]. HP-SEC was carried out using a Waters Insulin HMWP column, inner diameter of 7.8 mm, length 200 mm, Waters 510 HPLC pump, Waters 717 plus autosampler, Waters 486 tunable absorbance UV detector (Waters, Milford Massachusetts, USA)was used. Samples of 20 µl were injected, and separation was performed at a flow rate of 0.5 ml/min. Peaks were detected by UV absorption at 276 nm. Samples were eluted using mobile phase consisted of 40% of L-arginine 1 mg/ml solution, 50% of acetonitrile and 10% glacial acetic acid. The runtime was 30.5 minutes.

2.4 The effect of extremolytes on protein unfolding temperature (Tm)

2.4.1 Thermal shift assayThe unfolding temperature of lysozyme in solutions was analyzed by a thermal shift assay using a real-time PCR machine [16]. Solutions of 17.5 μl of 1.0 mg/ml lysozyme with or without stabilizers (1 M) and 7.5 μl of 300 fold diluted SYPRO Orange solution as a molecular probes were added to the wells of a 96-well thin-wall PCR plate (Bio-Rad). The plates were sealed with optical-quality sealing tape (Bio-Rad Laboratories BV, Veenendaal, The Netherlands), inserted into a real-time PCR machine (iCycler, Bio-Rad Laboratories BV, Veenendaal, The Netherlands) and heated from 20 to 90°C with a 0.2°C increaseper 20 s. The fluorescence changes of the SYPRO Orange probe in the wells of the plate were monitored simultaneously with a fluorescence detector (MyIQ single-colour RT-PCR detection system, Bio-Rad) at excitation and emission wavelengths of 490 and 575 nm, respectively. The midpoint of the transition was taken as the Tm.

2.4.2 Liquid Differential Scanning calorimetryThe Tmof insulin in solutions in a concentration of 1.5 mg/mL with or without extremolytes (0.1 M) was analyzed by differential scanning microcalorimetry. Data collection was performed using a VP-DSC differential scanning microcalorimeter (MicroCal, LLC, Northampton, MA) [17,18]. All insulin scans were performed with 2 mM PBS buffer pH 7.4 in the reference cell from 25 to 110°C at a scan rate of 90°C/hour and an excess pressure of 25-26 Psi. All samples and references were degassed immediately before use. A buffer-buffer reference scan was subtracted from each sample scan prior to concentration normalization. Data analysis was carried out using Origin 7.0 (OriginLab, Northampton, MA).

126

7


2.5 Interaction of extremolytes with protein determined by Isothermal Titration Calorimetry (ITC)Microcalorimetric titrations of extremolytes to lysozyme in citrate buffer pH 5.0 and insulin in PBS buffer were performed by using a MicroCal ITC 200 microcalorimeter (Northampton, MA 01060 USA). A solution of 20 mM of some selected extremolytes in 10 mM citrate buffer, pH 5.0 was placed in the syringe, while 300 µl of 1 mM lysozyme in 10 mM of citrate buffer, pH 5.0 was placed in the sample cell. The reference cell contained 300 µl of citrate buffer. Experiments were performed at 10, 25, and 55°C. The initial delay time was 60 s. The reference power and the filter were set to 5 µcal/s and 2 s respectively. A typical titration experiment consisted of 20 injections of 2 µl extremolytes solutions with duration of 4 s and the time interval between two consecutive injections was set to 180 s. During the experiments, the sample solution was continuously stirred at 1000 rpm. The effective heat of the protein-extremolytes interaction upon each titration step was corrected for dilution and mixing effects, as measured by titrating the extremolytes solution into the buffer and by titrating the buffer into the protein solution. The heats of bimolecular interactions were obtained by integrating the peak following each injection. All measurements were performed in triplicate. ITC data were analyzed by using the ITC non-linear curve fitting functions for one or two binding sites from MicroCal Origin 7.0 software (MicroCal Software, Inc.) [19].

3. results and dIscussIon3.1 The effect of extremolytes on the stability of proteins In order to get a clear and unambiguous picture of the effect of the different extremolytes on the stability of the two proteins it was decided to test the stability effects on each of the proteins with two different methods. For the lysozyme the Nile Red Fluorescence Spectroscopy and the activity assay were used, whereas for the insulin the RP-HPLC and HP-SEC were combined. In all studies the extremolytes concentration of 0,5 M was used.

3.1.1 Stability of lysozyme as measured by Nile Red Fluorescence SpectroscopyNile red fluorescence can be employed to probe changes in protein conformations that are related to the formation of hydrophobic surfaces, such as during aggregation or protein unfolding because of its sensitivity to the polarity of its environment [20,21].

The effect of a heat shock on lysozyme without extremolytes in citrate buffer solution pH 5.0 is shown in Figure 4A. A huge increase in the fluorescence intensity of Nile red was observed when lysozyme without extremolytes as stressed at 70°C for 10 minutes indicating substantial denaturation of lysozyme. Apparently, heating lysozyme without extremolytes for 10 minutes at 70 ºC caused collapse of its secondary and/or tertiary structure. Stressed lysozyme solutions in the presence of betaine(Figure 4B) or hydroxyectoine (Figure 4C) showed similar Nile red fluorescence spectra. However, when trehalosewas added, we observed a slight difference in the Nile red fluorescence spectra of the lysozyme sample before and after heat shock. Figure 4D shows that there was a minor shift in the maximum peak after stress, however, the huge increase in the fluorescence intensities of Nile red as found for lysozyme formulations without extremolytes or in the presence betaine or hydroxyectoine was not observed. This indicates that trehalose was able to inhibit substantial denaturation of lysozyme. The stabilization effect of trehalose might be due to the fact that

127

7

Figure 4. The effect of extremolytes on the tryptophanyl fluorescence of the lysozyme-nile red complex after stressing lysozyme at 70°C for 10 minutes in citrate buffer pH 5.0. Fluorescence spectra recorded of lysozyme unstressed (solid line) and stressed (dotted line) of lysozyme A: without extremolytes, and with B: betaine, C: hydroxyectoine, D: trehalose, E: ectoine, and F: firoin

128

7


trehalose has a propensity to depress the formation of aggregates and chemical reactions of lysozyme by inducing α-helical structures and some tertiary structures [22].

Also the Nile red fluorescence spectrum of the lysozyme solution formulated with ectoine showed a minor shift in the maximum peak after stress (Figure 4E). Furthermore, also significantly higher fluorescence intensity was observed in the stressed lysozyme samples containing ectoine than those containing trehalose. However, this increased intensity was much smaller than that found for formulations without extremolytes or in the presence betaine or hydroxyectoine. Therefore, these measurements indicate that ectoine protected lysozyme against denaturation, however, to a somewhat lesser extent than trehalose.

In conclusion, trehalose and ectoineare able to partially prevent the denaturation of lysozyme,. In contrast, when firoin was added to lysozyme solution, as the Nile red fluorescence spectra of unstressed and stressed lysozyme solution were fully identical (Figure 4 F) indicating that firoin completely inhibited the denaturation of lysozyme during heat shock.

3.1.2 Stability of lysozyme determined by using BioassayIn order to further evaluate the stabilizing effects of extremolytes during a heat shock at a temperature of 70ºC for 10 minutes and during storage at a temperature of 55ºC for 4 weeks, the biological activity of lysozyme was determined. After heat shock the ability of lysozyme to inactivate the bacterium M. Lysodeickticus decreased dramatically.

Figure 5 shows that lysozyme only maintained about 20-40% of its original activity. When betaine, trehalose, and ectoine were added, however, lysozyme maintained about 70% of its original activity. There was no significant difference observed between the stabilizing effects of hydroxyectoine, trehalose, and ectoine but it is difficult to draw a conclusion from these data since the level of significance was low. It seems, however, that ectoine stabilized lysozyme better than trehalose and hydroxyectoine, also hydroxyectoine was not significantly different from control.

Figure 5. The effect of lysozyme on M. Lysodeickticus (bioactivity) after stressed at 70°C for 10 minutes and the effect of extremolytes on the bioactivity of lysozyme. * is the level of significance

129

7

The bioactivity assay was also used to monitor the effects of extremolytes on the stability of lysozyme solution during incubation for 4 weeks at 55°C (accelerated stability study).

Figure 6 shows that the addition of the extremolytes betaine, hydroxyectoine, trehalose and ectoine destabilized lysozyme. These results are in contrast to the heat shock experiments where these extremolytes lead to minor or substantial stabilizing effects. On the other hand, the bioactivity assay indicated that firoin stabilized lysozyme during storage at a 55°C for 4 weeks, which is completely in line with the results of the heat shock experiments.

3.1.3 Stability of insulin as measured by RP-HPLC and HP-SECFigure 7A shows the results of insulin recovery analyzed by RP-HPLC method after incubation at 4°C and 40°C for 4 weeks. In PBS buffer alone approximately 95% insulin was recovered after 4 weeks at 4°C. The same result was obtained when hydroxyectoine, trehalose or ectoine was added to the solution. When incubated at 40°C for 4 weeks approximately 30% insulin was recovered from the PBS solution.. Addition of firoin, betaine and hydroxyectoine showed a destabilizing effect as these samples showed no insulin recovery at all after 4 weeks at 40°C. However, ectoine and trehalose showed significant stabilizing effects, resulting in an insulin recovery of 62% and 72%, respectively

Figure 7B shows the effects of extremolytes on the amount of insulin monomer remaining after incubation at 4°C and 40°C for 4 weeks as measured by HP-SEC. At 4°C, about 90% insulin remained as monomer in the absence or presence of extremolytes, although firoin resulted in a lower extent of monomer remaining. The results after incubation at 40°C for 4 weeks are in line with the RP-HPLC measurements. In the absence of extremolytes, insulin showed substantial degradation resulting in a remaining amount of insulin monomer of about 35%. The HP-SEC results confirm that betaine, hydroxyectoine and firoin destabilized insulin. However, ectoine and trehalose strongly stabilized insulin, resulting in a recovery of 80% and 88% insulin monomer, respectively.

Figure 6. The effect of extremolytes on the bioactivity of lysozyme during 4 weeks storage at 55 °C

130

7


In conclusion, the RP-HPLC and HP-SEC measurements indicate that trehalose and ectoine are able to protect insulin against chemical and physical degradation in liquid formulations, whereas betaine or hydroxyectoine and in particular firoin destabilized insulin.

3.2 The effect of extremolytes on protein unfolding temperature (Tm) The unfolding temperature (Tm) of lysozyme in solution was analyzed by a thermal shift assay using a real-time PCR machine. The Tm for lysozyme was found to be 82°C. While, the addition of firoin, resulted in a much higher increase in Tm i.e. 9°C.

These results may explain the improved stability of lysozyme in the presence of firoin. The firoin effect may be due to the increase of the melting temperature of lysozyme by 9ºC. It may be that a rise in Tm of 4ºC is not enough to protect lysozyme for a longer period of time at 55ºC, but that a rise of 9-ºC is enough to stabilize lysozyme for at least a week at 55ºC. Santoro et al showed that up to 8.2 molar of extremolytes, including betaine, were able to increase the Tm of lysozyme to about 23ºC [23]. It is possible that higher extremolytes concentrations are required to improve the stability of lysozyme during storage at a high temperature.

Figure 9 shows the Tm of insulin in the absence and presence of extremolytes as determined by liquid differential calorimetry. A slight but insignificant increase of the Tm of insulin was found when ectoine was added to the solution. Addition of trehalose and betaine had no effect on the Tm of insulin, whereas hydroxyectoine and in particular firoin even decrease the Tm of insulin. The decreased Tm found for firoin may explain the destabilizing effect of this extremolyte on insulin.

Figure 7. The effect of stabilizers on the concentration of insulin in a liquid formulation after storage at 4°C (light grey bars) or 40°C (dark grey bars) for 4 weeks determined by A. RP-HPLC, B HP-SEC

131

7

3.3 Interaction of proteins with extremolytesThe interaction of extremolytes with proteins was analyzed using Isothermal Titration Calorimetry at different temperatures for the most stable formulations.

Figure 10 shows the similarity in the magnitude of the reaction and dilution enthalpy for the titration of lysozyme by firoin and the titration of insulin by ectoine or trehalose at the temperatures of 10, 25, or 55°C. The results indicate that in aqueous solution extremolytes do not show any significant interaction with the proteins. The stabilizing effects are rather due to modification of the water properties. This is in line with the preferential hydration theory that the repulsion between the amide backbone of the protein and the extremolytes is due to the influence of the extremolytes on the water structure and therefore do not interact directly with the proteins [2].

Figure 8. The effect of stabilizers on the unfolding temperature (Tm) of lysozyme. Unfolding temperatures was measured as transition midpoint analyzed by thermal shift assay using RT-PCR machine of lysozyme on the concentration of 1.0 mg/ml with A: betaine, B: hydroxyectoine, C: trehalose, D: ectoine, and E: firoin

Figure 9. The effect of stabilizers on the unfolding temperature (Tm) of insulin. Unfolding temperatures was measured as transition midpoint analysed by liquid differential calorimetry of insulin on the concentration of 1.5 mg/ml with A: betaine, B: hydroxyectoine, C: trehalose, D: ectoine, and E: firoin

132

7


Figure 10. The effect of firoin (A) on the heat capacity of lysozyme, and the effect of ectoine (B) and trehalose (C) to the heat capacity of insulin determined by ITC

4 conclusIonThis study clearly shows that there are no extremolytes that can be used as a universal stabilizer for all proteins in aqueous solution. We even found that certain extremolytes, (firoin) can act as a stabilizer for a particular protein (lysozyme), but destabilizes another protein (insulin). Even worse, certain extremolytes (betaine) can stabilize a protein (lysozyme) under certain conditions (heat shock) but destabilize the same protein under other stress conditions (accelerated stability conditions). This implies not only that for each protein the stabilizing effects of different extremolytes should be screened but also that the envisaged storage conditions should be taken into account. Furthermore, for screening extremolytes for the stabilization of proteins, measuring the Tm of the protein can be useful to predict stabilizing effects but only if the extremolyte induces a substantial change in the Tm.

acknowledgment This study was performed within the framework of the Dutch Top Institute Pharma (projectnumber D6−202).

The authors want to thank G.K. Schuurman-Wolters, A. Kedrov and Prof. A.J.M. Driessen at the Groningen Institute Biomolecular Sciences & Biotechnology Groningen for helpful discussions on L-DSC and ITC.

133

7

references

1. G. Nemethy, Orientation of amino acid side chains: intraprotein and solvent interactions, Methods. Enzymol. 127 (1986) 183-96.

2. G. Lentzen, T. Schwarz, Extremolytes: Natural compounds from extremophiles for versatile applications, Appl. Microbiol. Biotechnol. 72 (2006) 623-34.

3. S. Knapp, R. Ladenstein, E.A. Galinski, Extrinsic protein stabilization by the naturally occurring osmolytes beta-hydroxyectoine and betaine, Extremophiles 3 (1999) 191-8.

4. A. Hedoux, J.F. Willart, L. Paccou, Y. Guinet, F. Affouard, A. Lerbret, M. Descamps, Thermostabilization mechanism of bovine serum albumin by trehalose, J. Phys. Chem. B 113 (2009) 6119-26.

5. T.Q. Faria, S. Knapp, R. Ladenstein, A.L. Macanita, H. Santos, Protein stabilisation by compatible solutes: effect of mannosylglycerate on unfolding thermodynamics and activity of ribonuclease A, ChemBioChem 4 (2003) 734-41.

6. N. Borges, A. Ramos, N.D. Raven, R.J. Sharp, H. Santos, Comparative study of the thermostabilizing properties of mannosylglycerate and other compatible solutes on model enzymes, Extremophiles 6 (2002) 209-16.

7. E.A. Galinski, M. Stein, B. Amendt, M. Kinder, The kosmotropic (structure-forming) effect of compensatory solutes, Comp. Biochem. Physiol. 117A (1997) 357-65.

8. T. Arakawa, S.N. Timasheff, The stabilization of proteins by osmolytes, Biophys. J. 47 (1985) 411-4.

9. D.W. Bolen, G.D. Rose, Structure and Energetics of the Hydrogen-Bonded Backbone in Protein Folding, Annu. Rev. Biochem. 77 (2008) 339-62.

10. A. Linhananta, G. Amadei, T. Miao, Computer simulation study of folding thermodynamics and kinetics of proteins in osmolytes and denaturants, Journal of Physics: Conference Series 341 (2012).

11. T. Arakawa, D. Ejima, Y. Kita, K. Tsumoto, Small molecule pharmacological chaperones: From thermodynamic stabilization to pharmaceutical drugs, Biochim. Biophys. Acta 1764 (2006) 1677-87.

12. M. Sutter, S. Oliveira, N.N. Sanders, B. Lucas, A. van Hoek, M.A. Hink, A.J. Visser, S.C. De Smedt, W.E. Hennink, W. Jiskoot, Sensitive spectroscopic detection of large

and denatured protein aggregates in solution by use of the fluorescent dye Nile red, J. Fluoresc. 17 (2007) 181-92.

13. D. Shugar, The measurement of lysozyme activity and the ultra-violet inactivation of lysozyme, Biochim. Biophys. Acta 8 (1952) 302-9.

14. X. Xu, Y. Fu, H. Hu, Y. Duan, Z. Zhang, Quantitative determination of insulin entrapment efficiency in triblockcopolymeric nanoparticles by high-performance liquid chromatography, J. Pharm. Biomed. Anal. 41 (2006) 266-73.

15. C.M. Yu, C.Y. Chin, E.I. Franses, N.H. Wang, In situ probing of insulin aggregation in chromatography effluents with spectroturbidimetry, J. Colloid. Interface Sci. 299 (2006) 733-9.

16. U.B. Ericsson, B.M. Hallberg, G.T. Detitta, N. Dekker, P. Nordlund, Thermofluor-based high-throughput stability optimization of proteins for structural studies, Anal. Biochem. 357 (2006) 289-98.

17. V.V. Plotnikov, J.M. Brandts, L.N. Lin, J.F. Brandts, A new ultrasensitive scanning calorimeter, Anal. Biochem. 250 (1997) 237-44.

18. K. Huus, S. Havelund, H.B. Olsen, M. van de Weert, S. Frokjaer, Thermal dissociation and unfolding of insulin, Biochemistry 44 (2005) 11171-7.

19. C. Hoffmann, A. Blume, I. Miller, P. Garidel, Insights into protein-polysorbate interactions analysed by means of isothermal titration and differential scanning calorimetry, Eur. Biophys. J. 38 (2009) 557-68.

20. A. Hawe, M. Sutter, W. Jiskoot, Extrinsic fluorescent dyes as tools for protein characterization, Pharm. Res. 25 (2008) 1487-99.

21. D.L. Sackett, J. Wolff, Nile red as a polarity-sensitive fluorescent probe of hydrophobic protein surfaces, Anal. Biochem. 167 (1987) 228-34.

22. T. Ueda, M. Nagata, T. Imoto, Aggregation and chemical reaction in hen lysozyme caused by heating at pH 6 are depressed by osmolytes, sucrose and trehalose, J. Biochem 130 (2001) 491-6.

23. M.M. Santoro, Y. Liu, S.M. Khan, L.X. Hou, D.W. Bolen, Increased thermal stability of proteins in the presence of naturally occurring osmolytes, Biochemistry 31 (1992) 5278-83.

134

summary, concludIng remarks and global perspectIve

&

summaryThe general objective of this thesis is to develop liquid formulations for polypeptide based medicines that are stable under tropical condition.

Chapter 2 of this thesis discusses peptide instabilities in aqueous solutions and strategies to improve peptide stability in parenteral formulations. The most prevalent degradation pathways at accelerated temperature and certain pH are hydrolysis including deamidation, isomerization and racemization, oxidation including light and metal-induced oxidation, β-elimination, disulfide exchange, dimerization and further aggregation. Adequate knowledge about amino acid residues involved in peptide degradation can be used to develop strategies to improve peptide stability in aqueous solutions. The most common approaches to protect peptide degradation in aqueous formulations are formulating peptide at an optimum pH, the proper choice of buffers, antioxidants, chelating agents, removal of oxygen, protection from light, the use of metal ions, organic solvents, or surfactants.

The focus of this thesis is on the stabilization of oxytocin (a nonapeptide) in formulations for injection. The major approach has been to investigate the stabilization of this peptide by divalent metal ions in combination with a certain buffer. The effect of monovalent and divalent metal ions in combination with buffers at a pH of 4.5 on the stability of oxytocin in aqueous solutions is described in Chapter 3. The chloride salts of monovalent metal ions (Na+ and K+) and divalent metal ions (Ca2+, Mg2+, and Zn2+) in combination with citrate or acetate buffer were investigated. Utilizing RP-HPLC and HP-SEC the effect of combinations of buffers and metal ions on the stability of aqueous oxytocin solutions after 4 weeks of storage at either 4°C or 55°C was quantified. Addition of the monovalent ions, Na+ and K+, to acetate- or citrate-buffered solutions did not increase oxytocin stability, nor did the addition of divalent metal ions to acetate buffered solutions. However, the stability of oxytocin in aqueous formulations was improved in the presence of 5 and 10 mM citrate buffer in combination with at least 2 mM CaCl2, MgCl2, or ZnCl2 and depended on the divalent metal ion concentration. Isothermal titration calorimetric (ITC) measurements were predictive for the stabilization effects observed during the stability study. Formulations in citrate buffer that had an improved stability displayed a strong interaction between oxytocin and Ca2+, Mg2+, or Zn2+, whereas formulations in acetate buffer did not. In conclusion, our study showed that divalent metal ions in combination with citrate buffer strongly improved the stability of oxytocin in aqueous solutions.

In chapter 4, various degradation products of oxytocin in citrate-buffered solution after thermal stress at a temperature of 70°C for 5 days were identified. The changes in degradation pattern caused by the presence of divalent metal ions were also determined. Degradation products of oxytocin in the citrate buffered formulations with and without divalent metal ions were analyzed using liquid chromatography−mass spectrometry/mass spectrometry (LC−MS/MS). In the presence of divalent metal ions, almost all degradation products, in particular citrate adduct, tri- and tetrasulfide, and dimers, were greatly reduced in intensity. Formulations containing divalent metal ions in combination with citrate buffer suppressed the formation of degradation products N-citryl oxytocin, tri/tetrasulfide and dimers. The suppression occurred on the disulfide bridge between Cys1 and Cys6. Cysteine is susceptible to oxidation and β-elimination, and degradation of oxytocin involving cysteine leads to dimerization, formation of tri/tetrasulfide and β-elimination followed by thio-ether formation. No significant difference in the stabilizing effects was found among the divalent metal ions Ca2+, Mg2+, and Zn2+, suggesting that divalency is the most important

138

&

SuMMary, concludInG reMarkS and Global perSpectIve

property of the metal contributing to stabilization of the oxytocin-metal-citrate cluster. The suppressed degradation pathways all involve the cysteine residues. We therefore postulate that cysteine-mediated intermolecular reactions are suppressed by complex formation of the divalent metal ion and citrate with oxytocin, thereby inhibiting the formation of citrate adducts and reactions of the cysteine thiol group in oxytocin. Citrate has two opposite effects on oxytocin stability. First, it is reactive itself and can attack the N-terminal amino group from the cysteine residue to form an adduct. Secondly, it protects oxytocin from degradation in the presence of divalent metal ions.

In chapter 5, the effect of divalent metal ions (Ca2+, Mg2+ and Zn2+) on the stability of oxytocin in aspartate buffer (pH 4.5) is described and their interaction with the peptide in aqueous solution was determined. RP-HPLC and HP-SEC results indicated that after 4 weeks of storage at 55°C all tested divalent metal ions improved the stability of oxytocin in aspartate buffered solutions. However, the stabilizing effects of Zn2+ were by far superior compared to both Ca2+ and Mg2+. LC-MS/MS results showed that the combination of aspartate and Zn2+ in particular suppressed the formation of peptide dimers. As shown by ITC, Zn2+ interacted with oxytocin in the presence of aspartate buffer while Ca2+ or Mg2+ did not. In conclusion, the stability of oxytocin in the aspartate buffered-solution is strongly improved by the presence of zinc ions, and the stabilization effect is correlated with the ability of the divalent metal ions in aspartate buffer to interact with oxytocin. Aspartate presumably sequesters the zinc ion inside the ring structure of oxytocin protecting the intramolecular disulfide bridge from reactions leading to degradation.

In chapter 6, the conformation of oxytocin in aspartate buffer in the presence of Mg2+ or Zn2+, was investigated by 2D NMR spectroscopy, i.e. NOESY, TOCSY, 1H-13C HSQC and 1H-15N HSQC. Almost all 1H, 13C and 15N resonance could be assigned using HSQC spectroscopy of oxytocin with neither 13C nor 15N enrichment. 1H-13C and 1H-15N HSQC spectra showed that Zn2+ caused changes in the chemical shifts of almost all amino acid residues of oxytocin, whereas Mg2+ only induces minor chemical shift changes in several but not all amino acid residues. On the other hand, NOESY spectra exhibited almost the same NOEs in the presence and absence of Zn2+. Our findings indicate the carboxylate group of aspartate neutralizes the positive charge of the N terminus of Cys1, allowing the interactions with Zn2+ to become more favorable. These interactions may explain the protection of the disulfide bridge against intermolecular reactions that lead to dimerization and inactivation.

Chapter 7, describes a study on the ability of various extremolytes to stabilize model proteins (lysozyme and insulin) in aqueous solution. The effects of the extremolytes, betaine, hydroxyectoine, trehalose, ectoine, and firoin on the stability of lysozyme were determined by Nile red Fluorescence Spectroscopy and a bioactivity assay. Insulin stability was determined by RP-HPLC and HP-SEC. The effects of extremolytes on the unfolding temperature of the proteins were analyzed using a thermal shift assay for lysozyme and liquid differential scanning microcalorimetry for insulin. The interaction between extremolytes and proteins was studied by ITC. During storage at 70°C for 10 min, firoin protected lysozyme against inactivation better than the other extremolytes. During storage at 55°C for 4 weeks, firoin also acted as a stabilizer, however, betaine, hydroxyectoine, trehalose and ectoine, destabilized lysozyme. These findings indicate that some extremolytes can stabilize proteins under certain stress conditions but destabilize the same proteins under other stress conditions. The increased stability caused by firoin is explained by the observed increased unfolding temperature of lysozyme in the presence of firoin. After storage at 40°C

139

&

for 4 weeks, trehalose and ectoine protected insulin against degradation, while betaine, hydroxyectoine and in particular firoin were found to destabilize insulin. Furthermore, it was shown that firoin sharply decreased the unfolding temperature of insulin. The interaction of firoin with lysozyme and ectoine or trehalose with insulin was negligible as determined by ITC. Thus, firoin appears to be an excellent stabilizer for lysozyme but strongly destabilizes insulin, whereas ectoine showed the opposite behaviour. This study clearly shows that there is not one extremolyte that can acts as a universal stabilizer for proteins in aqueous solution. Therefore, to develop stable protein formulations using extremolytes, one should consider the envisaged storage conditions besides screening the stabilizing effects of different extremolytes. Furthermore, for screening extremolytes for the stabilization of proteins, measuring the Tm of the protein can be useful to predict stabilizing effects but only if the extremolyte induces a substantial change in the Tm.

concludIng remarks In this thesis an attempt is made to develop a heat-stable liquid oxytocin formulation. Several formulations containing specific combinations of divalent metal ions with buffers were found to stabilize oxytocin in aqueous solution. Different analytical tools such as LC-MS (MS), ITC and NMR were utilized to gain information on the mechanism of stabilization at a molecular level. However, it is still difficult to correlate the results with the mechanism by which buffers and divalent metal ions may contribute to the stabilization of oxytocin.

Although the overall picture was rather complex and mechanisms varied between the different ions and buffers, it is clear that stabilization of the Cys1,6 disulfide bridge in the oxytocin molecule is of paramount importance for stabilization of the peptide in aqueous solution. This was best illustrated in an NMR study showing that the carboxylate group of aspartate neutralizes the positive charge of the N terminus of Cys1, allowing the interactions with Zn2+ to become more favorable. These interactions may explain the protection of the disulfide bridge against intermolecular reactions that lead to dimerization and inactivation.

Zn2+ causes a shift in the conformational equilibrium of oxytocin, which is evident from the observed changes in the chemical shifts of almost all resonances, but the exact nature of these changes is not clear yet. A final explanation of these observations must await a more detailed, dynamic description possibly from molecular dynamic simulations steered by distance- and angle-restraints derived from the NMR spectra.

Molecular dynamic simulations could be an interesting approach to provide insight on a molecular level in the stabilization mechanisms of the different divalent metal ions and buffers. With this type of studies it may be elucidated which conformational changes are caused by the different metal ions and how these conformational changes are affected by the buffers in the solution. When furthermore a relation between the oxytocin conformation and its stability can be established, it may be feasible to define an optimum combination of metal ion and buffer that keeps the oxytocin in the most stable conformation. Another approach that may finally result in a stable oxytocin formulation is to use the knowledge gained in this thesis in a more or less guided extensive trial-and-error development formulation study.

Cleavage of the disulfide bridge was found to be the major degradation pathway for oxytocin, and certain carboxylate buffers were found to support complex formation with certain divalent metal ions, which lead to stabilization. Especially the effects of different carboxylate buffers in combination with different divalent metal ions may be an interesting

140

&

SuMMary, concludInG reMarkS and Global perSpectIve

topic in such a study. Many different carboxylate buffers exist and only three were tested in the work described in this thesis. The complexity of the role of the buffer is nicely illustrated by the dual role that was found for citrate, being both a stabilizer and an adduct former, the differences found for the different ions in aspartate buffer and the absence of these differences in citrate or acetate buffer. It is unknown whether other carboxylate buffers may have better stabilizing properties on oxytocin and how their effect can be further improved with different metal ions. In this respect it is interesting to mention here that a recent study in our group revealed that malonate buffer had a higher stabilizing effect than aspartate or citrate buffer. At this moment a real time stability study with a formulation containing this buffer and zinc ions is ongoing. The six months results that were obtained so far are promising and this formulation may be suitable for use in tropical developing countries.

global perspectIveAs described in the introduction of this thesis the starting point for our research was the insufficient access of mothers in the tropical developing countries to oxytocin, which leads to a yearly death rate of approximately 150.000 mothers. The availability of a suitable, affordable, liquid oxytocin formulation would potentially prevent their death. When starting the research we assumed that the fact that none of the current oxytocin formulations could be stored outside the cold chain was a major reason for this problem and that a heat stable formulation would prevent these deaths. This starting point and the results of our studies raise two major questions:1. Was a liquid heat stable oxytocin formulation developed in this study?2. What is the best way forward to reduce maternal death due to oxytocin insufficient

availability in tropical regions, i.e. sub-Saharan Africa? Unfortunately, thus far we were not able to find a liquid-oxytocin formulation which

meets the standard requirement of the pharmacopoeia i.e. storage stability for at least one year at 40°C or two years at 30°C. However, the formulations, which involve zinc ions and malonate buffers, did demonstrate promising results. But as long as real time stability data for a period of at least one to two years are not available, definite conclusions cannot be drawn.

From a technical point of view several adequate solutions to solve the immediate need for a heat stable oxytocin formulation are available:1. In those places where a refrigerator is not available, storage of the current oxytocin

formulation at ambient conditions is feasible, as long as the expired material is simply replaced every six months.

2. Introduce refrigerators on solar energy, in those places where they are currently not available.

3. A stable freeze dried oxytocin formulation (to be reconstituted before use) has been developed by us in collaboration with MSD and could be introduced immediately.

4. The development of a stable spray dried oxytocin formulation for pulmonary administration.

5. The introduction of the developed formulations involving zinc ions and malonate buffer is most likely the best solution. However, assessment on the real time stability for at least another six months (until November 2012) is required before drawing a definite conclusion.

141

&

When these five solutions are evaluated the first two solutions require funding to replace the oxytocin stocks every half a year and to install energy power systems.

With regard to the third option, a fast introduction would not only depend on the willingness to accept the related increase in process costs and costs of goods. Much more important will be the willingness of the responsible authorities to accept such a product on their markets, as a replacement for the current product, without a full registration dossier. If such a dossier is to be prepared the costs will rise tremendously and they will surely be too high for any of the relevant countries in this project. It should in this respect also be realized that the proposed replacement from a scientific point of view would not have any relevant associated safety risk since it only replaces one parenteral formulation for another parenteral formulation.

The fourth solution involves an alternative dosage form. A pulmonary dry powder inhalation system may be a good alternative to the parenteral administration from a scientific point of view. But for this kind of alternatives there is a justified request for a full registration dossier showing equivalence in efficacy and safety with the current injection product. The development studies needed to fill such a dossier would take years and the costs would be unacceptably high. Based on these considerations any alternative that goes beyond a solution for injection (or a product that is to be reconstituted to such a solution) would require too much development time at the cost of too many unnecessary loss of life.

The fifth solution may potentially be the most ideal solution, but also for this solution authorities should be willing to limit their requirements regarding the registration dossier.

Finally, any solutions clearly require nothing more than good will and funding. Oxytocin, as we described, offers the most apparent and economical solution, where the money should be spent on rather than for the costs of the circus of politicians and officials, who travel around the globe to discuss “the problem”.

142

samenvattIng, conclusIes, aanbevelIngen,

en mondIaal perspectIef

&

samenvattIngDe algemene doelstelling van dit proefschrift (beschreven in hoofdstuk 1) is om vloeibare farmaceutische formuleringen te ontwikkelen voor op polypeptides gebaseerde geneesmiddelen die stabiel zijn onder tropische omstandigheden.

Hoofdstuk 2 bespreekt de instabiliteit van peptides in waterige oplossingen en mogelijkheden om de stabiliteit van peptides in parenterale formuleringen te verbeteren. Hydrolyse, oxidatie en aggregatie zijn de meest voorkomende degradatieprocessen die in versneld houdbaarheidsonderzoek (verhoogde temperatuur) met peptides worden waargenomen. Hydrolyse kan leiden tot deamidatie, isomerisatie en racemerisatie. Oxdatie is vaak geïnduceerd en gekatalyseerd door licht en sporen metaalionen en kan leiden tot β-eliminatie, disulfide uitwisseling en dimerisatie. Gedegen kennis over de aminozuurresiduen die betrokken zijn bij de degradatieprocessen van peptides kan worden gebruikt om strategieën te ontwikkelen om de stabiliteit van peptides in waterige oplossingen te verbeteren. Er zijn verschillende mogelijkheden om de stabiliteit van peptides in waterige formuleringen te verbeteren. Hiertoe behoren onder andere het bepalen van de optimale pH, verwijdering van zuurstof, bescherming tegen licht en de optimale toepassing van hulpstoffen in de formulering zoals buffers, anti-oxidanten, chelatiemiddelen, metaalionen, organisch oplosmiddelen, en oppervlakte-actieve stoffen.

De focus van dit proefschrift ligt op de stabilisatie van oxytocine (een nonapeptide) in een formulering voor injectie. Met name onderzochten we het stabiliseren van oxytocine door het combineren van verschillende metaalionen en buffers.

In hoofdstuk 3 beschrijven we het effect van monovalente en divalente metaalionen in combinatie met buffers bij pH 4,5 op de stabiliteit van oxytocine in waterige oplossingen. De chloridezouten van monovalente (Na+ en K+) en divalente metaalionen (Ca2+, Mg2+, en Zn2+) in combinatie met citraat- of acetaatbuffer werden onderzocht. Het effect van combinaties van buffer met metaalionen op de stabiliteit van waterige oxytocine-oplossingen na 4 weken bewaren bij 4°C of 55°C is gekwantificeerd met behulp van twee analytische technieken: Reversed-Phase High Performance Liquid Chromatography (RP-HPLC) en High-Performance Size exclussion Chromatography (HP-SEC). Het toevoegen van de monovalente metaalionen, Na+ en K+ aan acetaat- of citraat-gebufferde oplossingen en het toevoegen van de divalente metaalionen Ca2+, Mg2+, en Zn2+ aan acetaat- gebufferde oplossingen leidde niet tot een verbetering van stabiliteit van oxytocine. De stabiliteit in waterige formuleringen werd echter wel sterk verbeterd door in een 5 of 10 mM citraatbuffer in combinatie met ten minste 2 mM CaCl2, MgCl2, or ZnCl2 en was afhankelijk van de concentratie divalente metaalionen. Isothermal titration calorimetry (ITC) bleek een techniek te zijn met voorspellende waarde voor de waargenomen effecten. In formuleringen in citraatbuffer met een sterk verbeterde stabiliteit, werd een sterke interactie gevonden tussen oxytocine en Ca2+, Mg2+, of Zn2+. In formuleringen in acetaat buffer vonden we dit niet. Samenvattend kunnen we zeggen dat de combinatie van divalente metaalionen met citraatbuffer de stabiliteit van oxytocine in waterige oplossingen sterk verbeterde.

Hoofdstuk 4 beschrijft een onderzoek waarin de degradatieproducten van oxytocineformuleringen in citraatbuffer werden geïdentificeerd in de aan- en afwezigheid van divalente metaalionen, na 5 dagen blootstelling aan een temperatuur van 70°C. De degradatieproducten werden geanalyseerd met behulp van vloeistofchromatografie-massa spectrometrie/massa-spectrometrie (LC-MS/MS). In de aanwezigheid van divalente metaalionen werd de vorming van bijna alle degradatieproducten, in het bijzonder van

146

&

SaMenvattInG, concluSIeS, en MondIaal perSpectIef

citraatadduct, tri-en tetrasulfide en dimeren, sterk verminderd. Dit werd vooral veroorzaakt door stabilisatie van de disulfidebrug tussen de aminozuren Cys1 en Cys6. Cysteine is gevoelig voor oxidatie en β-eliminatie, en bij de degradatiereacties van oxytocine waarbij cysteine betrokken was leidde dit tot de vorming van dimeren, tri- en tetrasulfide en thio-ether. We vonden geen substantieel verschil tussen het stabiliserende effect van de divalente metaalionen, Ca2+, Mg2+, en Zn2+. Dit suggereert dat divalentie de meest belangrijke eigenschap van het metaalion is die leidt tot de stabilisatie van het oxytocine-metaal-citraatcluster. Daarom veronderstellen we dat cysteine-gemedieerde intermoleculaire reacties worden geremd door de vorming van een complex tussen de divalente metaalionen, citraat en oxytocine. Citraat heeft twee tegenovergestelde effecten op de stabiliteit van ocytocine. Enerzijds kan het met de N-terminale aminogroep van cysteineresidue reageren waardoor een adduct wordt gevormd. Anderzijds kan citraat in aanwezigheid van divalente metaalionen oxytocine juist stabiliseren.

In Hoofdstuk 5 wordt het effect van divalent metaalionen (Ca2+, Mg2+ en Zn2+) op de stabiliteit van oxytocine in aspartaatbuffer (pH 4,5) beschreven. Met behulp van RP-HPLC en HP-SEC metingen toonden we aan dat na 4 weken bewaren bij 55°C de combinatie van alle drie de geteste divalente metaalionen en aspartaatbuffer, leidde tot een verbeterde stabiliteit van oxytocine. De stabiliserende werking van Zn2+ was veel beter dan van Ca2+ of Mg2+. LC-MS/MS resultaten toonden aan dat de combinatie van aspartaat en Zn2+ vooral de vorming van dimeren tegenging. Uit ITC metingen bleek dat Zn2+ in de aanwezigheid van aspartaatbuffer interacties aangaat met oxytocine terwijl dit met Ca2+ of Mg2+ niet het geval was. Samenvattend kan worden gezegd dat de stabiliteit van oxytocine in de met aspartaat gebufferde oplossing sterk kan worden verbeterd door de aanwezigheid van zinkionen, en dat dit effect correleert met het vermogen tot interactie tussen de metaalionen en het oxytocine. Waarschijnlijk wordt de disulfidebrug gestabiliseerd doordat aspartaat complexering van zinkionen in de ringstructuur van oxytocine mogelijk maakt.

In Hoofdstuk 6 werd de conformatie van oxytocine in aspartaatbuffer in aanwezigheid van Mg2+ of Zn2+, onderzocht met behulp van een aantal 2D-NMR technieken: NOESY, TOCSY, 1H-13C HSQC en 1H-15N HSQC. In het HSQC-spectra van oxytocine dat niet was verrijkt met 13C of 15N konden bijna alle 1H, 13C and 15N resonanties worden toegewezen.

Uit deze spectra bleek dat Zn2+ veranderingen van de chemical shifts van vrijwel alle negen aminozuren van oxytocine veroorzaakt, terwijl Mg2+ slechts geringe veranderingen in chemical shifts in sommige aminozuren induceert. Anderzijds vertoonden de NOESY-spectra nagenoeg dezelfde NOEs in de aan- en afwezigheid van Zn2+. Dit duidt erop aan dat de carboxylaatgroep van aspartaat de positieve lading van de N-terminus van Cys1 neutraliseert, waardoor de interactie met Zn2+ gunstiger wordt. Deze interacties kunnen de bescherming van de disulfide brug in oxytocine verklaren.

In Hoofdstuk 7 bestuderen we de stabilisatie van twee modeleiwitten (lysozyme en insuline) in waterige oplossing met behulp van verschillende extremolieten. Het effect van de extremolieten, betaine hydroxyectoine, trehalose, ectoine en firoine op de stabiliteit van lysozyme werd bepaald met behulp van Nile Red fluorescentiespectroscopie en het bepalen van de biologische activiteit van het enzym. De stabiliteit van insuline werd bepaald met behulp van RP-HPLC en HP-SEC. Het effect van extremolieten op de ontvouwingstemperatuur (Tm) van de eiwitten werd geanalyseerd met behulp van een thermal shift assay voor lysozyme en liquid differential scanning microcalorimetry voor insuline. Tijdens incubatie bij 70°C gedurende 10 minuten werd lysozyme beter gestabiliseerd door fiorine dan door de andere geteste extremolieten. Ook tijdens de bewaren bij 55°C gedurende 4 weken, stabiliseerde

147

&

fiorine het eiwit. Betaine, hydroxyectoine, trehalose en ectoine bleken lysozyme onder deze condities echter te destabiliseren. Deze uitkomsten geven aan dat sommige extremolieten eiwitten kunnen stabiliseren onder bepaalde stress-omstandigheden, terwijl dezelfde eiwitten juist worden gedestabiliseerd onder andere stress-omstandigheden. De verbeterde stabiliteit door firoine kan worden verklaard door de waargenomen verhoogde Tm van lysozyme in de aanwezigheid van firoine. Na 4 weken bewaren bij 40°C bleek insulin te worden gestabiliseerd door trehalose en ectoine, terwijl het eiwit juist gedestabiliseerd werd door betaine, hydroxyectoine en met name door firoine. Verder vonden we dat de Tm van insuline sterk afnam in de aanwezigheid van firoine. De interactie van firoine met lysozyme en ectoine of de interactie van trehalose met insuline zoals bepaald met ITC was verwaarloosbaar. Samenvattend kunnen we zeggen dat firoine een uitstekende stabilisator is voor lysozyme, maar daarintegen een destabilisator is voor insuline, terwijl ectoine het tegenovergestelde gedrag vertoont. Uit onze studie blijkt duidelijk dat er geen extremoliet is die kan fungeren als een universele stabilisator voor eiwitten in waterige oplossing. Om een stabiele eiwitformulering met extremolieten te ontwikkelen, moet daarom rekening worden gehouden met de beoogde bewaarcondities. Voor het screenen van extremolieten op hun vermogen tot stabilisatie van eiwitten kan het nuttig om de Tm van de eiwitoplossing met en zonder extremoliet te bepalen, maar alleen als de extremoliet een substantiële verandering in de Tm veroorzaakt.

conclusIes en aanbevelIngen Dit proefschrift beschrijft onderzoek dat gericht is op het ontwikkelen van een thermostabiele waterige oxytocineformulering. Diverse combinaties van divalente metaalionen met buffers bleken oxytocine in waterig milieuw te kunnen stabiliseren. Verschillende analysemethoden, zoals LC-MS (MS), ITC en NMR, werden toegepast om het mechanisme van stabilisatie op moleculair niveau op te helderen. Het is echter nog lastig om de resultaten hiervan direct te correleren aan het exacte mechanisme dat aan de stabilisatie ten grondslag ligt.

Hoewel het algemene beeld vrij complex was en het mechanisme voor de verschillende gebruikte ionen en buffers anders, is wel duidelijk dat de stabilisatie van de Cys1,6 disulfidebrug van het oxytocinemolecuul van essentieel belang is voor het stabiliseren van het peptide in waterige oplossing. Dit wordt het beste geïllustreerd in het NMR-onderzoek waaruit blijkt dat de carboxylaatgroep van aspartaat de positieve lading van N-terminus van Cys1 neutraliseert, waardoor de interactie met zinkionen beter wordt. Deze interactie verklaart de bescherming van de disulfidebrug tegen intermoleculaire reacties die resulteren in dimerisatie en inactivatie.

Zn2+ veroorzaakt een verandering van de conformatie van oxytocine, hetgeen blijkt uit van de veranderingen in de chemical shifts van bijna alle aminozuren. De precieze aard van deze veranderingen is echter nog niet helemaal duidelijk maar kan mogelijk opgehelderd worden met een dynamische beschrijving van moleculaire dynamische simulaties gestuurd door afstand- en hoek-beperkingen die afgeleid kunnen worden uit NMR spectra.

Het toepassen van moleculair dynamische simulaties kan een interessante benadering zijn om inzicht te krijgen in het stabilisatiemechanisme van de divalente metaalionen en de buffers op moleculair niveau. Bij dit type studies kan worden opgehelderd welke conformatieveranderingen worden veroorzaakt door de verschillende metaalionen en hoe deze conformatieveranderingen worden beïnvloed door de buffers in de oplossing. Wanneer er ook een relatie is tussen de conformatie van oxytocine en de stabiliteit, is het

148

&

SaMenvattInG, concluSIeS, en MondIaal perSpectIef

mogelijk om een optimale combinatie van metaalionen en buffer te bepalen die leidt tot de meest stabiele conformatie van oxytocine. Een andere aanpak die uiteindelijk zou kunnen resulteren in een stabiele oxytocineformulering is de kennis uit dit proefschrift te gebruiken in meer of minder door ‘trial and error’ geleide formuleringsstudie.

Splitsing van de disulfidebrug bleek de belangrijkste degradatieroute voor oxytocine te zijn. Bepaalde carboxylaatbuffers in combinatie met bepaalde divalente metaalionen bleken te comlexeren met oxytocine hetgeen leidde tot stabilisatie van de disulfidebrug. Vooral de effecten van dit soort combinaties kunnen een interessant onderwerp voor verdere studies. Er bestaan veel verschillende carboxylaatbuffers en slecht drie werden onderzocht in het werk dat beschreven is in dit proefschrift. De complexiteit van de rol van de buffer wordt goed geïllustreerd door de dubbele rol die citraat kan spelen (stabilisator en adduc vormer), door de diverse effecten van de verschillende divalente metaalionen in aspartaatbuffer, en door de verschillen in effect van citraat- en acetaatbuffer in combinatie met divalente metaalionen. Het is onbekend of andere carboxylaatbuffers oxytocine beter kunnen stabiliseren en of de stabiliteit verder kan worden verbeterd met andere metaalionen. In dit verband is het interessant om te vermelden dat uit een recente studie binnen onze onderzoeksgroep gebleken is dat malonaatbuffer in combinatie met zinkionen oxytocine beter stabiliseert dan aspartaat- en citraatbuffers in combinatie met zinkionen. Op dit moment is een real-time stabiliteitsstudie met formuleringen die gebaseerd zijn op malonaatbuffer en zinkionen gaande. De resultaten na een bewaartermijn van zes maanden zijn veelbelovend waardoor deze formulering geschikt zou kunnen zijn voor gebruik in tropische ontwikkelingslanden.

mondIaal perspectIefZoals beschreven in de inleiding van dit proefschrift was de drijfveer voor ons onderzoek onvoldoende toegang in de tropische ontwikkelingslanden tot oxytocine, hetgeen leidt tot een jaarlijks sterfte van ongeveer 150.000 moeders. De beschikbaarheid van een geschikte vloeibare oxytocineformulering zou mogelijk hun dood kunnen voorkomen. Bij de start van het onderzoek zijn we ervan uitgegaan dat de belangrijkste reden voor dit probleem het feit is dat geen van huidige oxytocineformuleringen kan worden opgeslagen buiten de “cold-chain” en dat er met een warmte-stabiele formulering veel sterfgevallen zouden kunnen worden voorkomen. Dit uitgangspunt en de resultaten van ons onderzoek leidt tot twee belangrijke vragen:1. Is er in dit onderzoek een warmte-stabiele vloeibare oxytocine ormulering ontwikkeld?2. Wat is de beste manier om maternale sterfte als gevolg van een gebrek aan oxytocine in

de tropen, bijvoorbeeld sub-Sahara Afrika, te verminderen? Helaas zijn we in onze studie niet in staat geweest om een vloeibare oxytocineformulering

te vinden, die voldoet aan de standaardeis van veel farmacopees, zoals stabiliteit van ten minste één jaar bij 40°C of twee jaar bij 30°C. De formuleringen gebaseerd op de combinatie van zinkionen en malonaatbuffer zijn echter wel veelbelovend. Maar zolang real-time stabiliteitsgegevens over een periode van ten minste één of twee jaar niet beschikbaar zijn, kunnen er nog geen definitieve conclusies worden getrokken.

Vanuit technologisch oogpunt is er een aantal strategieën waarmee de onmiddellijke behoefte aan een warmtebesteding oxytocineformulering op adequate wijze opgelost zou kunnen worden:

149

&

1. Op plaatsen waar geen koelkast beschikbaar is, is het opslaan van de huidige conventionele oxytocineformuleringen bij omgevingsomstandigheden mogelijk, op voorwaarde dat ze elke zes maanden worden vervangen.

2. De introductie van koelkasten op zonne-energie op die plaatsen waar koelkasten op dit moment niet voorhanden zijn.

3. Een warmtestabiele gevriesdroogde oxytocineformulering (te reconstitueren vlak voor gebruik) is door ons ontwikkeld in samenwerking met MSD en kan direct worden toegepast.

4. De ontwikkeling van een stabiele gesproeidroogde oxytocineformulering voor pulmonale toediening.

5. De introductie van de eerder genoemde formulering gebaseerd op een combinatie van zinkionen en malonaatbuffer is waarschijnlijk de beste oplossing. Een definitieve conclusie over de geschiktheid van deze formulering kan pas worden getrokken na afronding van een real-time stabiliteitsstudie van twee jaar. De resultaten van dit onderzoek zullen in november 2012 beschikbaar komen.Als we deze vijf mogelijke oplossingen voor het probleem evalueren, kunnen we

concluderen dat voor de eerste twee alleen voldoende financiële middelen beschikbaar moeten komen om de oxytocinevoorraden elke half jaar te vervangen of om koelkasten op zonne-energie aan te schaffen.

Een snelle introductie van een te reconstitueren gevriesdroogd poeder op de markt, de derde oplossing, is niet alleen afhankelijk van de bereidheid van de industrie om een stijging van proceskosten en materiaalkosten te accepteren. Veel belangrijker is de bereidheid van de verantwoordelijke autoriteiten om een dergelijk product op hun markt te toe te laten, ter vervanging van het huidige conventionele product, zonder dat daar een volledig registratiedossier voor aangelegd hoeft te worden. Als namelijk een dergelijk dossier moet worden voorbereid zullen de kosten enorm stijgen en zeker te hoog worden voor de betrokken landen. In dit verband moeten we ons realiseren dat de voorgestelde vervanging vanuit wetenschappelijk oogpunt niet tot relevante veiligheidsrisico’s leidt, aangezien alleen de huidige parenterale formulering wordt vervangen door een andere parenterale formulering.

De vierde oplossing bestaat uit een alternatieve doseringsvorm. Vanuit wetenschappelijk oogpunt kan een droog poeder voor pulmonale toediening een goed alternatief zijn voor de parenterale toediening. Voor dit soort alternatieven is echter een volledig registratiedossier nodig waaruit de gelijkwaardigheid in werkzaamheid en veiligheid ten opzichte van het huidige injecteerbare product blijkt. De ontwikkelingsstudies die nodig zijn voor een dergelijk dossier kunnen jaren duren en de kosten zullen onaanvaardbaar hoog zijn. Op basis van deze overwegingen zal voor elk alternatief voor een injectie (of voor een reconstitueerbaar product) te veel ontwikkelstijd nodig zijn. Dit zal helaas ten koste gaan van te veel mensenlevens.

De vijfde oplossing is mogelijk de meest ideale, maar ook hiervoor zullen autoriteiten bereid moeten zijn om hun eisen ten aanzien van het registratiedossier te beperken.

Ten slotte, om het grote probleem op te lossen is niet meer nodig dan goede wil en financiering. De strategiën zoals we in het proefschrift hebben beschreven bieden de meest voor de hand liggende oplossing. Hier zou geld aan moeten worden besteed, in plaats van aan het circus van politici en ambtenaren die over de hele wereld reizen om “het probleem” te bespreken.

150

acknowledgements

&

Completing this thesis would not have been possible without the aid and support of countless people over the past four years. I must first address the sincerest gratitude to Professor Henderik W. Frijlink, my promotor, who allowed me to experience the life of a PhD student. His involvement and crucial contribution has exceptionally inspired and enriched my growth as a scientist. I am grateful in every possible way and hope to keep up our collaboration in the future. To Dr. Wouter L. J. Hinrichs, my co-promotor, I am deeply indebted and thankful for being patience and relentless effort throughout my research and during the writing of this thesis, as well as providing unflinching encouragement and support in various ways. I would also like to thank Dr. Herman J. Woerdenbag and Dr. Gerad J. Bolhuis, who initiated the first contact that led to my PhD studentship in Groningen. As a Dutch Top Institute Pharma fellow, I also thank Louise Lammers and John den Engelsman, for their valuable suggestions.

I was extraordinary fortunate in having Professor Geny M.M. Groothuis, Professor Arnold J.M. Driessen, and Professor Wim Jiskoot as members of the reading committee, for their constructive comments that significantly improved the content of this thesis.

The completion of this thesis was made attainable with support and contribution from colleagues and friends, whom provided valuable discussion, asisstance with instruments, molecular modelling, and production process. On these, I shall address sincere thanks to Jean-Piere Amorij, Andrea Hawe, Robert Poole, Bazak Kukrer, Hjalmar Permentier, Alexej Kedrov, Angela Casini, Frans Mulder, Ruud Scheek, Alia Oktaviani, Peter van der Moelen, Jamshed Anwar, Hans de Waard, Vinay Saluja, Jan Fisher, Anko Eissen, Marinella Visser, Andy-Mark Thunnissen, Ali Rohman, Eni Ratnaningsih, Faizah Fulyani, Ryanto Boediono, M. Khalid, Alexej Kedrov, Annie van Dam, Ghea Schuurman, Syarif Riyadi, Caroline Visser, Hans van Doorne, Wangsa Tirta Ismaya, Wouter Tonis, and Jan Ettema.

Collective and individual acknowledgment are also owed to my colleagues at the Department of Pharmaceutical Technology and Biopharmacy, University of Groningen, whose present refreshed, helpful and memorable at any time. Many thanks address to Paul, Dotie, Gieta, Doreenda, Peter, Stieneke, Lida, Fesia, Parinda, TT, Taufan, Senthil, Wouter T, Anne, Niels, Floris, Marcell, Elham, and Maarten. For Anne de Boer, to be my company on the trip to the hospital and cheering me up when I had the bike accident. Also thanks to the talented photographer BT, who provided nice picture for the thesis cover, which I am proud of. And to Milica, my “roomy”, it is really at my pleasure to have her as the paranymph.

I also benefited by outstanding works from graduate and undergraduate students, whose works contributed to this thesis: Dewi, Jenny, Imma, Stashek, Ilja, Marriel, Ilvy, Esther, Erwin, and Marieke. My special thanks to Phillips, for his “Innovative” idea.

To Henk, Kathy, Rika, Joke, Heleen, Anneke, Annete, Tim, Lisanne, Yvonne, Rijn and Sonja, I am thankful for their kind help with the administration works.

I would also acknowledge the Rector and the Dean of the Faculty of Pharmacy of the University of Surabaya for granting the academic leave permit during my PhD study in Groningen, The Netherlands.

Furthermore, I would also like to thank members of the Indonesian Students Association in Groningen which has given me the opportunity to involve as well as contribute in developing the organization. It is my wish that our splendid relationship would remain growing.

Sincere thanks for my cousin, FS Widoyono and his family: Indah, Pandu, Pandji who helped me a lot with my daily routine. My companion in arms: Mariana, Kenzie, Ono, Neng, Insanu, Uyung, Adit, Wisnu, Adhi, Yota, Puri, Aramel, Muiz, Robby, Astri, Rahma, Lia,

154

&

acknowledGeMentS

Mutia, Shinta, Kadek, Wiwin, Arvie, and my beloved best friends: Fanny, Poppy, Awalia, Yayok, Desti, Yuli, Aini, Klara, Tara, Sita, Ismail, Reza, Pandji, Ratna, and many others, whom have been becoming part of my life during my PhD studentship. The joys and the sorrows we share certainly help me to learn how to become a better me.

My parents deserve my highest gratitude, for their inseparable support and prayers. Spiritual bond and affection between us will never end at whatever cost, even when this life ends, someday. Dadang, Heru, and Ichwan, many thanks for being such a supportive and caring siblings. Also for the in-laws with their thoughtful support.

Words fail me to express my appreciation to my little family: my beloved husband, Iskandar Zulkarnain, for his support, caring and gently love. My beloved daughter, Viny, my ray of sunshine in the cloudiest day, thank you for being always on my side. We are one and hand by hand together reaching our goals. I feel blessed that, despite my limited time as a Doctoral student, I am still allowed to spend enough time to accompany you finishing your Middle Years Program at International School of Groningen.

Lastly, I would like to thank everyone who was important to the flourishing realization of this thesis, as well as expressing my request for forgiveness that I would not be able to mention personally one by one.

155

innovative strategies for stabilization formulations · rijksuniversiteit groningen innovative...

Documents