Innovative Strategies for Stabilization of Therapeutic - TI Pharma
Innovative Strategies for Stabilization of Therapeutic - TI Pharma
Innovative Strategies for Stabilization of Therapeutic - TI Pharma
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
InnovatIve strategIes <strong>for</strong> stabIlIzatIon<br />
<strong>of</strong> therapeutIc peptIdes In aqueous<br />
<strong>for</strong>mulatIons<br />
Christina Avanti
Paranimfen: Milica Stankovic<br />
Leviny Zachreina<br />
The research presented in this thesis was per<strong>for</strong>med within the<br />
framework <strong>of</strong> project D6-202 <strong>of</strong> the Dutch Top Institute <strong>Pharma</strong>.<br />
Printing <strong>of</strong> this thesis was supported by generous<br />
contributions from:<br />
University <strong>of</strong> Groningen<br />
Faculty <strong>of</strong> Mathematics and Natural Sciences <strong>of</strong> the<br />
University <strong>of</strong> Groningen<br />
ISBN: 978-94-6182-122-5<br />
© Copyright 2012 C. Avanti<br />
All rights reserved. No part <strong>of</strong> this thesis may be reproduced,<br />
stored in a retrieval system or transmitted in any <strong>for</strong>m or<br />
by any means, mechanically, by photocopying, recording or<br />
otherwise, without the written permission <strong>of</strong> the author.<br />
Cover design: Bao Tung Pham and Christina Avanti<br />
Layout & printing: Off Page, Amsterdam
RIJKSUNIVERSITEIT GRONINGEN<br />
InnovatIve strategIes <strong>for</strong> stabIlIzatIon<br />
<strong>of</strong> therapeutIc peptIdes In aqueous<br />
<strong>for</strong>mulatIons<br />
Proefschrift<br />
ter verkrijging van het doctoraat in de<br />
Wiskunde en Natuurwetenschappen<br />
aan de Rijksuniversiteit Groningen<br />
op gezag van de<br />
Rector Magnificus, dr. E. Sterken,<br />
in het openbaar te verdedigen op<br />
maandag 2 juli 2012<br />
om 16.15 uur<br />
door<br />
Christina Avanti<br />
geboren op 3 april 1968<br />
Kota Baru, Indonesië
Promotor: Pr<strong>of</strong>. dr. H.W. Frijlink<br />
Copromotor: Dr. W.L.J. Hinrichs<br />
Beoordelingscommissie: Pr<strong>of</strong>. dr. G.M.M. Groothuis<br />
Pr<strong>of</strong>. dr. A.J.M. Driessen<br />
Pr<strong>of</strong>. dr. W. Jiskoot
contents<br />
PART ONE INTRODUC<strong>TI</strong>ON<br />
Chapter 1 General Introduction 9<br />
Chapter 2 Current <strong>Strategies</strong> <strong>for</strong> <strong>Stabilization</strong> <strong>of</strong> <strong>Therapeutic</strong> Peptides in<br />
Aqueous Formulations 13<br />
PART TWO THE USE OF DIVALENT METAL IONS AND CITRATE BUFFERS<br />
TO STABILIZE OXYTOCIN IN AQUEOUS SOLU<strong>TI</strong>ON<br />
Chapter 3 A New Strategy to Stabilize Oxytocin in Aqueous Solutions :<br />
I. The Effects <strong>of</strong> Divalent Metal Ions and Citrate Buffer 35<br />
Chapter 4 A New Strategy To Stabilize Oxytocin in Aqueous Solutions:<br />
II. Suppression <strong>of</strong> Cysteine-Mediated Intermolecular Reactions<br />
by a Combination <strong>of</strong> Divalent Metal Ions and Citrate 51<br />
PART THREE THE USE OF DIVALENT METAL IONS AND ASPARTATE<br />
BUFFER TO STABILIZE OXYTOCIN IN AQUEOUS SOLU<strong>TI</strong>ON<br />
Chapter 5 Insight into the Stability <strong>of</strong> the Zinc-Aspartate-Oxytocin<br />
Complex 79<br />
Chapter 6 Aspartate buffer and divalent metal ions affect the oxytocin<br />
con<strong>for</strong>mation in aqueous solution and protect it from<br />
degradation 95<br />
PART FOUR THE USE OF EXTREMOLYTES TO STABILIZE PROTEIN IN<br />
AQUEOUS SOLU<strong>TI</strong>ON<br />
Chapter 7 Extremolytes: Are There Universal Stabilizers <strong>for</strong> Proteins in<br />
Aqueous Solution? 121<br />
Summary, Concluding Remarks, and Global Perspective 137<br />
Samenvatting, Conclusies, Aanbevelingen, en Mondiaal Perspectief 145<br />
Acknowledgments 153
<strong>for</strong> all the women in existence…<br />
<strong>for</strong> a better chance to live the life with the new born …
Every year 166 000 women die <strong>of</strong> bleeding after child birth,<br />
and more than 50% <strong>of</strong> these deaths occur in sub-Saharan Africa<br />
(Clyburn et. al., 2007)
general IntroductIon<br />
1
1<br />
10<br />
general IntroductIon<br />
Although successful developments in the field <strong>of</strong> peptide synthesis increased the availability<br />
<strong>of</strong> peptide drugs [1], a significant part <strong>of</strong> the world’s population is still facing serious<br />
problems associated with insufficient access to several essential drugs which are peptide in<br />
nature. The major cause <strong>for</strong> this problem is the lack <strong>of</strong> stable <strong>for</strong>mulations that withstand<br />
transport, storage and distribution, particularly in rural and remote areas <strong>of</strong> developing<br />
(tropical) countries that lack a cold chain.<br />
An important active pharmaceutical peptide is oxytocin, a nonapeptide which is the<br />
drug <strong>of</strong> first choice to treat bleeding after childbirth or post-partum hemorrhage (PPH) [2].<br />
Although the availability <strong>of</strong> this drug has greatly declined maternal mortality rates in<br />
the developed world, PPH remains a leading cause <strong>of</strong> maternal mortality elsewhere [3].<br />
Current commercial <strong>for</strong>mulations <strong>of</strong> oxytocin are insufficiently heat-stable to withstand<br />
tropical conditions [4,5]. Cleavage <strong>of</strong> the disulfide bridge was found to be the major<br />
degradation pathway [6,7].<br />
The aim <strong>of</strong> the present thesis was to develop heat-stable <strong>for</strong>mulations <strong>for</strong> polypeptide<br />
drugs, in particular oxytocin and investigate the mechanism <strong>of</strong> stabilization based on<br />
degradation products <strong>for</strong>med during thermal stress be<strong>for</strong>e and after <strong>for</strong>mulation.<br />
This thesis is divided in four parts. In part 1 (Chapter 2) a general introduction is given<br />
on the stability <strong>of</strong> peptide drugs in aqueous solution and current stabilization technologies.<br />
Part 2 (Chapter 3 and 4) describes the use <strong>of</strong> a combination <strong>of</strong> divalent metal ions and citrate<br />
buffer to improve stability <strong>of</strong> oxytocin in aqueous solution, whereas part 3 (Chapter 5 and 6)<br />
is about the effect <strong>of</strong> zinc ions to improve stability <strong>of</strong> oxytocin in aspartate buffer. Finally,<br />
part 4 (Chapter 7) describes an attempt to find the universal stabilizer <strong>for</strong> polypeptides by<br />
investigating various extremolytes using lysozyme and insulin as model peptides.<br />
In Chapter 2, we review the most common degradation pathways <strong>of</strong> peptides, such<br />
as hydrolysis, deamidation, isomerization, racemization, oxidation, disulfide exchange,<br />
dimerization, and further aggregation that cause loss <strong>of</strong> potency <strong>of</strong> pharmaceutical peptides.<br />
In this chapter we also review different strategies to overcome these instabilities, such as pH<br />
optimization, the use <strong>of</strong> buffers, antioxidant and other additives.<br />
In Chapter 3, we describe an investigation on the effect <strong>of</strong> monovalent (Na + and K + ) and<br />
divalent (Ca 2+ , Mg 2+ , and Zn 2+ ) metal ions in combination with citrate and acetate buffers<br />
at pH <strong>of</strong> 4.5 on the stability <strong>of</strong> oxytocin in aqueous solution. The effect <strong>of</strong> combinations <strong>of</strong><br />
buffers and metal ions on the stability <strong>of</strong> aqueous oxytocin solutions was determined by<br />
reversed-phase high per<strong>for</strong>mance liquid chromatography (RP-HPLC) and size exclusion<br />
chromatography (HP-SEC) after 4 weeks <strong>of</strong> storage at either 4°C or 55°C. We also measured<br />
the interaction between oxytocin and Ca 2+ , Mg 2+ , or Zn 2+ in citrate buffer in comparison<br />
with acetate buffer by using isothermal titration calorimetry (ITC).<br />
In Chapter 4, we identified various degradation products <strong>of</strong> oxytocin in citrate-buffered<br />
solution after thermal stress at a temperature <strong>of</strong> 70 °C <strong>for</strong> 5 days and the differences in degradation<br />
pattern in the presence and absence <strong>of</strong> divalent metal ions. Degradation products <strong>of</strong> oxytocin in<br />
the citrate buffer <strong>for</strong>mulation with and without divalent metal ions were analyzed using liquid<br />
chromatography−mass spectrometry/mass spectrometry (LC−MS/MS).<br />
In Chapter 5, we investigated the effects <strong>of</strong> various metal ions (Ca 2+ , Mg 2+ and Zn 2+ ) on<br />
the stability <strong>of</strong> oxytocin in aspartate buffer pH 4.5 and determined their interaction with<br />
the peptide in aqueous solution. The effect <strong>of</strong> combinations <strong>of</strong> various metal ions on the<br />
stability <strong>of</strong> oxytocin in aspartate buffer solutions was determined by RP-HPLC and HP-SEC
after 4 weeks <strong>of</strong> storage at either 4°C or 55°C. We also investigated which degradation<br />
products <strong>of</strong> oxytocin were <strong>for</strong>med in the aspartate buffer <strong>for</strong>mulation with and without<br />
divalent metal ions using LC−MS/MS and determined the interaction between oxytocin<br />
and Ca 2+ , Mg 2+ , or Zn 2+ in aspartate buffer by using ITC.<br />
In Chapter 6, we further explored the mechanism <strong>of</strong> stabilization <strong>of</strong> oxytocin by the<br />
combination <strong>of</strong> Zn 2+ and aspartate buffer. There<strong>for</strong>e, we investigated the con<strong>for</strong>mation <strong>of</strong><br />
oxytocin in aspartate buffer in the presence <strong>of</strong> Zn 2+ in comparison with Mg 2+ using 2D<br />
NMR spectroscopy, i.e. NOESY, TOCSY, 1 H- 13 C HSQC and 1 H- 15 N HSQC with neither 13 C<br />
nor 15 N enrichment.<br />
In Chapter 7, we describe a study on the effects <strong>of</strong> extremolytes on the stabilization <strong>of</strong> two<br />
model proteins (the larger peptides) lysozyme and insulin in aqueous solutions. The effects <strong>of</strong><br />
different extremolytes (betaine, hydroxyectoine, trehalose, ectoine, and firoin) on the stability<br />
<strong>of</strong> lysozyme were determined by Nile red Fluorescence Spectroscopy and a bioactivity assay.<br />
Insulin stability was determined by RP-HPLC and HP-SEC. The effects <strong>of</strong> extremolytes on the<br />
unfolding temperature <strong>of</strong> the proteins were analyzed using a thermal shift assay <strong>for</strong> lysozyme<br />
and liquid differential scanning microcalorimetry <strong>for</strong> insulin. The interaction between<br />
extremolytes and protein was studied by isothermal titration calorimetry (ITC).<br />
references<br />
1. B.A. Moss, Peptide Synthesis, in: S. Frokjaer,<br />
L. Hovgaard (Eds.), <strong>Pharma</strong>ceutical<br />
Formulation Development <strong>of</strong> Peptides<br />
and Proteins, CRC Press, United States <strong>of</strong><br />
America, (2000) 1-12.<br />
2. R.A. Dyer, D. van Dyk, A. Dresner, The use <strong>of</strong><br />
uterotonic drugs during caesarean section,<br />
Int. J. Obstet. Anesth. 19 (2010) 313-19.<br />
3. P. Clyburn, S. Morris, J. Hall, Anaesthesia<br />
and safe motherhood, Anaesthesia 62 (2007)<br />
21-5.<br />
4. H.V. Hogerzeil, G.J.A. Walker, M.J. De Goeje,<br />
Stability <strong>of</strong> injectable ocytocics in tropical<br />
climates, World Health Organization,<br />
Geneva WHO/DAP/93.6. (1993).<br />
5. A.N.J.A.D. Groot, T.B. Vree, H.V. Hogerzeil,<br />
G.J.A. Walker, WHO Action Programme on<br />
Essential Drugs: Stability <strong>of</strong> Oral Oxytocics<br />
in Tropical Climates : Results <strong>of</strong> Simulation<br />
Studies on Oral Ergometrine, Oral<br />
Methylergometrine, Buccal Oxytocin and<br />
Buccal Desamino-Oxytocin, World Health<br />
Organization, Geneva, (1994).<br />
6. A. Hawe, R. Poole, S. Romeijn, P. Kasper,<br />
R. van der Heijden, W. Jiskoot, Towards<br />
heat-stable oxytocin <strong>for</strong>mulations: analysis<br />
<strong>of</strong> degradation kinetics and identification<br />
<strong>of</strong> degradation products, Pharm. Res. 26<br />
(2009) 1679-88.<br />
7. C. Avanti, H.P. Permentier, A.V. Dam, R.<br />
Poole, W. Jiskoot, H.W. Frijlink, W.L.J.<br />
Hinrichs, A new strategy to stabilize oxytocin<br />
in aqueous solutions: II. Suppression <strong>of</strong><br />
cysteine-mediated intermolecular reactions<br />
by a combination <strong>of</strong> divalent metal ions and<br />
citrate, Mol. <strong>Pharma</strong>ceutics 9 (2012) 554-62.<br />
General IntroductIon<br />
1<br />
11
Christina Avanti 1 , Wim Jiskoot 2 , Wouter L. J. Hinrichs 1 , and Henderik W. Frijlink 1<br />
1 Department <strong>of</strong> <strong>Pharma</strong>ceutical Technology and Biopharmacy,<br />
University <strong>of</strong> Groningen, Groningen, The Netherlands<br />
2 Division <strong>of</strong> Drug Delivery Technology,<br />
Leiden/Amsterdam Center <strong>for</strong> Drug Research, Leiden University, Leiden, The Netherlands
current strategIes <strong>for</strong> stabIlIzatIon<br />
<strong>of</strong> therapeutIc peptIdes In aqueous<br />
<strong>for</strong>mulatIons<br />
2
2<br />
abstract<br />
Parenteral administration is one <strong>of</strong> the most used routes to obtain systemic delivery <strong>of</strong><br />
peptide drugs. Although most <strong>of</strong> therapeutic peptides are <strong>for</strong>mulated in the dry powder<br />
<strong>for</strong> reconstitution, from economic point <strong>of</strong> view aqueous liquid <strong>for</strong>mulations are preferred.<br />
However, therapeutic peptides are <strong>of</strong>ten unstable in the aqueous <strong>for</strong>mulation <strong>of</strong> the<br />
injection. This review will focus on the <strong>for</strong>mulation strategies that can be applied to stabilize<br />
therapeutic peptides in aqueous solution. We have organized this review as follows: first the<br />
main peptide stability problems in solution are described followed by a discussion on the<br />
known strategies to reduce peptide degradation, including recent research on improving<br />
peptide stability in liquid <strong>for</strong>mulations.
1. IntroductIon<br />
There has been a rapid expansion in the use <strong>of</strong> peptides as potential drugs since the successful<br />
chemical synthesis <strong>of</strong> oxytocin by duVigneaud in 1953 [1]. The following decades witnessed<br />
the discovery <strong>of</strong> a large number <strong>of</strong> peptides as active pharmaceutical ingredients, and this<br />
process is likely to continue in the future. Currently over a hundred peptide drug candidates<br />
are in development <strong>for</strong> a wide variety <strong>of</strong> diseases such as several <strong>for</strong>ms <strong>of</strong> cancer and metabolic<br />
diseases [2]. Examples <strong>of</strong> several peptides in the Dutch market are described in Table 1.<br />
Peptides differ from proteins in that they are smaller and typically lack a defined tertiary<br />
structure, but a clear distinction is difficult to make and several arbitrary definitions exist.<br />
For instance, Malavolta [3] defined peptides as molecules containing fewer than 40 amino<br />
acid residues, while proteins contain 50 residues or more, with in between a category<br />
called polypeptides (40-49 residues). In another source, amino acids joined together in<br />
chains <strong>of</strong> 50 amino acids or less are defined as peptides, 50-100 amino acids are defined<br />
as polypeptides, and over 100 amino acids are defined as proteins [4]. Although also other<br />
definitions <strong>of</strong> the term peptide have been proposed, in this review we will focus on peptides<br />
containing less than 50 amino acid residues. Within this definition peptides can range in<br />
size from dimers containing two (modified) amino acid residues, such as enalapril and<br />
lisinopril [5], through small peptides, such as oxytocin [6] and octreotide [7] containing<br />
fewer than 10 residues, to relatively large polypeptides, such as calcitonin (32 residues) [8].<br />
A number <strong>of</strong> hormones, enzymes, antitumor agents, antibiotics and neurotransmitters<br />
are peptides. Peptides regulate many physiological processes, acting at some sites as<br />
endocrine or paracrine signals and at other sites as neurotransmitters or growth factors.<br />
Nowadays, peptides are used as therapeutic agents in diverse disease areas such as<br />
neurological, endocrinological and hematological disorders [9].<br />
<strong>Therapeutic</strong> peptides pose a number <strong>of</strong> challenges <strong>for</strong> pharmaceutical scientists regarding<br />
their <strong>for</strong>mulation and delivery. The sensitivity <strong>of</strong> many peptides to enzymatic breakdown<br />
(e.g. in the gastro-intestinal tract) [10] and their poor ability to pass absorbing membranes<br />
typically results in a poor bioavailability following non-parenteral administration [11].<br />
Moreover, the lack <strong>of</strong> physical and chemical stability may lead to significant degradation<br />
during processing and storage <strong>of</strong> the (aqueous) <strong>for</strong>mulations. Enalapril and lisinopril do<br />
not have the typical delivery issues and oral delivery is well possible. There<strong>for</strong>e we do not<br />
discuss those modified dipeptides in this review.<br />
The lack <strong>of</strong> oral efficacy <strong>of</strong> peptides has prompted the examination <strong>of</strong> other noninvasive<br />
routes <strong>for</strong> peptide drug administration [4,12]. These routes include buccal [13-15],<br />
rectal [16,17], vaginal [18,19], percutaneous [20], ocular [21,22], transdermal [23,24],<br />
nasal [25-27] and the pulmonary route [28,29]. However, many <strong>of</strong> these routes <strong>of</strong><br />
administration are still under investigation and they may be insufficiently efficient,<br />
especially when a rapid effect is desired. There<strong>for</strong>e, <strong>for</strong> delivery <strong>of</strong> therapeutic peptides<br />
<strong>for</strong>mulations the invasive parenteral routes are still <strong>of</strong>ten preferred [30]. Intravenous<br />
administration is the most efficient way to deliver peptide drugs directly into the systemic<br />
circulation, since no absorption process is involved. Intramuscular routes can also be used,<br />
however, the absorption is slower than after intravenous administration. Both intravenous<br />
and intramuscular routes do not easily allow self-administration and patient experiences<br />
pain after injection. The subcutaneous route can be employed when a more sustained action<br />
is acceptable or required and this route is more suitable <strong>for</strong> self-administration.<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
15
2<br />
16<br />
Table 1 Several therapeutic peptide-based products in the Dutch market in 2010-2011<br />
Generic<br />
names<br />
Trade<br />
names Principal<br />
Dosage<br />
<strong>for</strong>m<br />
Self-life and<br />
storage cond.<br />
pH<br />
(adj. agent)<br />
Thyrotropin Releasing Hormones<br />
Protirelin<br />
Antibiotic peptides<br />
Relefact<br />
TRH®<br />
Sanovi-Aventis Liquid inj. 3 yrs at 15-25°C 6.5 (Phosphate)<br />
Daptomycin Cubicin® Novartis<br />
Powder<br />
<strong>for</strong> inj.<br />
3 yrs at 2-8°C 4-5 (NaOH)<br />
Teicoplanin Targocid® San<strong>of</strong>i-Aventis<br />
Powder<br />
<strong>for</strong> inj.<br />
4 yrs at 2-8°C 3.8 to 6.5<br />
Platelet aggregates inhibitors<br />
Eptifibatide Integrilin® GlaxoSmithKline<br />
Liquid <strong>for</strong><br />
infusion<br />
3 yrs at 2-8°C 5.35 (citrate)<br />
Somatostatin analog<br />
Octreotide Sandostatin® Novartis Liquid inj. 3 yrs at 2-8°C 4.2 (lact/carb.)<br />
Vasopressins and analog<br />
Desmopressin DDAVP® Ferring Liquid inj. 4 yrs at 2-8°C 4-5<br />
Octostim® Ferring Liquid inj. 4 yrs at 2-8°C 4 (HCl)<br />
Minrin® Ferring Liquid inj. 4 yrs at 2-8°C 4 (HCl)<br />
Felypressin<br />
Oxytocins<br />
Citanest 3%<br />
Octapressin®<br />
Densply Liquid inj. 3 yrs at 15-25°C 3.2<br />
Oxytocin Syntocinon® Defiante Liquid inj. 4 yrs at 2-8°C 4 (acetate)<br />
carbetocin<br />
Oxytocin antagonist<br />
Pabal® Ferring Liquid inj. 2 yrs 2-8°C 3.8 (acetate)<br />
Atosiban Tractocile ®® Ferring Liquid inj. 4yrs at 2-8°C 4.5 (HCl)<br />
Gonadotropin Releasing Hormone (GNRH)/Luteinizing Releasing Hormone (LHRH) agonists<br />
Goserelin Zoladex® Div Liquid inj. 3 yrs < 25°C<br />
Gonadorelin<br />
Relefact<br />
LH-RH®<br />
San<strong>of</strong>i-Aventis Liquid inj. 15-25°C<br />
Powder<br />
Triptorelin<br />
Decapeptyl<br />
-CR®<br />
Ipsen<br />
and solvent<br />
<strong>for</strong> solution<br />
<strong>for</strong> inj.<br />
3 yrs at 2-8°C<br />
nafarelin Synarel® Pfizer<br />
Liquid<br />
nasal spray<br />
Powder<br />
2 yrs < 25*C 5-7 (acetate)<br />
Leuprolide Eligard ® San<strong>of</strong>i-Aventis<br />
and solvent<br />
<strong>for</strong> solution<br />
<strong>for</strong> inj.<br />
2 yrs at 2-8°C<br />
Cetrorelix Cetrotide® Serono<br />
Powder<br />
<strong>for</strong> inj.<br />
3 yrs at 15-25°C<br />
Source: http://www.geneesmiddelenrepertorium.nl/repertorium<br />
continued next page
Table 1 continued Several therapeutic peptide-based products in the Dutch market in 2010-2011<br />
Generic<br />
names<br />
Trade<br />
names Principal<br />
Dosage<br />
<strong>for</strong>m<br />
Self-life and<br />
storage cond.<br />
pH<br />
(adj. agent)<br />
Non-Steroidal Anti-Inflammatory Drugs<br />
Ziconotide<br />
Calcitonins<br />
Prialt® Elan Liquid inj. 3 yrs at 2-8°C 4-5 (HCl/NaOH)<br />
Salmon<br />
Calcitonin<br />
Calcitonin<br />
Sandoz®<br />
Novartis Liquid inj. 5 yrs at 2-8°C 3.3-3.7<br />
Human Parathyroid Hormone [hPTH (1–34)<br />
Teriparatide Forsteo® Eli Lily Liquid inj. 2 yrs at 2-8°C 4 (acetate)<br />
Fusion inhibitor <strong>of</strong> Human Immunodeficiency Virus (HIV) with Cluster Difference 4 (CD4) cells<br />
Powder<br />
Enfuvirtide Fuzeon® Roche<br />
and solvent<br />
<strong>for</strong> solution<br />
<strong>for</strong> inj.<br />
4 yrs at 2-8°C 9-9.5 (carbonate)<br />
Adrenocorticotropin Hormone (ACTH) and derivatives<br />
Corticorelin<br />
CRH -<br />
Ferring®<br />
Ferring<br />
Powder<br />
<strong>for</strong> inj.<br />
3 yrs < 25°C<br />
Source: http://www.geneesmiddelenrepertorium.nl/repertorium<br />
Because <strong>of</strong> their instability, most peptide drugs have to be stored and transported at<br />
low temperatures. This has a big impact on the access to pharmaceutical active peptides,<br />
particularly in rural and tropical areas that lack a cold chain [31,32]. There<strong>for</strong>e, an urgent<br />
need exists <strong>for</strong> new strategies to tackle the problems associated with peptide instability,<br />
especially <strong>for</strong> aqueous injectable <strong>for</strong>mulations which are preferred over freeze-dried<br />
<strong>for</strong>mulations. Although freeze-dried <strong>for</strong>mulations seem an ideal strategy to maintain the<br />
peptide´s structural integrity [9], un<strong>for</strong>tunately, this strategy is economically not ideal.<br />
Particularly <strong>for</strong> developing countries, the lyophilization process may be too expensive.<br />
Production costs are high and the products that have to be stored and transported have a<br />
volume and mass up to twice as large as that <strong>of</strong> a liquid <strong>for</strong>mulation, since the storage and<br />
transport includes both the vials containing the lyophilized powder and sterile water <strong>for</strong><br />
reconstitution. There<strong>for</strong>e, liquid <strong>for</strong>mulations are preferred.<br />
Stability <strong>of</strong> peptide drugs in liquid <strong>for</strong>mulations is the most important factor to be<br />
considered in the design and development <strong>of</strong> liquid parenteral peptide <strong>for</strong>mulations. Table 1<br />
displays the shelf-life and storage conditions <strong>of</strong> the various marketed peptides, showing that<br />
the majority <strong>of</strong> the <strong>for</strong>mulations is to be stored under refrigerated conditions. This is a clear<br />
indication <strong>for</strong> the poor stability <strong>of</strong> many products. Loss <strong>of</strong> potency <strong>of</strong> a peptide commonly<br />
finds its origin in physical degradation processes such as adsorption [33], aggregation and<br />
chemical degradation pathways (e.g. hydrolysis, oxidation, deamidation [34]. This process<br />
is not only specific <strong>for</strong> small peptide or protein but <strong>for</strong> peptides in general. The chemical<br />
instability <strong>of</strong> peptides poses a problem in their development as active pharmaceutical<br />
ingredients. There<strong>for</strong>e, a better understanding <strong>of</strong> the underlying mechanisms <strong>of</strong> instability<br />
<strong>of</strong> a certain peptide is essential to design rational strategies that can be investigated during<br />
the pharmaceutical development process in order to optimize the stability <strong>of</strong> the peptide in<br />
the final <strong>for</strong>mulation [9].<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
17
2<br />
18<br />
2. peptIde InstabIlIty and the possIble<br />
causes <strong>of</strong> degradatIon<br />
During <strong>for</strong>mulation, processing, storage and transportation <strong>of</strong> a peptide drug product,<br />
the peptide is exposed to conditions that can have significant effects on its chemical and<br />
physical integrity and many peptides have to be transported and stored under a “cold chain”<br />
regime [35]. There are several degradation pathways peptides may undergo when they are<br />
<strong>for</strong>mulated as an aqueous solution. The major pathways <strong>of</strong> peptide degradation are broadly<br />
divided into: chemical and physical instability.<br />
Chemical instability can be defined as any process involving modification <strong>of</strong> the<br />
peptide by covalent bond <strong>for</strong>mation or covalent bond cleavage, generating new chemical<br />
entities [36]. Chemical instability pathways include hydrolysis, oxidation, racemization,<br />
isomerization, deamidation, disulfide exchange and β-elimination [37]. Physical instability<br />
refers to any changes to the higher order structure (dimerization and further aggregation),<br />
i.e. changes in noncovalent bonds, not necessarily involving covalent bond modifications.<br />
This physical instability can result in denaturation which may lead to adsorption to surfaces,<br />
aggregation and precipitation [9]. In Table 2, reported degradation pathways <strong>of</strong> various<br />
peptides are presented.<br />
2.1 Hydrolytic pathways<br />
Hydrolysis is one <strong>of</strong> the potential degradation pathways <strong>of</strong> peptides. In aqueous<br />
solution peptide bonds can undergo chemical reactions, usually through an attack <strong>of</strong><br />
an electronegative atom (nucleophile) on the carbonyl carbon, resulting in cleavage <strong>of</strong><br />
peptide bonds. In an alkaline aqueous environment, hydroxyl ions are better nucleophiles<br />
than polar molecules such as water. Under acidic conditions, the carbonyl group becomes<br />
protonated, which leads to a much easier nucleophile attack [58].<br />
2.1.1 Acid/base catalyzed hydrolysis<br />
The stability <strong>of</strong> peptides in aqueous solution strongly depends on pH. It was reported that<br />
at pH range <strong>of</strong> 1-3 gonadorelin and triptorelin mostly undergo acid-catalysed hydrolysis<br />
via deamidation <strong>of</strong> the C-terminal amide to <strong>for</strong>m free acid deamidated products. In the<br />
pH range 5−6, the peptide backbones <strong>of</strong> gonadorelin and triptorelin are hydrolyzed at the<br />
N-terminal side <strong>of</strong> the serine residue probably involving nucleophilic addition <strong>of</strong> the serine<br />
hydroxyl group to the neighboring amide bond <strong>for</strong>ming a cyclic intermediate resulting<br />
in fragmentation [47,59]. At pH values over 7, base- catalyzed epimerization is the main<br />
pathway <strong>of</strong> degradation. Serine is most likely involved in base-catalyzed epimerization<br />
through a carbanion intermediate. Its ability to <strong>for</strong>m a relatively stable six membered<br />
intermediate with a hydrogen bridge explains the relative high rate <strong>of</strong> racemization <strong>of</strong> the<br />
L-serine residue compared to other amino acid residues. Parallel to the epimerization,<br />
base-catalysed hydrolysis <strong>of</strong> gonadorelin and triptorelin occurs. Epimerization <strong>of</strong> serine is<br />
considered the most important degradation reaction <strong>for</strong> hydroxyl-catalyzed degradation <strong>of</strong><br />
gonadorelin and triptorelin [47,48,60].<br />
Acid/base catalyzed hydrolysis has also been observed in the degradation <strong>of</strong> somatostatin<br />
analog octastatin in aqueous <strong>for</strong>mulations. This hydrolysis is also affected by the buffer<br />
species. Jang et. al. [43] reported that the degradation rate <strong>of</strong> octastatin was higher in<br />
phosphate-buffered solution than in glutamate buffers. Increasing buffer concentrations<br />
resulted in a greater degradation <strong>of</strong> octastatin in phosphate-containing buffer, presumably
Table 2. Reported degradation pathways <strong>of</strong> different peptides<br />
Peptide<br />
Number <strong>of</strong><br />
amino acid<br />
residues Degradation pathways Reference<br />
Thyrotropin Releasing Hormones 3 Hydrolysis [38]<br />
Ceftazidime 5 Hydrolysis [39,40]<br />
Eptifibatide 6<br />
Hydrolysis, isomerization, deamidation,<br />
oxidation, and dimerization<br />
[41]<br />
Octreotide 8 Hydrolysis [42,43]<br />
Oxytocin 9<br />
Hydrolysis, deamidation, oxidation,<br />
beta-elimination, and dimerization<br />
[31,44]<br />
Desmopressin 9<br />
Beta-elimination, deamidation, disulfide<br />
exchange, racemization, and oxidation<br />
[45]<br />
Leuprolide 10<br />
Hydrolysis, isomerization, oxidation,<br />
and aggregation<br />
[46]<br />
Goserelin 10 Acid-base catalyzed hydrolysis [47]<br />
Gonadorelin 10 Acid-base catalyzed hydrolysis [47,48]<br />
Triptorelin 10 Acid-base catalyzed hydrolysis [47,48]<br />
Somatostatin and analogues 14 Hydrolysis, [42,43]<br />
Calcitonin 32<br />
Deamidation, oxidation, and<br />
aggregation/fibrilation<br />
[8,49,50]<br />
Human Brain Natriuretic Peptide<br />
[hBNP (1–32)]<br />
32<br />
Aggregation,<br />
deamidation, and oxidation<br />
[51]<br />
Human Parathyroid Hormone<br />
[hPTH (1–34)<br />
34 Aggregation, deamidation, and oxidation [51]<br />
Enfuvirtide 36 Deamidation and aggregation [52,53]<br />
Adrenocorticotropin Hormone<br />
(ACTH)<br />
39<br />
Deamidation <strong>of</strong> Asn residues, and<br />
racemization<br />
[54]<br />
Corticotropin Heleasing factor 41 Oxidation [55]<br />
Amyloid-β (Aβ) Peptides 36-43<br />
Metal-catalyzed oxidation,<br />
dimerization and aggregation<br />
[56,57]<br />
because <strong>of</strong> catalytic effects <strong>of</strong> phosphate ions. In contrast, in the glutamate-buffered<br />
solution, increasing buffer concentrations caused greater stabilization, presumably due to<br />
a stabilizing effect <strong>of</strong> glutamic acid by both ionic and hydrophobic interactions between<br />
octastatin and glutamate ions.<br />
2.1.2 Deamidation <strong>of</strong> Asn and Gln residues<br />
Deamidation is regarded as the most common chemical degradation pathway <strong>for</strong> peptides and<br />
proteins. Peptides containing asparagine (Asn) and glutamine (Gln) residues are known to<br />
undergo spontaneous deamidation to <strong>for</strong>m aspartic acid (Asp) and glutamic acid (Glu) residues<br />
under physiological conditions. Under acidic conditions (pH below 3), deamidation <strong>of</strong> Asn<br />
residues is reported to proceed through direct hydrolysis <strong>of</strong> the Asn side chain amide to <strong>for</strong>m<br />
Asp. Similarly, Gln residues are converted to Glu by acid catalyzed direct hydrolysis [36].<br />
At alkaline and neutral conditions, the deamidation <strong>of</strong> Asn residues predominantly<br />
occurs through a cyclic imide intermediate <strong>for</strong>med by an intermolecular reaction in which<br />
the carbonyl carbon in the side chain <strong>of</strong> Asn residue is attacked by the nitrogen in the amino<br />
acid residue following Asn. There<strong>for</strong>e, the rate <strong>of</strong> deamidation via this pathway depends on<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
19
2<br />
20<br />
the nature <strong>of</strong> the amino acid residue on the carboxyl side <strong>of</strong> the Asn residue [54,61]. At<br />
the same condition, deamidation <strong>of</strong> Gln residues is less common than <strong>for</strong> Asn, because<br />
cyclization <strong>of</strong> Asn residues leads to a five-membered ring. With Gln, the corresponding<br />
intermediate is a six-membered ring, which <strong>for</strong>mation is thermodynamically less favorable<br />
than the smaller, five-membered, ring [36].<br />
The high rate <strong>of</strong> deamidation is caused by the high degree <strong>of</strong> peptide chain flexibility.<br />
The rate <strong>of</strong> deamidation also depends on the amino acid sequence in the peptides [62].<br />
The amino acid residues following asparagine (Asn) such as glycine (Gly), alanine (Ala),<br />
serine (Ser), and aspartic acid (Asp) may seriously increase the reaction rate in the<br />
degradation <strong>of</strong> therapeutic peptides [63]. Catak et al. [64] reported that deamidation<br />
in aqueous solution can occur either through direct hydrolysis or through succinimidemediation.<br />
These two reactions are competitive even in the absence <strong>of</strong> acid or base<br />
catalysis (Figure 1).<br />
Adrenocorticotropic hormone (ACTH), a 39-amino acid polypeptide with a single Asn<br />
residue, was shown to be degraded via deamidation <strong>of</strong> Asn residues [54,65]. The deamidation<br />
<strong>of</strong> Asn or Gln residues was also observed in the degradation <strong>of</strong> salmon calcitonin [49].<br />
The nonapeptide oxytocin is another example <strong>of</strong> a peptide that can undergo deamidation<br />
and involves the hydrolysis <strong>of</strong> Asn 5 and Gln 4 side chain amides. It was also reported that<br />
Gly-NH 2 <strong>of</strong> oxytocin undergoes deamidation at a pH <strong>of</strong> 2. [31].<br />
2.1.3 Isomerization <strong>of</strong> Asp residues<br />
Deamidation <strong>of</strong> the Asn residue in neutral conditions proceeds through the <strong>for</strong>mation <strong>of</strong><br />
a cyclic succinimide intermediate by the intermolecular attack <strong>of</strong> the backbone nitrogen<br />
atom on the carbonyl group <strong>of</strong> the Asn side chain [66]. Hydrolysis <strong>of</strong> the succinimide<br />
intermediate generates the <strong>for</strong>mation <strong>of</strong> isoaspartic (isoAsp) and aspartic acid (Asp)<br />
Figure 1. Deamidation pathways <strong>of</strong> Asn residue via A. direct hydrolysis and B. succinimide<br />
mediation [64].
esidues. Furthermore, direct dehydration occurs in a trans<strong>for</strong>mation <strong>of</strong> aspartic acid (Asp)<br />
into its iso<strong>for</strong>m isoAsp through the same cyclic succinimide intermediate [67]. Additionally,<br />
L-succinimide may be racemized to D-succinimide to <strong>for</strong>m the D-Asp and D-isoAsp<br />
enantiomers [61,68]. Thus, the mechanisms <strong>for</strong> the deamidation and the isomerization<br />
reactions are similar since they both proceed via an intra-molecular cyclic succinimide<br />
intermediate. The <strong>for</strong>mation <strong>of</strong> succinimide intermediate is the rate-limiting step in both<br />
the deamidation and the isomerization reactions at physiological pH [69].<br />
Under acidic conditions, cleavage at the Asp residue is the most important degradation<br />
pathway <strong>of</strong> recombinant human parathyroid hormone (rhPTH). While at pH values above<br />
5 the Asn deamidation is the major degradation route [70].<br />
2.2 Oxidation pathways<br />
Oxidation is another primary chemical degradation pathway that can occur in a peptide.<br />
Oxidation <strong>of</strong> organic molecules is defined as an increase in oxygen or a decrease in hydrogen<br />
content [71]. Alternatively, oxidation can be defined as a reaction that increases the content<br />
<strong>of</strong> more electronegative atoms in a molecule, in which the electronegative heteroatoms are<br />
generally oxygen or halogens.[72]. Any peptide that contains cysteine (Cys), methionine (Met),<br />
histidine (His), tyrosine (Tyr) and tryptophan (Trp) residues can potentially be damaged due<br />
to their high reactivity with various reactive oxygen species (ROS). Cys and Met are at risk<br />
<strong>for</strong> oxidation because <strong>of</strong> their sulfur atoms, and His, Tyr and Trp because <strong>of</strong> their aromatic<br />
rings [73]. Figure 2 illustrates the oxidation reaction <strong>of</strong> Met and His.<br />
Figure 2. Oxidation reaction <strong>of</strong> A: methionine to methionine sulfoxide in acidic condition<br />
(HA) and B: histidine through an oxometallacycyclic intermediate to <strong>for</strong>m various<br />
degradation products, such as 2-oxomidazoline, asparagine, and aspartate (adapted from<br />
Li, et. al, 1995)<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
21
2<br />
22<br />
Oxidation can be induced by contaminating oxidants and light exposure. Oxidation<br />
can be catalyzed by the presence <strong>of</strong> (traces <strong>of</strong>) transition metal ions during processing and<br />
storage. Oxidation <strong>of</strong> peptides may furthermore be influenced by pH, temperature, and<br />
buffer composition. [63].<br />
2.2.1 Auto-oxidation<br />
A rapid oxidation reaction <strong>of</strong> a peptide may occur when the ROS is able to access the side<br />
chain, particularly <strong>for</strong> peptides with Met residues autooxidation may be a major degradation<br />
pathway [36]. Adrenocorticotropic hormone (ACTH) and human atrial natriuretic peptide<br />
(ANP) are examples <strong>of</strong> peptides that undergo spontaneous Met-oxidation [74].<br />
Cysteine residues may also undergo spontaneous oxidation to <strong>for</strong>m the molecular<br />
byproducts sulfinic acid and cysteic acid in the presence <strong>of</strong> metal ions or nearby thiol<br />
groups. Cysteine oxidation involves a nucleophilic attack <strong>of</strong> thiolate ions on disulfide bonds,<br />
generating new disulfide bonds and other thiolate ions. The newly <strong>for</strong>med thiolate ions can<br />
subsequently react with another disulfide bond to <strong>for</strong>m cysteine [75].<br />
2.2.2 Metal ion-catalyzed oxidation<br />
Metal-catalyzed oxidation occurs when a redox active metal ion binds to the peptide. The<br />
amino acid residues that are most susceptible to metal ion catalyzed oxidations are His, Arg,<br />
Lys, Pro, Met, and Cys. Of these amino acids, His and Cys are sensitive to oxidative damage,<br />
as the ROS generated at the metal center does not have to diffuse very far be<strong>for</strong>e reacting<br />
with the peptide [36,76]. Several metals, such as iron and copper have a strong catalytic<br />
power to generate highly reactive hydroxyl radicals if reacted with hydrogen peroxide<br />
(Fenton reaction). The hydroxyl radicals may react with His residue to <strong>for</strong>m 2-oxo-His[77].<br />
Metal ion-catalysis has been observed in the oxidation <strong>of</strong> amyloid-β (Aβ) peptides which<br />
mediates the neurotoxicity <strong>of</strong> Alzheimer’s disease. The amino acid residue involved in this<br />
pathway is His [56].<br />
2.2.3 Light-induced oxidation<br />
In 2007, Kerwin and Remmel [78] summarized light-induced degradation in<br />
biopharmaceuticals. These reactions may occur at several points from production to<br />
delivery <strong>of</strong> the products. Light-induced oxidation is initiated when a compound absorbs<br />
a certain wavelength <strong>of</strong> light, which provides energy to raise the molecule to an excited<br />
state. The excited molecule can then transfer that energy to molecular oxygen, converting<br />
it to reactive singlet oxygen atoms. This is how tryptophan, histidine, and tyrosine can<br />
be modified under light in the presence <strong>of</strong> oxygen [37,79]. Tyrosine photo-oxidation can<br />
produce mono-, di-, tri-, and tetrahydroxyl tyrosine as byproducts [80]. Aggregation is<br />
observed in some peptides due to cross-linking between oxidized tyrosine residues [81].<br />
2.3 β-elimination reactions<br />
Disulfide-bridge peptides might undergo destruction to <strong>for</strong>m free thiol groups through<br />
β-elimination. This reaction is commonly observed when peptides are incubated at<br />
higher pH values or elevated temperature [82,83]. At neutral and alkaline pH levels<br />
peptides can undergo a disulfide exchange reaction which is catalyzed by thiol groups<br />
[84]. When cystine-containing proteins are heated at 100°C, even at a neutral pH, they<br />
undergo destruction to <strong>for</strong>m free thiols [85]. Salmon calcitonin (sCT) degrades via<br />
β-elimination at a disulfide bridge between the cysteine residues at positions 1 and 7. The
insertion <strong>of</strong> an extra sulfur to <strong>for</strong>m a trisulfide bridge has also been reported as a result<br />
<strong>of</strong> a β-elimination reaction [8].<br />
2.4 Disulfide exchange reactions<br />
Disulfide exchange reactions contribute to the <strong>for</strong>mation <strong>of</strong> dimers and larger aggregates.<br />
These reactions may occur at the cysteine–cysteine link. In a study on the degradation <strong>of</strong><br />
sCT, dimeric products were found to be <strong>for</strong>med via disulfide exchange reactions, however,<br />
the disulfide-linked dimers may undergo further disulfide exchange reactions that will<br />
eventually regenerate sCT monomers [8]. Disulfide interchange at aqueous acid conditions<br />
proceeds through sulfenium ions, which arise from the acid hydrolysis <strong>of</strong> the disulfide bond<br />
[86]. Several studies on the disulfide exchange reaction and the importance <strong>of</strong> disulfidebridges<br />
<strong>for</strong> the stability <strong>of</strong> peptides have been reported [8,82,86,87].<br />
2.5 Dimerization, aggregation, and precipitation<br />
The Cys residues in the oxytocin molecule are able to <strong>for</strong>m disulfide bonds due to thiol<br />
exchange between two oxytocin molecules [31,44]. Thiol exchange occurs at neutral and<br />
alkaline conditions via a nucleophilic attack <strong>of</strong> free thiolate on one <strong>of</strong> the sulfur atoms within<br />
the disulfide bridge. At lower pH values the disulfide exchange progresses via a sulfenium<br />
cation, <strong>for</strong>med following protonation <strong>of</strong> the disulfide bridge [36,86]. In either case, disulfide<br />
exchange can result in dimerization and progressive aggregation [31]. As mentioned above,<br />
peptide molecules are also able to <strong>for</strong>m dimers due to light or metal induced oxidation<br />
<strong>of</strong> tyrosine residues to <strong>for</strong>m dityrosine-linked dimers [31,88,89]. Aggregation may be<br />
induced by several stress condition, such as heating, freezing or agitation. Aggregates can<br />
<strong>for</strong>m either via covalent bonds, such as disulfide bonds, ester, or amide linkage or noncovalent<br />
bonds occurring via hydrophobic interaction or charge-charge complexation. The<br />
relative weakness <strong>of</strong> non-covalent protein bonds can lead to aggregate disruption during<br />
the analytical process [90].<br />
There is no single pathway by which peptides can <strong>for</strong>m aggregates [36]. Both mechanisms<br />
can occur simultaneously leading to the <strong>for</strong>mation <strong>of</strong> either soluble or insoluble aggregates<br />
[91]. Furthermore higher concentrations <strong>of</strong> the peptide in solution will bring aggregates more<br />
easily to a later stage <strong>of</strong> physical instability i.e. precipitation [12]. Calcitonin, β-amyloid peptide,<br />
leuprolide, and deterelix have been reported to <strong>for</strong>m gel-like aggregates. The gel <strong>for</strong>mation is a<br />
function <strong>of</strong> peptide concentration, temperature, time, pH, and agitation. However the chemical<br />
stability and dimerization <strong>of</strong> leuprolide was not affected by gelation [92].<br />
3. strategIes to Improve peptIde stabIlIty<br />
In aqueous <strong>for</strong>mulatIons<br />
Many ef<strong>for</strong>ts have been made to improve the stability <strong>of</strong> peptides in aqueous solution.<br />
Examples are: the use <strong>of</strong> buffers, metal ions or organic solvents, and oxygen removal. However,<br />
an optimal use <strong>of</strong> these strategies requires first <strong>of</strong> all knowledge <strong>of</strong> the peptide´s structure and<br />
thorough understanding <strong>of</strong> the possible degradation pathways <strong>of</strong> the specific peptide.<br />
Examining the amino acid sequence <strong>of</strong> a peptide can provide an insight into how the<br />
molecule may degrade. This will be <strong>of</strong> real use <strong>for</strong> peptides <strong>of</strong> which the amino acid sequence<br />
is generally exposed on the aqueous environment and there<strong>for</strong>e able to undergo chemical<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
23
2<br />
24<br />
degradation. The degradation pathways and involved amino acid sequences as well as the<br />
<strong>for</strong>mulation strategies to inhibit the specific degradation pathway are summarized in Table 3.<br />
3.1 Optimizing hydrolytic stability<br />
3.1.1 pH optimization and buffer species<br />
Stability <strong>of</strong> peptides depends on the pH, there<strong>for</strong>e the common strategy to avoid instability<br />
<strong>of</strong> peptides in aqueous solution is adjusting the pH using buffers. In table 1 the pH values<br />
and the pH adjusting agent or buffer <strong>of</strong> several marketed products are given. Considering<br />
modern <strong>for</strong>mulation development strategies it may be assumed that these values represent<br />
a value which is (close to) the optimum pH stability <strong>of</strong> the different peptides. The acceptable<br />
pH range <strong>for</strong> (slow) intravenous administration is 3-10.5 and 4-9 <strong>for</strong> other parenteral routes<br />
to minimize discom<strong>for</strong>t at the injection site [93,94]. There<strong>for</strong>e, it is important to study the<br />
pH stability <strong>of</strong> a peptide in the range <strong>of</strong> pH 3-10 with different buffers in the early stages<br />
<strong>of</strong> <strong>for</strong>mulation development [95,96]. Hydrolysis <strong>of</strong> the side chain amide on glutamine and<br />
asparagine residues could be reduced by <strong>for</strong>mulating at a pH below neutral. However,<br />
the pH should not be lower than 3 to prevent direct hydrolysis <strong>of</strong> the Asn and Gln side<br />
chains and to minimize hydrolytic fragmentation [36]. Oxytocin was described to have the<br />
highest stability at pH 4.5 [31]. Calcitonin undergoes hydrolysis at basic pH, but no such<br />
degradation is observed at a pH <strong>of</strong> 7 even at room temperature [97]. Other mechanisms<br />
<strong>for</strong> stabilization by buffers have also been reported: In some cases they can act as radical<br />
scavengers [98]. Even more important is the fact that some buffers are able to bind directly<br />
to peptides, thereby increasing their con<strong>for</strong>mational stability [36]. Citric acid has been<br />
reported to bind with oxytocin <strong>for</strong>ming N-cytril oxytocin which increases the stability <strong>of</strong><br />
oxytocin in the presence <strong>of</strong> divalent metal ions [44].<br />
The pH <strong>of</strong> the <strong>for</strong>mulation can also substantially affect deamidation. Deamidation<br />
is known to be sensitive to both the pH and the composition and concentration <strong>of</strong> the<br />
buffer. In general, <strong>for</strong>mulations with a pH in the range from 3 to 5 minimize peptide<br />
deamidation [65,98,105]. Buffer species can also affect deamidation. It has been reported<br />
that the rate <strong>of</strong> deamidation is faster in phosphate and bicarbonate buffers than in acetate<br />
and pyruvate buffers [96]. Isomerization is greatest at low pH values <strong>for</strong> asparagine and<br />
glutamine residues [106].<br />
3.1.2 Ionic strength<br />
The total concentration <strong>of</strong> dissolved electrolytes might affect the rate <strong>of</strong> hydrolysis. The<br />
increase in ionic strength could have a stabilizing or destabilizing effect on a peptide,<br />
depending on the nature <strong>of</strong> the charge–charge interactions within the peptide [107-109].<br />
However, in another study, no significant effect <strong>of</strong> the ionic strength on the rate <strong>of</strong><br />
deamidation or hydrolysis in small peptides was found [110].<br />
3.1.3 Cosolvent<br />
Cosolvents such as low molecular weight polyethylene glycol (PEG) were shown to reduce<br />
the aggregation <strong>of</strong> several peptides [9,12,111]. Peptide deamidation can be modestly<br />
inhibited in an aqueous solution by the addition <strong>of</strong> glycerol [100], or ethanol [112]. Addition<br />
<strong>of</strong> organic solvents decreases the dielectric constant <strong>of</strong> an aqueous solution. Reduction<br />
<strong>of</strong> the solvent´s dielectric strength leads to significantly lower rates <strong>of</strong> isomerization and<br />
deamidation [69]. The use <strong>of</strong> polyols includes polyhydric alcohols and carbohydrates
Table 3. Degradation pathways and possible stabilization strategies <strong>for</strong> peptides including amino acid<br />
residues involved in the degradation.<br />
Degradation pathway <strong>Stabilization</strong> strategy<br />
Chemical Instability<br />
Acid/base catalyzed hydrolysis pH, buffer species, co-solvents<br />
Amino acid<br />
residue(s) involved Ref.<br />
Ser Trp<br />
Asn-Pro<br />
Asn-Tyr<br />
[47,60]<br />
[42]<br />
Deamidation pH 3-5, increased solvent viscosity Asn, Gln<br />
[62,69]<br />
[99,100]<br />
Isomerization pH 6
2<br />
26<br />
adjusting the primary and secondary packaging to protect from light, and the use <strong>of</strong> antioxidants<br />
in the <strong>for</strong>mulation. Watermann and co-workers have made a guideline on the use <strong>of</strong> excipients to<br />
optimize oxidative stability including their recommended concentrations [101].<br />
3.2.1 Buffers<br />
The side chains <strong>of</strong> tryptophan, methionine and cysteine and, to a lesser extent, tyrosine<br />
and histidine are potential oxidation sites. The modification <strong>of</strong> indole group <strong>of</strong> Trp, thiol<br />
group <strong>of</strong> Cys, imidazole group <strong>of</strong> His, and phenolic side chain <strong>of</strong> Tyr by the reactive oxygen<br />
species is most significant at neutral and alkaline condition. There<strong>for</strong>e, a lower pH reduces<br />
oxidation <strong>of</strong> peptides containing these amino acids [114]. In contrast, the thioether group<br />
<strong>of</strong> methionine can be readily oxidized by certain reagents under acidic conditions [58].<br />
3.2.2 Air exclusion<br />
Removal <strong>of</strong> oxygen by bubbling another gas, <strong>for</strong> example nitrogen, through the solution<br />
may be effective to exclude oxygen. For some oxidation-sensitive peptide drugs, processing<br />
and filling steps should be carried out in the presence <strong>of</strong> an inert gas such as nitrogen, argon<br />
or helium. To further minimize air oxidation violent stirring should be avoided [115].<br />
3.2.3 Antioxidants<br />
Antioxidants are commonly used to protect peptides from oxidation during processing<br />
and <strong>for</strong>mulation. Some antioxidants, however, can be problematic in peptide<br />
<strong>for</strong>mulations [37,114]. For example, bisulfite is not a suitable agent because it is a strong<br />
nucleophile, which may interact with peptides. Furthermore, addition <strong>of</strong> antioxidants to<br />
trace metal ions contaminated peptide solutions will not protect the peptide from oxidative<br />
modification. In contrast, it can even accelerate the oxidation process, as demonstrated by<br />
the use <strong>of</strong> ascorbic acid which promotes rather than inhibits oxidation <strong>of</strong> the Met residue <strong>of</strong><br />
small model peptides in the presence <strong>of</strong> metal ions [116,117]. Methionine, a sulfur containing<br />
amino acid, is readily oxidized to methionine sulfoxide by many reactive oxygen species and<br />
oxidizes more easily than the peptide, and there<strong>for</strong>e can act as a sacrificial antioxidant [118].<br />
3.2.4 Chelating agents<br />
Another method to reduce free radical oxidation is to include chelating agents. Chelating<br />
agents are used to inhibit oxidation by complexation <strong>of</strong> trace metal ions. The most commonly<br />
used chelating agents in pharmaceutical <strong>for</strong>mulations are ethylenediaminetetraacetic acid<br />
(EDTA), desferal, diethylenetriaminepentaacetic acid (DTPA), inositol hexaphosphate,<br />
ethylenediamine bis(o-hydroxyphenylacetic acid), tris(hydroxymethyl)aminomethane<br />
(TRIS), citric acid, and tartaric acid. EDTA has been recommended as a chelating agent <strong>for</strong><br />
copper ions, whereas desferal, DTPA, inositol hexaphosphate and ethylenediamine bis(ohydroxy-phenylacetic<br />
acid) were recommended <strong>for</strong> iron ions [113]. It is well known that<br />
chelating agents are generally effective in stabilizing peptides against oxidation. However,<br />
it cannot be assumed that addition <strong>of</strong> a certain chelating agents will be able to interact<br />
with all trace metal ions and completely eliminate oxidation. Under certain circumstances,<br />
chelating agents may even accelerate the oxidation process [37,119].<br />
3.2.5 Polyols<br />
Polyols such as mannitol, trehalose, sucrose, maltose, and raffinose can prevent oxidation<br />
<strong>of</strong> therapeutic peptides. For example, mannitol was shown to inhibit the iron-catalyzed<br />
oxidation <strong>of</strong> Met-containing peptides [102]. Sucrose has been shown to decrease the
ate <strong>of</strong> oxidation <strong>of</strong> parathyroid hormone hPTH (1–34) and brain natriuretic hormones<br />
hBNP (1–32) in liquid <strong>for</strong>mulations [51].<br />
3.3 Protection against disulfide exchange reaction<br />
Formulating octreotide in glycine with pharmaceutically acceptable salts and HCl was<br />
found to be effective in protecting its disulfide-bridge from cleavage, and was also reported<br />
to be better tolerated than the <strong>for</strong>mulation in acetate, lactate or bicarbonate [120]. Divalent<br />
metal ions in combination with specific buffers may protect peptide drugs against disulfide<br />
exchange reaction. Recently, it was shown that Zn 2+ , Ca 2+ and Mg 2+ ions in combination<br />
with dicarboxylic and tricarboxylic acids improved the stability <strong>of</strong> oxytocin [44,104].<br />
The addition <strong>of</strong> at least 2 mM CaCl 2 , MgCl 2 , or ZnCl 2 to 5 and 10 mM citrate-buffered<br />
solutions at pH 4.5 increased oxytocin stability in aqueous <strong>for</strong>mulations and the stability is<br />
further increased with increasing concentration <strong>of</strong> the divalent metals ions up to 50 mM.<br />
The improved stability is due to the reactivity <strong>of</strong> carboxyl ions in citrate buffer which can<br />
attack the N-terminal group <strong>of</strong> Cys to <strong>for</strong>m an adduct and together with divalent metal ions<br />
suppress the disulfide exchange reaction [44,104].<br />
3.4 Inhibition <strong>of</strong> dimerization, aggregation, and precipitation<br />
Dimerization and aggregation may involve the interchange <strong>of</strong> covalent bonds, such as<br />
disulfide bridges or non-covalent <strong>for</strong>ces such as hydrophobic interactions. Both soluble<br />
and insoluble aggregates may occur. A peptide in aqueous solution can be stabilized against<br />
dimerization and aggregation by optimizing the pH and ionic strength <strong>of</strong> the solution.<br />
Furthermore, dimerization can be prevented by preferential exclusion using sugars, amino<br />
acids, and/or polyols, and by using surfactants [36]. The use <strong>of</strong> a combination <strong>of</strong> divalent<br />
metal ions and citrate buffer has been found to inhibit the cysteine-mediated dimerization <strong>of</strong><br />
oxytocin [44]. PEG has been shown to reduce the aggregation <strong>of</strong> several peptides [9,12,111].<br />
The stabilizing effects increased with increasing concentration and increasing molecular<br />
weight [99,100]. Dicarboxylic amino acids such as aspartic and glutamic acid have been<br />
used to reduce aggregation [121] and glycine, arginine and lysine have also been reported to<br />
prevent aggregation [121-123]. Polysorbate 20 and 80 [124,125] can reduce agitation-induced<br />
aggregation <strong>of</strong> peptides. However, surfactants are reported to be less effective in reducing<br />
thermally-induced aggregation. There<strong>for</strong>e, the optimum concentration required to protect a<br />
specific peptide should be evaluated <strong>for</strong> each different type <strong>of</strong> stress that may be encountered.<br />
3.5 Preferential exclusion<br />
Another strategy to minimize peptide degradation is the use <strong>of</strong> extremolytes.<br />
Extremolytes are small organic molecules produced by extremophilic microorganisms,<br />
to protect their biological macromolecules from damage by external stresses [126].<br />
Examples <strong>of</strong> extremolytes are the polyol derivatives ectoin and hydroxyectoin [127],<br />
betain [128], trehalose [129], amino acids (e.g. proline), and the mannose derivative<br />
mannosylglycerate [130]. Mannosylglycerate was reported to have the ability to inhibit<br />
β-amyloid peptide aggregation [131]. Extremolytes are known to stabilize peptides by<br />
<strong>for</strong>ming solute hydrate clusters that are preferentially excluded from the hydrate shell <strong>of</strong><br />
the peptide as a result <strong>of</strong> the repulsive interactions between extremolytes and the backbone<br />
<strong>of</strong> the peptides. Accumulation <strong>of</strong> water near peptide domains assembles the peptide into<br />
a more compact structure with a reduced surface area [132-134]. However, there are no<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
27
2<br />
28<br />
extremolytes that can be used as a universal stabilizer <strong>for</strong> all peptides in aqueous solution.<br />
In addition, it should be realized that extremolytes may also act as destabilizer <strong>for</strong> certain<br />
peptides under specific conditions [135].<br />
Surfactants have also been described as stabilizers <strong>for</strong> peptides. Pluronic®<br />
F68 has<br />
been used to enhance stability <strong>of</strong> peptide drugs ceftazidime in parenteral solutions. In this<br />
approach, the peptides are kept in a microenvironment that shields them from exterior<br />
conditions, limiting the effect <strong>of</strong> moisture and pH [40].<br />
5. conclusIon<br />
In aqueous solutions peptides are <strong>of</strong>ten unstable. Peptides have unique structures that differ<br />
from proteins in that they are smaller and rarely have a tendency to physical degradation e.g.<br />
unfolding. Peptides <strong>of</strong>ten do not possess a higher order structure that can sequester reactive<br />
groups, the side chain <strong>of</strong> nearly all <strong>of</strong> the amino acid residues are fully solvent exposed,<br />
allowing maximal contact with solvents and the degradation rates appear to correlate<br />
with the degree <strong>of</strong> solvent exposure. Based on knowledge <strong>of</strong> the peptide’s structure and an<br />
understanding <strong>of</strong> the predominant degradation pathways, the strategies may be developed<br />
to achieve adequate stability <strong>of</strong> the <strong>for</strong>mulation. The degradation pathways <strong>of</strong> peptides are<br />
mainly dependent on the amino acid sequence. The most prominent degradation pathways<br />
<strong>for</strong> peptides are hydrolysis, oxidation, and dimerization. Formulating peptides in a specific<br />
pH with a specific buffer, avoiding oxygen reactive species, and minimizing solvent exposure<br />
eliminate chemical degradation. Increasing solution viscosity by using sugars or polymers<br />
reduces peptide mobility and further decelerates physical degradation.<br />
references<br />
1. V. du Vigneaud, C. Ressler, S. Trippett, The<br />
sequence <strong>of</strong> amino acids in oxytocin, with a<br />
proposal <strong>for</strong> the structure <strong>of</strong> oxytocin, J. Biol.<br />
Chem. 205 (1953) 949-57.<br />
2. J. Sterling, GEN’s 30th Anniversary: Peptide<br />
<strong>Therapeutic</strong>s, Editor’s Note, Genetic<br />
Engineering and Biotechnology News 31<br />
(2011).<br />
3. L. Malavolta, F.R. Cabral, Peptides:<br />
Important tools <strong>for</strong> the treatment <strong>of</strong> central<br />
nervous system disorders, Neuropeptides 45<br />
(2011) 309-16.<br />
4. V.H.L. Lee (Ed.), Peptide and Protein Drug<br />
Delivery, Marcel Dekker Inc., New York,<br />
1991.<br />
5. S.Y. Lin, S.L. Wang, Advances in simultaneous<br />
DSC-F<strong>TI</strong>R microspectroscopy <strong>for</strong> rapid<br />
solid-state chemical stability studies: Some<br />
dipeptide drugs as examples, Adv. Drug<br />
Deliv. Rev. 64 (2012) 461-78.<br />
6. T. Wieland, M. Bodanszky (Eds.), The Worlds<br />
<strong>of</strong> Peptides: A Brief History <strong>of</strong> Peptide<br />
Chemistry, Springer-Verlag, Berlin, 1991.<br />
7. J. Pless, The history <strong>of</strong> somatostatin analogs,<br />
J. Endocrinol. Invest. 28 (2005) 1-4.<br />
8. V. Windisch, F. DeLuccia, L. Duhau,<br />
F. Herman, J.J. Mencel, S.Y. Tang, M.<br />
Vuilhorgne, Degradation pathways <strong>of</strong><br />
salmon calcitonin in aqueous solution, J.<br />
Pharm. Sci. 86 (1997) 359-64.<br />
9. S. Frokjaer, L. Hovgaard (Eds.),<br />
<strong>Pharma</strong>ceutical Formulation Development<br />
<strong>of</strong> Peptides and Proteins, CRC Press, United<br />
States America, 2000.<br />
10. P.A. Zhou XH, Peptide and protein drugs.<br />
<strong>Therapeutic</strong> applications, absorption and<br />
parenteral administration, Int J Pharm 75<br />
(1991) 97-115.<br />
11. A.M. Papini, Peptide Chemistry Revolution,<br />
Chemistry Today 29 (2011).<br />
12. M.F. Powell, L.M. Sanders, A. Rogerson, V. Si,<br />
Parenteral peptide <strong>for</strong>mulations: chemical<br />
and physical properties <strong>of</strong> native luteinizing<br />
hormone-releasing hormone (LHRH) and<br />
hydrophobic analogues in aqueous solution,<br />
Pharm. Res. 8 (1991) 1258-63.<br />
13. F. Veuillez, Y.N. Kalia, Y. Jacques, J.<br />
Deshusses, P. Buri, Factors and strategies <strong>for</strong><br />
improving buccal absorption <strong>of</strong> peptides,<br />
Eur. J. Pharm. Biopharm. 51 (2001) 93-109.
14. D.H. Oh, K.H. Chun, S.O. Jeon, J.W. Kang,<br />
S. Lee, Enhanced transbuccal salmon<br />
calcitonin (sCT) delivery: Effect <strong>of</strong> chemical<br />
enhancers and electrical assistance on in<br />
vitro sCT buccal permeation, Eur. J. Pharm.<br />
Biopharm. 79 (2011) 357-63.<br />
15. K.V. Patel, N.D. Patel, H.D. Dodiya, P.K.<br />
Shelat, Buccal Bioadhesive Drug Delivery<br />
System: An Overview, Int. J. Pharm. Biol.<br />
Arch. 2 (2011) 600-9.<br />
16. A. Nasr, A.Y. Shahin, A.M. Elsamman,<br />
M.S. Zakherah, O.M. Shaaban, Rectal<br />
misoprostol versus intravenous oxytocin <strong>for</strong><br />
prevention <strong>of</strong> postpartum hemorrhage, Int.<br />
J. Gynaecol. Obstet. 105 (2009) 244-7.<br />
17. K. Morimoto, H. Akatsuchi, R. Aikawa, M.<br />
Morishita, K. Morisaka, Enhanced rectal<br />
absorption <strong>of</strong> [Asu’.‘]-eel calcitonin in rats<br />
using polyacrylic acid aqueous gel base, J.<br />
Pharm. Sci. 73 (1984) 1366-8.<br />
18. A.P. Sayani, Y.W. Chien, Systemic delivery<br />
<strong>of</strong> peptides and proteins across absorptive<br />
mucosae, Rev. Ther. Drug Carrier Syst. 13<br />
(1996) 85-184.<br />
19. J.L. Richardson, L. Illum, (D) Routes <strong>of</strong><br />
delivery: Case studies: (8) The vaginal route<br />
<strong>of</strong> peptide and protein drug delivery, Adv.<br />
Drug Deliv. Rev. 8 (1992) 341-66.<br />
20. N. Seo, M. Takigawa, The current status<br />
and future direction <strong>of</strong> percutaneous<br />
peptide immunization against melanoma, J.<br />
Dermatol. Sci. 48 (2007) 77-85.<br />
21. S.P. Read, S.M. Cashman, R. Kumar-Singh,<br />
A poly(ethylene) glycolylated peptide<br />
<strong>for</strong> ocular delivery compacts DNA into<br />
nanoparticles <strong>for</strong> gene delivery to postmitotic<br />
tissues in vivo, J. Gene Med. 12<br />
(2010) 86-96.<br />
22. D. Harris, J.H. Liaw, J.R. Robinson, (D)<br />
Routes <strong>of</strong> Delivery: Case Studies (7) Ocular<br />
delivery <strong>of</strong> peptide and protein drugs, Adv.<br />
Drug Deliv. Rev. 8 (1992) 331-9.<br />
23. H.E. Bodde, J.C. Verhoef, M. Ponec,<br />
Transdermal peptide delivery, Biochem.<br />
Soc. Trans. 17 (1989) 943-5.<br />
24. M.R. Prausnitz, A peptide chaperone<br />
<strong>for</strong> transdermal drug delivery, Nature<br />
Biotechnology 24 (2006) 416-7.<br />
25. A.E. Pontiroli, Peptide hormones: Review <strong>of</strong><br />
current and emerging uses by nasal delivery,<br />
Adv Drug Deliv Rev 29 (1998) 81-7.<br />
26. S. Chhajed, S. Sangale, S.D. Barhate,<br />
Advantageous Nasal Drug Delivery System:<br />
A Review, IJPSR 2 (2011) 1322-36.<br />
27. S. Turker, E. Onur, Y. Ozer, Nasal route and<br />
drug delivery systems, Pharm. World Sci. 26<br />
(2004) 137-42.<br />
28. S.A. Shoyele, A. Slowey, Prospects <strong>of</strong><br />
<strong>for</strong>mulating proteins/peptides as aerosols<br />
<strong>for</strong> pulmonary drug delivery, Int. J. Pharm.<br />
314 (2006) 1-8.<br />
29. R.I. Henkin, Inhaled insulinintrapulmonary,<br />
intranasal, and other routes<br />
<strong>of</strong> administration: mechanisms <strong>of</strong> action,<br />
Nutrition 26 (2010) 33-9.<br />
30. X.H. Zhou, A. Li Wan Po, Peptide and<br />
protein drugs: II. Non-parenteral routes <strong>of</strong><br />
delivery, Int. J. Pharm. 75 (1991) 117-30.<br />
31. A. Hawe, R. Poole, S. Romeijn, P. Kasper,<br />
R. van der Heijden, W. Jiskoot, Towards<br />
heat-stable oxytocin <strong>for</strong>mulations: analysis<br />
<strong>of</strong> degradation kinetics and identification<br />
<strong>of</strong> degradation products, Pharm. Res. 26<br />
(2009) 1679-88.<br />
32. International Conference on Harmonisation;<br />
Stability Data Package <strong>for</strong> Registration<br />
Applications in Climatic Zones III and IV;<br />
Stability Testing <strong>of</strong> New Drug Substances<br />
and Products; availability. Notice, Fed.<br />
Regist. 68 (2003) 65717-8.<br />
33. S.T. Anik, J.Y. Hwang, Adsorption <strong>of</strong> D-<br />
Nal(2)6LHRH, a decapeptide, onto glass<br />
and other surface., Int. J. Pharm. 16 (1983)<br />
181-190.<br />
34. V.G. Badelin, O.V. Kulikov, V.S. Vatagin,<br />
Physico-chemical properties <strong>of</strong> peptides and<br />
their solutions, Thermochimica Acta 169<br />
(1990) 81-93.<br />
35. M. Zahn, P.W. Kallberg, G.M. Slappendel,<br />
H.M. Smeenge, A risk-based approach to<br />
establish stability testing conditions <strong>for</strong><br />
tropical countries, J. Pharm. Sci. 95 (2006)<br />
946-65.<br />
36. M.C. Manning, D.K. Chou, B.M. Murphy,<br />
R.W. Payne, D.S. Katayama, Stability <strong>of</strong><br />
protein pharmaceuticals: an update, Pharm.<br />
Res. 27 (2010) 544-75.<br />
37. S. Li, C. Schoneich, R.T. Borchardt, Chemical<br />
instability <strong>of</strong> protein pharmaceuticals:<br />
Mechanisms <strong>of</strong> oxidation and strategies <strong>for</strong><br />
stabilization, Biotechnol. Bioeng. 48 (1995)<br />
490-500.<br />
38. T. Hashimoto, K. Ohki, N. Sakura, Hydrolytic<br />
cleavage <strong>of</strong> pyroglutamyl-peptide bond. I.<br />
The susceptibility <strong>of</strong> pyroglutamyl-peptide<br />
bond to dilute hydrochloric acid, Chem.<br />
Pharm. Bull. (Tokyo) 43 (1995) 2068-74.<br />
39. J.M. Hwang, T.E. Piccinini, C.J. Lammel,<br />
W.K. Hadley, G.F. Brooks, Effect <strong>of</strong> storage<br />
temperature and pH on the stability <strong>of</strong><br />
antimicrobial agents in MIC trays, J. Clin.<br />
Microbiol. 23 (1986) 959-61.<br />
40. C.A.D. Santos, G.B. Ribeiro, M.C. Knirsch,<br />
A.P. Junior, T.C.V. Penna, Influence <strong>of</strong><br />
Pluronic® F68 on ceftazidime biological<br />
activity in parenteral solutions, J. Pharm.<br />
Sci. 100 (2011) 715-20.<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
29
2<br />
30<br />
41. L. Zhao, S.H. Yalkowsky, <strong>Stabilization</strong> <strong>of</strong><br />
eptifibatide by cosolvents, Int. J. Pharm. 218<br />
(2001) 43-56.<br />
42. R. Krishnamoorthy, A.K. Mitra, Kinetics<br />
and mechanism <strong>of</strong> degradation <strong>of</strong> a cyclic<br />
hexapeptide (somatostatin analogue) in<br />
aqueous solution, Pharm. Res. 10 (1992)<br />
1314-20.<br />
43. S.W. Jang, B.H. Woo, J.T. Lee, S.C. Moon,<br />
K.C. Lee, P.P. DeLuca, Stability <strong>of</strong> Octastatin,<br />
a somatostatin analog cyclic octapeptide, in<br />
aqueous solution, Pharm. Dev. Technol. 2<br />
(1997) 409-14.<br />
44. C. Avanti, H.P. Permentier, A.V. Dam, R.<br />
Poole, W. Jiskoot, H.W. Frijlink, W.L.J.<br />
Hinrichs, A new strategy to stabilize oxytocin<br />
in aqueous solutions: II. Suppression <strong>of</strong><br />
cysteine-mediated intermolecular reactions<br />
by a combination <strong>of</strong> divalent metal ions and<br />
citrate, Mol. <strong>Pharma</strong>ceutics 9 (2012) 554-62.<br />
45. S.L. Law, K.J. Huang, V.H. Chou, Stability<br />
<strong>of</strong> desmopressin loaded in liposomes, J.<br />
Liposome Res. 13 (2003) 269-77.<br />
46. S.C. Hall, M.M. Tan, J.J. Leonard,<br />
C.L. Stevenson, Characterization and<br />
comparison <strong>of</strong> leuprolide degradation<br />
pr<strong>of</strong>iles in water and dimethyl sulfoxide, J.<br />
Pept. Res. 53 (1999) 432-41.<br />
47. M.A. Hoitink, J.H. Beijnen, M.U.S. Boschma,<br />
A. Bult, E. Hop, J. Nijholt, C. Versluis,<br />
Wiese,G.,Underberg,W.J.M., Identification<br />
<strong>of</strong> the degradation products <strong>of</strong> gonadorelin<br />
and three analogues in aqueous solution,<br />
Anal. Chem. 69 (1997) 4972-8.<br />
48. V.J. Helm, B.W. Muller, Stability <strong>of</strong><br />
gonadorelin and triptorelin in aqueous<br />
solution, Pharm. Res. 7 (1990) 1253-6.<br />
49. K.C. Lee, Y.J. Lee, H.M. Song, C.J. Chun, P.P.<br />
DeLuca, Degradation <strong>of</strong> synthetic salmon<br />
calcitonin in aqueous solution, Pharm. Res.<br />
9 (1992) 1521-3.<br />
50. S. Seyferth, G. Lee, Structural studies <strong>of</strong><br />
EDTA-induced fibrillation <strong>of</strong> salmon<br />
calcitonin, Pharm. Res. 20 (2003) 73-80.<br />
51. M. Kamberi, Y.J. Kim, B. Jun, C.M. Riley,<br />
The effects <strong>of</strong> sucrose on stability <strong>of</strong> human<br />
brain natriuretic peptide [hBNP (1-32)] and<br />
human parathyroid hormone [hPTH (1-<br />
34)], J. Pept. Res. 66 (2005) 348-56.<br />
52. H. Stocker, C. Kl<strong>of</strong>t, N. Plock, A. Breske, G.<br />
Kruse, C. Herzmann, H. Schulbin, P. Kreckel,<br />
C. Weber, F. Goebel, J. Roeling, S. Staszewski,<br />
A. Plettenberg, C. Moecklingh<strong>of</strong>f, K.<br />
Arasteh, M. Kurowski, <strong>Pharma</strong>cokinetics<br />
<strong>of</strong> enfuvirtide in patients treated in typical<br />
routine clinical settings, Antimicrob. Agents<br />
Chemother. 50 (2006) 667-73.<br />
53. R.W. Payne, R. Nayar, R. Tarantino, S. Del<br />
Terzo, J. Moschera, J. Di, D. Heilman, B. Bray,<br />
M.C. Manning, C.S. Henry, Second virial<br />
coefficient determination <strong>of</strong> a therapeutic<br />
peptide by self-interaction chromatography,<br />
Biopolymers 84 (2006) 527-33.<br />
54. N.P. Bhatt, K. Patel, R.T. Borchardt,<br />
Chemical pathways <strong>of</strong> peptide degradation.<br />
I. Deamidation <strong>of</strong> adrenocorticotropic<br />
hormone, Pharm. Res. 7 (1990) 593-9.<br />
55. I.Q. Assil, M.E. Shomali, A.B. Abou-Samra,<br />
An oxidation resistant radioligand <strong>for</strong><br />
corticotropin-releasing factor receptors,<br />
Peptides 22 (2001) 1055-61.<br />
56. K. Inoue, A. Nakagawa, T. Hino, H. Oka,<br />
Screening assay <strong>for</strong> metal-catalyzed<br />
oxidation inhibitors using liquid<br />
chromatography-mass spectrometry with<br />
an N-terminal beta-amyloid peptide, Anal.<br />
Chem. 81 (2009) 1819-25.<br />
57. D.P. Smith, G.D. Ciccotosto, D.J. Tew, M.T.<br />
Fodero-Tavoletti, T. Johanssen, C.L. Masters,<br />
K.J. Barnham, R. Cappai, Concentration<br />
dependent Cu2+ induced aggregation and<br />
dityrosine <strong>for</strong>mation <strong>of</strong> the Alzheimer’s<br />
disease amyloid-beta peptide, Biochemistry<br />
46 (2007) 2881-91.<br />
58. M.C. Manning, K. Patel, R.T. Borchardt,<br />
Stability <strong>of</strong> protein pharmaceuticals, Pharm<br />
Res 6 (1989) 903-18.<br />
59. R.G. Strickley, M. Brandl, K.W. Chan, K.<br />
Straub, L. Gu, High-per<strong>for</strong>mance liquid<br />
chromatographic (HPLC) and HPLCmass<br />
spectrometric (MS) analysis <strong>of</strong> the<br />
degradation <strong>of</strong> the luteinizing hormonereleasing<br />
hormone (LH-RH) antagonist RS-<br />
26306 in aqueous solution., Pharm. Res. 7<br />
(1990) 530-6.<br />
60. M.A. Hoitink, J.H. Beijnen, A. Bult,<br />
O.A.G.J. Van der Houwen, J. Nijholt,<br />
W.J.M. Underberg, Degradation kinetics <strong>of</strong><br />
gonadorelin in aqueous solution, J. Pharm.<br />
Sci. 85 (1996) 1053-59.<br />
61. T. Geiger, S. Clarke, Deamidation,<br />
isomerization, and racemization at<br />
asparaginyl and aspartyl residues in<br />
peptides: Succinimide-linked reactions that<br />
contribute to protein degradation, J. Biol.<br />
Chem. 262 (1987) 785-94.<br />
62. K. Patel, R.T. Borchardt, Chemical pathways<br />
<strong>of</strong> peptide degradation. III. Effect <strong>of</strong> primary<br />
sequence on the pathways <strong>of</strong> deamidation<br />
<strong>of</strong> asparaginyl residues in hexapeptides,<br />
Pharm. Res. 7 (1990) 787-93.<br />
63. R.W. Payne, M.C. Manning, Peptide<br />
<strong>for</strong>mulation: challenges and strategies,<br />
Innov. Pharm. Technol. 28 (2009) 64-8.<br />
64. S. Catak, G. Monard, V. Aviyente, M.F. Ruiz-<br />
Lopez, Deamidation <strong>of</strong> asparagine residues:<br />
direct hydrolysis versus succinimide-
mediated deamidation mechanisms, J Phys<br />
Chem A 113 (2009) 1111-20.<br />
65. K. Patel, R.T. Borchardt, Chemical pathways<br />
<strong>of</strong> peptide degradation. II. Kinetics <strong>of</strong><br />
deamidation <strong>of</strong> an asparaginyl residue in a<br />
model hexapeptide, Pharm. Res. 7 (1990)<br />
703-11.<br />
66. H. Yang, R.A. Zubarev, Mass spectrometric<br />
analysis <strong>of</strong> asparagine deamidation and<br />
aspartate isomerization in polypeptides,<br />
Electrophoresis 31 (2010) 1764-72.<br />
67. N.E. Robinson, Protein deamidation, Procl.<br />
Natl. Acad. Sci. 99 (2002) 5283-8.<br />
68. N.P. Sargaeva, A.A. Goloborodko,<br />
P.B. O’Connor, E. Moskovets, M.V.<br />
Gorshkov, Sequence-specific predictive<br />
chromatography to assist mass spectrometric<br />
analysis <strong>of</strong> asparagine deamidation and<br />
aspartate isomerization in peptides,<br />
Electrophoresis 32 (2011) 1962-9.<br />
69. A.A. Wakankar, R.T. Borchardt, Formulation<br />
considerations <strong>for</strong> proteins susceptible<br />
to asparagine deamidation and aspartate<br />
isomerization, J. Pharm. Sci. 95 (2006) 2321-<br />
36.<br />
70. Y. Nabuchi, E. Fujiwara, H. Kuboniwa, Y.<br />
Asoh, H. Ushio, The stability and degradation<br />
pathway <strong>of</strong> recombinant human parathyroid<br />
hormone: deamidation <strong>of</strong> asparaginyl<br />
residue and peptide bond cleavage at<br />
aspartyl and asparaginyl residues, Pharm.<br />
Res. 14 (1997) 1685-90.<br />
71. T.W.G. Solomons (Ed.), Organic Chemistry,<br />
5th ed., John Wiley and Sons, New York,<br />
1992.<br />
72. [72] J. March (Ed.), Advanced Organic<br />
Chemistry, 4th ed., John Wiley and Sons,<br />
New York, 1992.<br />
73. S. Li, C. Schoneich, R.T. Borchardt, Chemical<br />
pathways <strong>of</strong> peptide degradation. VIII.<br />
Oxidation <strong>of</strong> methionine in small model<br />
peptides by prooxidant/transition metal ion<br />
systems: influence <strong>of</strong> selective scavengers <strong>for</strong><br />
reactive oxygen intermediates, Pharm. Res.<br />
3 (1995) 348-55.<br />
74. Z. Guan, N.A. Yates, R. Bakhtiar, Detection<br />
and characterization <strong>of</strong> methionine<br />
oxidation in peptides by collisioninduced<br />
dissociation and electron capture<br />
dissociation, J. Am. Soc. Mass Spectrom. 14<br />
(2003) 605-13.<br />
75. J. Patel, R. Kothari, R. Tunga, N.M. Ritter,<br />
B.S. Tunga, Stability considerations <strong>for</strong><br />
biopharmaceuticals. Part 1 Overview <strong>of</strong><br />
protein and peptide degradation pathways<br />
, Bioprocess International (2011) 20-31.<br />
76. M. Khossravi, R.T. Borchardt, Chemical<br />
pathways <strong>of</strong> peptide degradation. X: Effect<br />
<strong>of</strong> metal-catalyzed oxidation on the solution<br />
structure <strong>of</strong> a histidine-containing peptide<br />
fragment <strong>of</strong> human relaxin, Pharm. Res. 17<br />
(2000) 851-8.<br />
77. M. Khossravi, R.T. Borchardt, Chemical<br />
pathways <strong>of</strong> peptide degradation: IX. Metalcatalyzed<br />
oxidation <strong>of</strong> histidine in model<br />
peptides, Pharm. Res. 15 (1998) 1096-1102.<br />
78. B.A. Kerwin, R.L. Remmele Jr, Protect<br />
from light: photodegradation and protein<br />
biologics, J. Pharm. Sci. 96 (2007) 1468-<br />
1479.<br />
79. A.J. Grosvenor, J.D.D. Morton J.M.,<br />
Pr<strong>of</strong>iling <strong>of</strong> residue-level photooxidative<br />
damage in peptides<br />
, Amino Acids 39 (2010) 285-96.<br />
80. J. Zhang, D.S. Kalonia, The effect <strong>of</strong><br />
neighboring amino acid residues and<br />
solution environment on the oxidative<br />
stability <strong>of</strong> tyrosine in small peptides, AAPS<br />
PharmSciTech 8 (2007) 176-83.<br />
81. N. Rathore, R.S. Rajan, Current perspectives<br />
on stability <strong>of</strong> protein drug products during<br />
<strong>for</strong>mulation, fill and finish operations,<br />
Biotechnol. Prog. 24 (2008) 504-14.<br />
82. A.K. Galande, J.O. Trent, A.F. Spatola,<br />
Understanding base-assisted desulfurization<br />
using a variety <strong>of</strong> disulfide-bridged peptides,<br />
Peptide Science 71 (2003) 534-51.<br />
83. S.L. Cohen, C. Price, J. Vlasak, Betaelimination<br />
and peptide bond hydrolysis:<br />
two distinct mechanisms <strong>of</strong> human IgG1<br />
hinge fragmentation upon storage, J. Am.<br />
Chem. Soc. 129 (2007) 6976-7.<br />
84. A.P. Ryle, F. Sanger, Disulphide interchange<br />
reactions, Biochem. J. 60 (1955) 535.<br />
85. D.B. Volkin, A.M. Klibanov, Thermal<br />
Destruction Processes in Proteins Involving<br />
Cystine Residues, J. Biol. Chem. 262 (1987)<br />
2945-50.<br />
86. R.E. Benesch, R. Benesch, The mechanism <strong>of</strong><br />
disulfide interchange in acid solution; role <strong>of</strong><br />
sulfenium ions, J. Am. Chem. Soc. 80 (1958)<br />
1666-9.<br />
87. K. Wakabayashi, H. Nakagawa, A. Tamura,<br />
S. Koshiba, K. Hoshijima, M. Komada, T.<br />
Ishikawa, Intramolecular disulfide bond is a<br />
critical check point determining degradative<br />
fates <strong>of</strong> ATP-binding cassette (ABC)<br />
transporter ABCG2 protein, J. Biol. Chem.<br />
282 (2007) 27841-6.<br />
88. D.A. Malencik, S.R. Anderson, Dityrosine<br />
<strong>for</strong>mation in calmodulin: conditions <strong>for</strong><br />
intermolecular cross-linking, Biochemistry<br />
33 (1994) 13363-72.<br />
89. I.C. Smith, R. Deslauriers, H. Saito, R.<br />
Walter, C. Garrigou-Lagrange, H. McGregor,<br />
D. Sarantakis, Carbon-13 NMR studies <strong>of</strong><br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
31
2<br />
32<br />
peptide hormones and their components,<br />
Ann. N Y Acad. Sci. 222 (1973) 597-627.<br />
90. M. Kamberi, P. Chung, R. DeVas, L. Li, Z.<br />
Li, X.S. Ma, S. Fields, C.M. Riley, Analysis <strong>of</strong><br />
non-covalent aggregation <strong>of</strong> synthetic hPTH<br />
(1-34) by size-exclusion chromatography<br />
and the importance <strong>of</strong> suppression <strong>of</strong><br />
non-specific interactions <strong>for</strong> a precise<br />
quantitation, J. Chromatogr. B. Analyt<br />
Technol. Biomed. Life. Sci. 810 (2004) 151-5.<br />
91. W. Wang, Protein aggregation and its<br />
inhibition in biopharmaceutics, Int. J.<br />
Pharm. 289 (2005) 1-30.<br />
92. M.M. Tan, C.A. Corley, C.L. Stevenson,<br />
Effect <strong>of</strong> gelation on the chemical stability<br />
and con<strong>for</strong>mation <strong>of</strong> leuprolide, Pharm. Res.<br />
15 (1998) 1442-8.<br />
93. G.A. Brazeau, B. Cooper, K.A. Svetic, C.L.<br />
Smith, P. Gupta, Current perspectives on<br />
pain upon injection <strong>of</strong> drugs, J. Pharm. Sci.<br />
87 (1998) 667-77.<br />
94. Extemp.ie, Section 2: Sterile <strong>Pharma</strong>ceuticals<br />
- Parenterals, accessed April, 01 2012.<br />
95. L. Jorgensen, S. Hostrup, E.H. Moeller,<br />
H. Grohganz, Recent trends in stabilising<br />
peptides and proteins in pharmaceutical<br />
<strong>for</strong>mulation - considerations in the choice<br />
<strong>of</strong> excipients, Expert Opin. Drug Deliv. 6<br />
(2009) 1219-30.<br />
96. J.L. Cleland, M.F. Powell, S.J. Shire, The<br />
development <strong>of</strong> stable protein <strong>for</strong>mulations:<br />
a close look at protein aggregation,<br />
deamidation, and oxidation, Crit. Rev. Ther.<br />
Drug Carrier Syst. 10 (1993) 307-77.<br />
97. M. Cholewinski, B. Lückel, H. Horn,<br />
Degradation pathways, analytical<br />
characterization, and <strong>for</strong>mulation strategies<br />
<strong>of</strong> a peptide and a protein: calcitonin and<br />
human growth hormone in comparison,<br />
Pharm. Acta Helv. 71 (1996) 405-19.<br />
98. N.E. Good, G.D. Winget, W. Winter, T.N.<br />
Connolly, S. Izawa, R.M. Singh, Hydrogen<br />
ion buffers <strong>for</strong> biological research,<br />
Biochemistry 5 (1966) 467-77.<br />
99. R. Li, M.J. Hageman, E.M. Topp, Effect <strong>of</strong><br />
viscosity on the deamidation rate <strong>of</strong> a model<br />
Asn-hexapeptide, J. Pept. Res. 59 (2002)<br />
211-20.<br />
100. R. Li, A.J. D’Souza, B.B. Laird, R.L. Schowen,<br />
R.T. Borchardt, E.M. Topp, Effects <strong>of</strong><br />
solution polarity and viscosity on peptide<br />
deamidation. 2000 Nov;56(5):326-334., J<br />
Pept Res 56 (2000) 326.<br />
101. K.C. Waterman, R.C. Adami, K.M. Alsante,<br />
J. Hong, M.S. Landis, F. Lombardo, C.J.<br />
Roberts, <strong>Stabilization</strong> <strong>of</strong> pharmaceuticals to<br />
oxidative degradation, Pharm. Dev. Technol.<br />
7 (2002) 1-32.<br />
102. S. Li, T.W. Patap<strong>of</strong>f, T.H. Nguyen, R.T.<br />
Borchardt, Inhibitory effect <strong>of</strong> sugars and<br />
polyols on the metal-catalyzed oxidation <strong>of</strong><br />
human relaxin, J. Pharm. Sci. 85 (1996) 868-<br />
72.<br />
103. L.P. Stratton, R.M. Kelly, J. Rowe, J.E. Shively,<br />
D.D. Smith, J.F. Carpenter, M.C. Manning,<br />
Controlling deamidation rates in a model<br />
peptide: Effects <strong>of</strong> temperature, peptide<br />
concentration, and additives, J. Pharm. Sci.<br />
90 (2001) 2141-8.<br />
104. C. Avanti, J.P. Amorij, D. Setyaningsih, A.<br />
Hawe, W. Jiskoot, J. Visser, A. Kedrov, A.J.<br />
Driessen, W.L. Hinrichs, H.W. Frijlink, A<br />
new strategy to stabilize oxytocin in aqueous<br />
solutions: I. The effects <strong>of</strong> divalent metal ions<br />
and citrate buffer, AAPS J. 13 (2011) 284-90.<br />
105. W. Wang, S. Martin-Moe, C. Pan, L. Musza,<br />
Y.J. Wang, <strong>Stabilization</strong> <strong>of</strong> a polypeptide in<br />
non-aqueous solvents, Int. J. Pharm. 351<br />
(2008) 1-7.<br />
106. Y.C.J. Wang, M.A. Hanson, Parenteral<br />
<strong>for</strong>mulations <strong>of</strong> proteins and peptides:<br />
stability and stabilizers, J. Parenter. Sci.<br />
Technol. 42 (1987) S2-S26.<br />
107. K. Zheng, C.R. Middaugh, T.J. Siahaan,<br />
Evaluation <strong>of</strong> the physical stability <strong>of</strong> the<br />
EC5 domain <strong>of</strong> E-cadherin: effects <strong>of</strong> pH,<br />
temperature, ionic strength, and disulfide<br />
bonds, J. Pharm. Sci. 98 (2009) 63-73.<br />
108. S.A. Bursakov, C. Carneiro, M.J. Almendra,<br />
R.O. Duarte, J. Caldeira, I. Moura, J.J. Moura,<br />
Enzymatic properties and effect <strong>of</strong> ionic<br />
strength on periplasmic nitrate reductase<br />
(NAP) from Desulfovibrio desulfuricans<br />
ATCC 27774, Biochem. Biophys. Res.<br />
Commun. 239 (1997) 816-22.<br />
109. P. O’Neill, S. Davies, E.M. Fielden, L.<br />
Calabrese, C. Capo, F. Marmocchi, G. Natoli,<br />
G. Rotilio, The effects <strong>of</strong> pH and various salts<br />
upon the activity <strong>of</strong> a series <strong>of</strong> superoxide<br />
dismutases, Biochem. J. 251 (1988) 41-6.<br />
110. R. Tyler-Cross, V. Schirch, Effects <strong>of</strong> amino<br />
acid sequence, buffers, and ionic strength on<br />
the rate and mechanism <strong>of</strong> deamidation <strong>of</strong><br />
asparagine residues in small peptides, J. Biol.<br />
Chem. 266 (1991) 22549-56.<br />
111. K. Neelon, H.J. Schreier, H. Meekins, P.M.<br />
Robinson, M.F. Roberts, Compatible solute<br />
effects on thermostability <strong>of</strong> glutamine<br />
synthetase and aspartate transcarbamoylase<br />
from Methanococcus jannaschii, Biochim.<br />
Biophys. Acta 1753 (2005) 164-73.<br />
112. T.V. Brennan, S. Clarke, Spontaneous<br />
degradation <strong>of</strong> polypeptides at aspartyl and<br />
asparaginyl residues: effects <strong>of</strong> the solvent<br />
dielectric, Protein Sci. 2 (1993) 331-8.
113. D.A. Parkins, U.T. Lashmar, The Formulation<br />
<strong>of</strong> Biopharmaceutical Products, Pharm. Sci.<br />
Technol. Today 3 (2000) 129-37.<br />
114. S. Li, C. Schoneich, R.T. Borchardt, Chemical<br />
instability <strong>of</strong> protein pharmaceuticals:<br />
Mechanisms <strong>of</strong> oxidation and strategies <strong>for</strong><br />
stabilization, Biotechnol. Bioeng. 48 (1995)<br />
490-500.<br />
115. S. Landi, H.R. Held, Effect <strong>of</strong> oxidation on<br />
the stability <strong>of</strong> tuberculin purified protein<br />
derivative (PPD), Dev. Biol. Stand. 58 (1986)<br />
545-52.<br />
116. C. Schoneich, F. Zhao, G.S. Wilson, R.T.<br />
Borchardt, Iron-thiolate-induced oxidation<br />
<strong>of</strong> methionine to methionine sulfoxide in<br />
small model peptides. Catalysis by histidine.,<br />
Biochim. Biophys. Acta 1158 (1993) 307-22.<br />
117. S. Li, C. Schoneich, G.S. Wilson, R.T.<br />
Borchardt, Chemical pathways <strong>of</strong> peptide<br />
degradation. V. Ascorbic acid promotes<br />
rather than inhibits the oxidation <strong>of</strong><br />
methionine to methionine sulfoxide in<br />
small model peptides, Pharm. Res. 10 (1993)<br />
1572-9.<br />
118. R.L. Levine, J. Moskovitz, E.R. Stadtman,<br />
Oxidation <strong>of</strong> methionine in proteins: roles in<br />
antioxidant defense and cellular regulation,<br />
IUBMB Life 50 (2000) 301-7.<br />
119. P.K. Tsai, D.B. Volkin, J.M. Dabora, K.C.<br />
Thompson, M.W. Bruner, J.O. Gress, B.<br />
Matuszewska, M. Keogan, J.V. Bondi, C.R.<br />
Middaugh, Formulation design <strong>of</strong> acidic<br />
fibroblast growth factor, Pharm. Res. 10<br />
(1993) 649-59.<br />
120. B.P. Obiols, G.J. Farres, F.J.C. Rodriguez,<br />
S.P. Fernandez, J.B. Cabado, Stable<br />
pharmaceutical <strong>for</strong>mulation <strong>for</strong> intravenous<br />
or intramuscular administration <strong>of</strong> active<br />
peptide compound (2003) March, 20, 2012.<br />
121. T. Arakawa, S.J. Prestrelski, W.C. Kenney, J.F.<br />
Carpenter, Factors affecting short-term and<br />
long-term stabilities <strong>of</strong> proteins, Adv. Drug<br />
Deliv. Rev. 10 (1993) 1-28.<br />
122. T. Matsuoka, S. Tomita, H. Hamada,<br />
K. Shiraki, Amidated amino acids are<br />
prominent additives <strong>for</strong> preventing heatinduced<br />
aggregation <strong>of</strong> lysozyme, J. Biosci.<br />
Bioeng. 103 (2007) 440-3.<br />
123. R. Quinn, J.D. Andrade, Minimizing the<br />
aggregation <strong>of</strong> neutral insulin solutions, J.<br />
Pharm. Sci. 72 (1983) 1472-3.<br />
124. D.K. Chou, R. Krishnamurthy, T.W.<br />
Randolph, J.F. Carpenter, M.C. Manning,<br />
Effects <strong>of</strong> Tween 20 and Tween 80 on the<br />
stability <strong>of</strong> Albutropin during agitation, J.<br />
Pharm. Sci. 94 (2005) 1368-81.<br />
125. A. Lahlou, B. Blanchet, M. Carvalho, P. P M.,<br />
A. Astier, Mechanically-induced aggregation<br />
<strong>of</strong> the monoclonal antibody cetuximab, Ann.<br />
Pharm. Fr. 67 (2009) 340-52.<br />
126. G. Lentzen, T. Schwarz, Extremolytes:<br />
Natural compounds from extremophiles<br />
<strong>for</strong> versatile applications, Appl. Microbiol.<br />
Biotechnol. 72 (2006) 623-34.<br />
127. I. Yu, Y. Jindo, M. Nagaoka, Microscopic<br />
understanding <strong>of</strong> preferential exclusion <strong>of</strong><br />
compatible solute ectoine: direct interaction<br />
and hydration alteration, J. Phys. Chem. B<br />
111 (2007) 10231-8.<br />
128. S. Knapp, R. Ladenstein, E.A. Galinski,<br />
Extrinsic protein stabilization by the<br />
naturally occurring osmolytes betahydroxyectoine<br />
and betaine, Extremophiles<br />
3 (1999) 191-8.<br />
129. A. Hedoux, J.F. Willart, L. Paccou, Y. Guinet,<br />
F. Affouard, A. Lerbret, M. Descamps,<br />
Thermostabilization mechanism <strong>of</strong> bovine<br />
serum albumin by trehalose, J. Phys. Chem.<br />
B 113 (2009) 6119-26.<br />
130. T.Q. Faria, S. Knapp, R. Ladenstein,<br />
A.L. Macanita, H. Santos, Protein<br />
stabilisation by compatible solutes: effect<br />
<strong>of</strong> mannosylglycerate on unfolding<br />
thermodynamics and activity <strong>of</strong> ribonuclease<br />
A, ChemBioChem 4 (2003) 734-41.<br />
131. J. Ryu, M. Kanapathipillai, G. Lentzen,<br />
C.B. Park, Inhibition <strong>of</strong> β-amyloid peptide<br />
aggregation and neurotoxicity by α-dmannosylglycerate,<br />
a natural extremolyte,<br />
Peptides 29 (2008) 578-84.<br />
132. E.A. Galinski, M. Stein, B. Amendt, M.<br />
Kinder, The kosmotropic (structure<strong>for</strong>ming)<br />
effect <strong>of</strong> compensatory solutes,<br />
Comp. Biochem. Physiol. 117A (1997) 357-<br />
65.<br />
133. T. Arakawa, S.N. Timasheff, The stabilization<br />
<strong>of</strong> proteins by osmolytes, Biophys. J. 47<br />
(1985) 411-4.<br />
134. D.W. Bolen, G.D. Rose, Structure and<br />
Energetics <strong>of</strong> the Hydrogen-Bonded<br />
Backbone in Protein Folding, Annu. Rev.<br />
Biochem. 77 (2008) 339-62.<br />
135. N. Borges, A. Ramos, N.D. Raven, R.J.<br />
Sharp, H. Santos, Comparative study<br />
<strong>of</strong> the thermostabilizing properties <strong>of</strong><br />
mannosylglycerate and other compatible<br />
solutes on model enzymes, Extremophiles 6<br />
(2002) 209-16.<br />
StabIlIzatIon <strong>of</strong> therapeutIc peptIdeS<br />
2<br />
33
Christina Avanti 1 , Jean-Pierre Amorij 1 , Dewi Setyaningsih 1 , Andrea Hawe 4 ,<br />
Wim Jiskoot 4 , Jan Visser 2 , Alexej Kedrov 3 , Arnold J. M. Driessen 3 ,<br />
Wouter L. J. Hinrichs 1 , and Henderik W. Frijlink 1<br />
1 Department <strong>of</strong> <strong>Pharma</strong>ceutical Technology and Biopharmacy,<br />
2 Department <strong>of</strong> <strong>Pharma</strong>cokinetics, Toxicology, and Targeting,<br />
3 Department <strong>of</strong> Molecular Microbiology,<br />
University <strong>of</strong> Groningen, Groningen, The Netherlands.<br />
4 Division <strong>of</strong> Drug Delivery Technology, Leiden/Amsterdam Center <strong>for</strong> Drug Research,<br />
Leiden University, Leiden, The Netherlands.
a new strategy to stabIlIze oxytocIn<br />
In aqueous solutIons: I. the effects <strong>of</strong><br />
dIvalent metal Ions and cItrate buffer<br />
AAP S Journal 2011; 13(2): 284–290<br />
3
3<br />
abstract<br />
In the current study, the effect <strong>of</strong> metal ions in combination with buffers (citrate, acetate,<br />
pH 4.5) on the stability <strong>of</strong> aqueous solutions <strong>of</strong> oxytocin was investigated. Both monovalent<br />
metal ions (Na + and K + ) and divalent metal ions (Ca 2+ , Mg 2+ , and Zn 2+ ) were tested all as<br />
chloride salts. The effect <strong>of</strong> combinations <strong>of</strong> buffers and metal ions on the stability <strong>of</strong> aqueous<br />
oxytocin solutions was determined by RP-HPLC and HP-SEC after 4 weeks <strong>of</strong> storage at<br />
either 4°C or 55°C. Addition <strong>of</strong> sodium or potassium ions to acetate- or citrate-buffered<br />
solutions did not increase stability, nor did the addition <strong>of</strong> divalent metal ions to acetate<br />
buffer. However, the stability <strong>of</strong> aqueous oxytocin in aqueous <strong>for</strong>mulations was improved in<br />
the presence <strong>of</strong> 5 and 10 mM citrate buffer in combination with at least 2 mM CaCl 2 , MgCl 2 ,<br />
or ZnCl 2 and depended on the divalent metal ion concentration. Isothermal titration<br />
calorimetric measurements were predictive <strong>for</strong> the stabilization effects observed during the<br />
stability study. Formulations in citrate buffer that had an improved stability displayed a<br />
strong interaction between oxytocin and Ca 2+ , Mg 2+ , or Zn 2+ , while <strong>for</strong>mulations in acetate<br />
buffer did not. In conclusion, our study shows that divalent metal ions in combination with<br />
citrate buffer strongly improved the stability <strong>of</strong> oxytocin in aqueous solutions.
1. IntroductIon<br />
According to the World Health Organization, half a million <strong>of</strong> women in Africa, Asia, and<br />
Latin America die each year due to problems during pregnancy and childbirth. At least 25%<br />
<strong>of</strong> those deaths can be attributed to bleeding after child birth (post-partum hemorrhage),<br />
mainly caused by failure <strong>of</strong> the uterus to contract adequately after child birth (atonicity) [1].<br />
The preferred drug to prevent post-partum hemorrhage is oxytocin. Oxytocin is a cyclic<br />
nonapeptide hormone [sequence: cyclo (Cys 1 -Tyr 2 -Ile 3 -Gln 4 -Asn 5 -Cys 6 ), -Pro 7 -Leu 8 -Gly 9 -NH 2 ],<br />
which is naturally produced in the hypothalamus. It is involved primarily in uterine contraction<br />
and stimulation <strong>of</strong> milk release from the mammary tissue [2]. Oxytocin, which is currently<br />
available in synthetic <strong>for</strong>m [3], has been widely used <strong>for</strong> indications such as induction <strong>of</strong><br />
labor, augmentation <strong>of</strong> labor, post-partum hemorrhage, or uterine atony, and also <strong>for</strong> other<br />
indications such as diabetes insipidus and vasodilatory shock. Reported additional functions<br />
<strong>for</strong> oxytocin include an antiduretic effect and blood vessel contraction [2,4,5].<br />
Un<strong>for</strong>tunately, oxytocin preparations are highly unstable at elevated temperatures,<br />
which is an issue particularly in tropical countries [6]. Stability studies conducted by<br />
Groot et al. [6] have shown that injectable oxytocin <strong>for</strong>mulations are rapidly degraded as<br />
the storage temperature rises to 30°C or higher. Oxytocic tablets <strong>for</strong> oral administration<br />
(ergometrine, ethylergometrine, oxytocin, and desamino-oxytocin) are also not stable<br />
under simulated tropical conditions. Because <strong>of</strong> its poor stability at elevated temperatures,<br />
the use <strong>of</strong> oxytocin in many developing countries is limited. Thus, there is a clear need <strong>for</strong> a<br />
heat-stable oxytocin <strong>for</strong>mulation, preferably an aqueous injectable solution, with improved<br />
thermal stability.<br />
One way to effectively improve stability <strong>of</strong> several peptides in aqueous solution is using<br />
metal salts in combination with a suitable buffer [7]. To investigate the effect <strong>of</strong> metal ions<br />
in buffered solutions on the stability <strong>of</strong> oxytocin, we screened various combinations <strong>of</strong><br />
unbuffered and buffered solutions with monovalent or divalent metal ions. Hawe et al. [8]<br />
observed that the degradation <strong>of</strong> oxytocin strongly depends on the pH <strong>of</strong> the <strong>for</strong>mulation,<br />
with the highest stability at pH 4.5. There<strong>for</strong>e, all <strong>for</strong>mulations will be set to pH 4.5. The<br />
purpose <strong>of</strong> this study is to investigate whether specific combinations <strong>of</strong> buffer and metal<br />
ions can stabilize oxytocin.<br />
2. materIals and method<br />
2.1 Materials<br />
The following materials were used in this study: oxytocin monoacetate powder (Diosynth,<br />
Oss, The Netherlands), citric acid, calcium chloride (Riedel-de Haen, Seelze, Germany),<br />
acetic acid, magnesium chloride, zinc chloride (Fluka, Steinheim, Germany), sodium<br />
hydroxide, sodium chloride, potassium chloride, sodium dihydrogen phosphate dihydrate,<br />
acetonitrile, <strong>for</strong>mic acid (Merck, Darmstadt, Germany) and Baxter Viavlo Ringer’s lactate<br />
solution <strong>for</strong> intravenous infusion (Baxter, Utrecht, The Netherlands).<br />
2.2 Formulation and Stability Study<br />
Oxytocin was <strong>for</strong>mulated at a concentration <strong>of</strong> 0.1 mg/ml in citrate (5 or 10 mM) or<br />
acetate (10 mM) buffer at pH 4.5 (pH adjusted with sodium hydroxide) with different<br />
additions <strong>of</strong> metal ions. pH samples were controlled and remained within ±0.1 pH units<br />
StabIlIty <strong>of</strong> the dIvalent Metal-cItrate-oxytocIn coMplex<br />
3<br />
37
3<br />
38<br />
during the stability study. The initial concentration <strong>of</strong> oxytocin was determined using UV<br />
spectrophotometry [9] at 280 nm with an extinction coefficient <strong>of</strong> 1.52 ml mg −1 cm −1 . All<br />
metal ion solutions were prepared using their chloride salts at concentrations <strong>of</strong> 2, 5, 10,<br />
and 50 mM. Control solutions were <strong>for</strong>mulated in water (pH 6.9±0.2) and Ringer’s lactate<br />
(6.4±0.2). Ringer’s lactate solution consists <strong>of</strong> 131 mM sodium, 5 mM potassium, 2 mM<br />
calcium, 111 mM chloride, and 29 mM bicarbonate (as lactate). In this report, the following<br />
codes were used: First character(s) refer to the type <strong>of</strong> buffer or water; CB (citrate),<br />
AC (acetate), RL (Ringer’s lactate), and W (water). Following digit(s) refer to buffer<br />
concentration in mM, following character(s) to the type <strong>of</strong> metal ion, and last digit(s) to<br />
metal ion concentration in mM. Thus, e.g., CB10Mg10 means 10 mM citrate buffer (pH 4.5)<br />
and 10 mM MgCl 2 . After preparation, the solutions were stored in 6R glass type 1 vials <strong>for</strong><br />
4 weeks at either 4°C or 55°C, and protected from light.<br />
Based on the results <strong>of</strong> the screening study, oxytocin <strong>for</strong>mulations in 10 mM citrate<br />
buffer pH 4.5 with 10 or 50 mM divalent metal salts were selected <strong>for</strong> a longer period<br />
<strong>of</strong> stability study <strong>for</strong> 6 months at 40°C according to ICH guidelines <strong>for</strong> long-term and<br />
accelerated stability study <strong>for</strong> climatic zone III and IV [10].<br />
2.3 Reversed-Phase High-Per<strong>for</strong>mance Liquid Chromatography<br />
The recovery <strong>of</strong> oxytocin (remaining oxytocin as percentage <strong>of</strong> initial amount) was<br />
determined by RP-HPLC. RPHPLC was per<strong>for</strong>med according to the procedure described<br />
by Hawe et al. [8] An Alltima C-18 RP column with 5 μm particle size, inner diameter <strong>of</strong><br />
4.6 mm, and length <strong>of</strong> 150 mm (Alltech, Ridderkerk, Netherlands), a Waters (Millipore)<br />
680 Automated Gradient Controller, two Waters 510 HPLC pumps, a Waters 717 Plus<br />
Autosampler, and a Waters 486 Tunable Absorbance UV Detector were used. Samples <strong>of</strong><br />
20 μl were injected and the separation was carried out at a flow rate <strong>of</strong> 1.0 ml/min and UV<br />
detection at 220 nm. Samples were eluted using 15% (v/v) acetonitrile in 65 mM phosphate<br />
buffer pH 5.0 as solvent A and 60% (v/v) acetonitrile in 65 mM phosphate buffer pH 5.0 as<br />
solvent B. The acetonitrile concentration was linearly increased from 15% at the beginning,<br />
to 20% at 10 min, to 30% at 20 min, and finally to 60% at 25 min.<br />
2.4 Size Exclusion HPLC<br />
The fraction <strong>of</strong> monomeric oxytocin (percentage <strong>of</strong> total remaining oxytocin) was assessed<br />
by Size Exclusion HPLC (HP-SEC). HP-SEC was carried out using a Superdex peptide<br />
10/300 GL column (GE Healthcare Inc., Brussels, Belgium) on an isocratic HPLC system,<br />
according to the method previously reported by Hawe et al. [8]. A Waters 510 pump, a<br />
Waters 717 plus auto sampler, a Waters 474 Scanning Fluorescence Detector and Waters 484<br />
Tunable Absorbance Detector (Waters, Mil<strong>for</strong>d Massachusetts, USA) were used. Samples<br />
<strong>of</strong> 50 μl were injected, and separation was per<strong>for</strong>med at a flow rate <strong>of</strong> 1 ml/min. Peaks<br />
were detected by UV absorption at 274 nm, as well as fluorescence detection at excitation<br />
wavelength <strong>of</strong> 274 nm and emission wavelength <strong>of</strong> 310 nm. The mobile phase consisted <strong>of</strong><br />
30% acetonitrile and 70% 0.04 M <strong>for</strong>mic acid.<br />
2.5 Isothermal Titration Calorimetry<br />
Isothermal titration calorimetry (ITC) was used to investigate the interaction between<br />
oxytocin and divalent metal ions in the presence <strong>of</strong> citrate buffer and acetate buffer.<br />
Microcalorimetric titrations <strong>of</strong> divalent metal ions to oxytocin were conducted by using a
MicroCal ITC 200 Microcalorimeter (Northampton, MA 01060 USA). A solution <strong>of</strong> 300 μL<br />
<strong>of</strong> 5 mM oxytocin in 10 mM <strong>of</strong> either citrate or acetate pH 4.5 was placed in the sample cell,<br />
while 30 μL <strong>of</strong> 125 mM divalent metal chloride either calcium, magnesium, or zinc in 10 mM<br />
citrate or acetate buffer pH 4.5 was placed in the syringe. The reference cell contained 300 μL<br />
<strong>of</strong> the corresponding buffer. Experiments were per<strong>for</strong>med at 55°C. Automated titrations<br />
were conducted up to a divalent metal ion/oxytocin molar ratio <strong>of</strong> 5:1. The effective heat <strong>of</strong><br />
the peptide-metal ion interaction upon each titration step was corrected <strong>for</strong> dilution and<br />
mixing effects, as measured by titrating the divalent metal ion solution into buffer and by<br />
titrating buffer into oxytocin solution. To investigate the possibility <strong>of</strong> oxytocin or metal ion<br />
binding to the buffer components, control experiments were per<strong>for</strong>med in water. The heats<br />
<strong>of</strong> bimolecular interactions were obtained by integrating the peak following each injection.<br />
All measurements were per<strong>for</strong>med in triplicate.<br />
ITC data were analyzed by using the ITC non-linear curve fitting functions <strong>for</strong> one<br />
or two binding sites from MicroCal Origin 7.0 s<strong>of</strong>tware (MicroCal S<strong>of</strong>tware, Inc.). The<br />
calculated curve was determined by the best-fit parameter, which was used to determine<br />
the molar enthalpy change <strong>for</strong> binding and the corresponding association constant, K a . The<br />
molar free energy <strong>of</strong> binding ΔG° and the molar entropy change ΔS° were derived from the<br />
fundamental equations <strong>of</strong> thermodynamics ΔG° = -RTlnK a and ΔG° = ΔH°- TΔS°.<br />
3. results<br />
3.1 Influence <strong>of</strong> Divalent Metal Ions on Oxytocin Stability in<br />
Unbuffered Solutions<br />
First, the effect <strong>of</strong> divalent metal ions on the stability <strong>of</strong> oxytocin in water without any buffer<br />
salt was investigated. RP-HPLC results (Fig. 1a) showed that after 4 weeks <strong>of</strong> storage at 4°C,<br />
oxytocin recovery was almost 100% in the presence <strong>of</strong> 2–50 mM zinc or 50 mM calcium<br />
ions. No stabilizing effect was observed from the presence <strong>of</strong> magnesium (2–50 mM)<br />
and calcium (2–10 mM), where the recovery was reduced to about 65%, similar to levels<br />
found <strong>for</strong> oxytocin solutions in water. HPSEC results (Fig. 1b) showed a similar trend in<br />
the recovery <strong>of</strong> monomeric oxytocin. However, when the solutions were stored at 55°C,<br />
both RP-HPLC and HP-SEC measurements showed substantial degradation <strong>of</strong> oxytocin<br />
after 4 weeks. These results demonstrate that divalent metal ions in non-buffered aqueous<br />
oxytocin <strong>for</strong>mulations have only a limited stabilizing effect at elevated temperature.<br />
3.2 Oxytocin Stability in Buffered Solutions<br />
To determine the effect <strong>of</strong> buffer on stability, oxytocin was <strong>for</strong>mulated in 5 and 10 mM<br />
citrate buffer and 10 mM acetate buffer.As a reference, the stability <strong>of</strong> oxytocin in pure water<br />
and in Ringer’s lactate buffer was investigated. Figure 2a shows the oxytocin recovery in<br />
RP-HPLC after 4 weeks <strong>of</strong> storage either at 4°C or 55°C in the buffered solutions. Compared<br />
to pure water, the stability <strong>of</strong> oxytocin was substantially increased at 4°C in the presence <strong>of</strong><br />
the buffer salts. After storage at 55°C, the recovery <strong>of</strong> oxytocin after 4 weeks in citrate and<br />
acetate buffer was much higher as compared to water or Ringer’s lactate solution. However,<br />
the recovery <strong>of</strong> oxytocin was still poor. In addition, only about 20% <strong>of</strong> oxytocin remained<br />
in its monomeric <strong>for</strong>m after storage (Fig. 2b).<br />
StabIlIty <strong>of</strong> the dIvalent Metal-cItrate-oxytocIn coMplex<br />
3<br />
39
3<br />
40<br />
Figure 1. Recovery <strong>of</strong> oxytocin in the presence <strong>of</strong> divalent metal ions in non-buffered pure<br />
water, stored <strong>for</strong> 4 weeks at pH 4.5 and a temperature <strong>of</strong> 4°C (light gray bars) or 55°C (dark<br />
gray bars). The divalent metal ions (Ca 2+ , Mg 2+ , and Zn 2+) were used in concentrations <strong>of</strong><br />
2, 5, 10, and 50 mM. a recovery determined by RP-HPLC. b oxytocin monomer recovery<br />
determined by HP-SEC. The results are depicted as averages <strong>of</strong> three independent<br />
measurements±SD<br />
3.3 Oxytocin Stability in Buffered Solutions Containing Monovalent<br />
Metal Ions<br />
To investigate the stability <strong>of</strong> oxytocin in the presence <strong>of</strong> a combination <strong>of</strong> buffer and<br />
monovalent metal ions, citrate and acetate buffer were used at a concentration <strong>of</strong> 10 mM,<br />
in combination with the monovalent metal ions, sodium and potassium, added at a<br />
concentration <strong>of</strong> 10 and 20 mM (excluding sodium from the buffer component). The
Figure 2. Recovery <strong>of</strong> oxytocin in pure water, with or without a buffer. Citrate buffer at a<br />
concentration <strong>of</strong> 5, 10, or 50 mM, acetate buffer at a concentration <strong>of</strong> 10 mM, or Ringer’s<br />
lactate solution were used. The <strong>for</strong>mulations contained no metal ions and were stored <strong>for</strong><br />
4 weeks at pH 4.5 or 6.4 <strong>for</strong> Ringer’s lactate solution at a temperature <strong>of</strong> 4°C (light gray<br />
bars) or 55°C (dark gray bars). a recovery determined by RP-HPLC. b oxytocin monomer<br />
recovery determined by HP-SEC. The results are depicted as averages <strong>of</strong> three independent<br />
measurements±SD<br />
presence <strong>of</strong> monovalent metal ions had only a minor effect on the stability <strong>of</strong> oxytocin<br />
(stored at 55°C). Although the oxytocin recovery was slightly improved compared to<br />
oxytocin in the presence <strong>of</strong> buffer alone, the maximum recovery <strong>of</strong> oxytocin was only 35%<br />
in 10 mM acetate buffer with 10 mM sodium chloride. In addition, only about 30% <strong>of</strong><br />
oxytocin in this <strong>for</strong>mulation remained monomeric (data not shown). These results clearly<br />
indicate that the presence <strong>of</strong> a combination <strong>of</strong> buffer and monovalent metal ions is not<br />
sufficient to substantially stabilize oxytocin in aqueous solution.<br />
StabIlIty <strong>of</strong> the dIvalent Metal-cItrate-oxytocIn coMplex<br />
3<br />
41
3<br />
42<br />
3.4 Oxytocin Stability in Citrate-Buffered Solutions Containing<br />
Divalent Metal Ions<br />
To study the stability <strong>of</strong> oxytocin in the presence <strong>of</strong> citrate buffers and divalent metal ions,<br />
<strong>for</strong>mulations containing 5 and 10 mM citrate buffer in combination with divalent metal ions<br />
(calcium, magnesium, and zinc) added at concentrations <strong>of</strong> 2, 5, 10, or 50 mM were used.<br />
The results <strong>of</strong> RP-HPLC and HP-SEC <strong>of</strong> <strong>for</strong>mulations in 10 mM citrate buffer are presented<br />
in Fig. 3a and b. The stability <strong>of</strong> oxytocin solutions was clearly improved when <strong>for</strong>mulating<br />
them with citrate buffer in combination with calcium ions. The oxytocin stability increased<br />
Figure 3. Effect <strong>of</strong> Ca 2+ (squares), Mg 2+ (circles), and Zn 2+ (triangles) concentration on the<br />
recovery <strong>of</strong> oxytocin in citrate buffer at the concentration <strong>of</strong> 10 mM after 4 weeks <strong>of</strong> storage<br />
at either 4°C or 55°C and pH 4.5. Solid symbols denoted 4°C storage, while open symbols<br />
correspond to 55°C. a recovery determined by RP-HPLC. b oxytocin monomer recovery<br />
determined by HP-SEC. The results are depicted as averages <strong>of</strong> three independent<br />
measurements±SD
with increasing calcium ion concentrations. The recovery <strong>of</strong> oxytocin and the remaining<br />
percentage <strong>of</strong> oxytocin monomers after 4 weeks <strong>of</strong> storage at 55°C were increased up to<br />
almost 80% in the presence <strong>of</strong> 50 mM calcium.<br />
Similar results were obtained <strong>for</strong> the combination <strong>of</strong> citrate with magnesium. The<br />
degradation <strong>of</strong> oxytocin in citrate buffer at 5 and 10 mM decreased with an increasing<br />
concentration <strong>of</strong> magnesium ions. Formulations with zinc ions in citrate buffer also<br />
preserved oxytocin during storage. These combinations exert even a stronger effect on<br />
oxytocin stability than combinations <strong>of</strong> citrate buffer and calcium or magnesium ions.<br />
Stability was strongly improved at zinc concentrations as low as 2 or 5 mM. Both oxytocin<br />
recovery (RP-HPLC) and the monomeric oxytocin fraction (HP-SEC) were substantially<br />
higher (up to 90%) in the presence <strong>of</strong> 10 mM zinc ions (CB5Zn10) after storage <strong>for</strong> 4 weeks<br />
at 55°C. When citrate buffer was used at a concentration <strong>of</strong> 5 mM, similar results were<br />
found (data not shown).<br />
Beside citrate, we also carried out several further experiments to investigate whether<br />
divalent metal ions affect the stability in the presence <strong>of</strong> acetate buffer. Acetate buffer at a<br />
concentration <strong>of</strong> 10 mM was used with and without calcium, magnesium, and zinc ions at<br />
various concentrations (2, 5, 10, 50 mM). The combination <strong>of</strong> these ions with acetate buffer<br />
was found to be less efficient in stabilizing oxytocin (data not shown).<br />
3.5 Long-term Stability <strong>of</strong> Oxytocin in Selected Formulations<br />
Containing Citrate and Divalent Metal Ions<br />
For the combination <strong>of</strong> citrate with Ca 2+ , Zn 2+ , and Mg 2+ , a long-term stability study<br />
<strong>for</strong> 6 months at 40°C was conducted. A temperature <strong>of</strong> 40°C was chosen to simulate<br />
tropical conditions [10,11]. The long-term stability study at 40°C clearly demonstrates<br />
the synergistic stabilizing effect <strong>of</strong> citrate buffer and divalent metal ions. Even though the<br />
oxytocin recovery decreased gradually with time, the recovery <strong>of</strong> oxytocin (Fig. 4a) and the<br />
remaining percentage <strong>of</strong> oxytocin monomers (Fig. 4b) after 6 months storage at 40°C were<br />
increased up to 80%in the presence <strong>of</strong> 50 mM calcium, and even higher (up to 90%) in the<br />
presence <strong>of</strong> 50 mM magnesium.<br />
Formulations with 10 mM zinc ions in citrate buffer exerted the same effect on oxytocin<br />
stability as combinations <strong>of</strong> citrate buffer and 50 mM magnesium ions. This shows that zinc<br />
ions at lower concentrations have already a higher impact on increasing oxytocin stability<br />
compared with calcium or magnesium ions.<br />
This result also confirms the short-term stability study that showed up to 90%remaining<br />
oxytocin in the presence <strong>of</strong> 10mM zinc ions after storage <strong>for</strong> 4 weeks at 55°C.<br />
3.6 ITC to Study the Interaction between Oxytocin and Divalent<br />
Metal Ions<br />
To examine the interaction between oxytocin and the metal ions, ITC experiments were<br />
carried out that are summarized in Table 1. The titration <strong>of</strong> calcium ions into an oxytocin<br />
solution in citrate buffer resulted in an exothermic reaction (Fig. 5a) with a K a value <strong>of</strong><br />
about 400 M −1 and an apparent ion/oxytocin stoichiometry close to 4:1 (Table 1).<br />
Remarkably, when titrating magnesium into oxytocin we observed heat absorption (Fig. 5a)<br />
and this endothermic reaction occurred with an identical apparent stochiometry (4:1) and a similar<br />
K a value <strong>of</strong> about 200 M −1 as with calcium ions. However, with magnesium ions the ITC trace was<br />
complex and showed an exothermic phase at magnesium concentrations below 5 mM. Due to<br />
StabIlIty <strong>of</strong> the dIvalent Metal-cItrate-oxytocIn coMplex<br />
3<br />
43
3<br />
44<br />
Figure 4. Oxytocin recovery over time storage at 40°C and pH 4.5 in the presence <strong>of</strong> 10<br />
mM citrate buffer, without (star) and with divalent metal ions. Ca 2+ (square), Mg 2+ (triangle),<br />
and Zn 2+ (circle) were used in concentrations <strong>of</strong> 10 mM (open symbols), and 50 mM (solid<br />
symbols). a recovery determined by RP-HPLC. b oxytocin monomer recovery determined<br />
by HP-SEC. The results are depicted as averages <strong>of</strong> three independent measurements±SD<br />
the rapid saturation <strong>of</strong> this early phase—typically, within three injections—and the low enthalpy<br />
<strong>of</strong> the ion/oxytocin interaction, it was not possible to quantitatively analyze the exothermic stage.<br />
To analyze the following endothermic phase we omitted the first 4-titration steps and fitted the<br />
remaining data using a single site model. A similar dual-phase response was observed <strong>for</strong> the<br />
zinc: oxytocin interaction (Fig. 5a, open triangles), but the exothermic and endothermic stages<br />
were well-resolved and suitable <strong>for</strong> the analysis using a two sites model (Table 1). In contrast, in<br />
the presence <strong>of</strong> acetate buffer there was no measurable interaction between oxytocin and calcium,<br />
magnesium, or zinc ion (Fig. 5b).
Table 1. Thermodynamics <strong>of</strong> Divalent Metal binding to Oxytocin as Determined by Isothermal Titration<br />
Calorimetry in 10 mM Citrate Buffer<br />
Metal Phase N (sites) Ka (M−1 ) ΔH°(cal mol−1 ) ΔS° (cal/mol/deg)<br />
Ca2+ 1 0.26 400 −2,100 5.6<br />
Mg2+ 1 n.d. n.d.
3<br />
46<br />
a b<br />
Figure 5. Least squares fit <strong>of</strong> the data from calorimetric titration pr<strong>of</strong>iles <strong>of</strong> aliquots <strong>of</strong> 125<br />
mM divalent metal ions: Ca 2+ (solid square), Mg 2+ (open square), and Zn 2+ (open triangle)<br />
into 5 mM oxytocin in 10 mM a citrate buffer and b acetate buffer pH 4.5. The heat absorbed<br />
per mol <strong>of</strong> titrant is plotted versus the ratio <strong>of</strong> the total concentration <strong>of</strong> divalent metal ions<br />
to the total concentration <strong>of</strong> oxytocin<br />
from water molecules. These con<strong>for</strong>mational changes can increase the stability <strong>of</strong> oxytocin in<br />
aqueous medium as they will prevent dimerization and further aggregation by hydrophobic<br />
interactions among oxytocin molecules. Metal salts are <strong>of</strong>ten used to stabilize peptides or<br />
proteins by chelation or ionic interactions [22]. Wang et al. examined the peptide (P66)<br />
stability in the presence <strong>of</strong> ZnCl 2 , MgCl 2 , and CaCl 2 in non aqueous solution, and found that<br />
in the presence <strong>of</strong> 1 mM ZnCl 2 , P66 was significantly stabilized. However in the aqueous<br />
solution (pure water), these ions did not show any stabilizing effect [22]. In our experiments,<br />
the addition <strong>of</strong> calcium, magnesium, and zinc ions in combination with citrate buffer had<br />
a large impact on oxytocin stability in contrast to similar experiments in pure water or<br />
acetate buffer. This study suggests that there is a synergistic effect between citrate buffer and<br />
the divalent metal ions, possibly due to the protection <strong>of</strong> the disulfide bridge by complex<br />
<strong>for</strong>mation <strong>of</strong> divalent metal ion and citrate with oxytocin which suppressed intermolecular<br />
reaction leading to tri/tetrasulfide <strong>for</strong>mation as well as dimerization (unpublished data).<br />
ITC is a sensitive method <strong>for</strong> studying the thermodynamics <strong>of</strong> binding events and<br />
quantifying binding reactions. When divalent metal ion are added to oxytocin, the ITC<br />
data indicate an interaction between oxytocin and Ca 2+ , Mg 2+ , or Zn 2+ ions in the presence<br />
<strong>of</strong> citrate buffer. Both Mg 2+ and Zn 2+ ions demonstrated complex, dual-phase interaction<br />
pr<strong>of</strong>ile, while a single phase was observed <strong>for</strong> Ca 2+ . Each interaction was entropy driven,<br />
while both exothermic and endothermic reactions were observed. It may be speculated that<br />
the solvation effect, i.e., release <strong>of</strong> structured water molecules plays a key role in binding,<br />
while the specific ion–oxytocin interaction further contributes to the complex stability.<br />
The latter is also predicted by molecular dynamic simulations [20,21]. Remarkably, no
interaction between oxytocin and either <strong>of</strong> the tested ions was detected in the acetate buffer<br />
or deionized water. These observations underscore the role <strong>of</strong> a particular environment<br />
in the ion–oxytocin interaction and agree well with our findings on the peptide stability.<br />
Isothermal titration calorimetric measurements were predictive <strong>for</strong> the effects observed<br />
during the stability study.<br />
In conclusion, this study shows that with a combination <strong>of</strong> divalent metal salts and citrate<br />
buffer, the stability <strong>of</strong> oxytocin in aqueous solution can be strongly improved. The increased<br />
stability <strong>of</strong> oxytocin aqueous <strong>for</strong>mulations was achieved in the presence <strong>of</strong> citrate acid buffer and<br />
2 mM or more <strong>of</strong> the salts CaCl 2 , MgCl 2 , or ZnCl 2 . The oxytocin stability is further increased with<br />
increasing concentration <strong>of</strong> the divalent metals ions up to 50 mM. In combination with citrate<br />
buffer, Zn 2+ has a superior stabilizing effect as compared with Ca 2+ or Mg 2+ .<br />
acknowledgments<br />
The authors want to thank MSD Oss <strong>for</strong> providing oxytocin <strong>for</strong> the study. This study was<br />
per<strong>for</strong>med within the framework <strong>of</strong> the Dutch Top Institute <strong>Pharma</strong> project: number<br />
D6–202. Open Access This article is distributed under the terms <strong>of</strong> the Creative Commons<br />
Attribution Noncommercial License which permits any noncommercial use, distribution,<br />
and reproduction in any medium, provided the original author(s) and source are credited.<br />
references<br />
1. K.S. Khan, D. Wojdyla, L. Say, A.M.<br />
Gulmezoglu, P.F. Van Look, WHO analysis<br />
<strong>of</strong> causes <strong>of</strong> maternal death: a systematic<br />
review, Lancet 367 (2006) 1066-74.<br />
2. V.S. Ananthanarayanan, K.S. Brimble,<br />
Interaction <strong>of</strong> oxytocin with Ca2+: I. CD and<br />
fluorescence spectral characterization and<br />
comparison with vasopressin, Biopolymers<br />
40 (1996) 433-43.<br />
3. E.H. Bishop, Synthetic oxytocin; a clinical<br />
evaluation, Obstet Gynecol 11 (1958) 290-4.<br />
4. K.P. Conrad, M. Gellai, W.G. North, H.<br />
Valtin, Influence <strong>of</strong> oxytocin on renal<br />
hemodynamics and electrolyte and water<br />
excretion, Am. J. Physiol. 251 (1986) F290-6.<br />
5. A.V. Somlyo, C.Y. Woo, A.P. Somlyo,<br />
Responses <strong>of</strong> Nerve-Free Vessels to<br />
Vasoactive Amines and Polypeptides, Am J<br />
Physiol 208 (1965) 748-53.<br />
6. A.N.J.A.D. Groot, T.B. Vree, H.V. Hogerzeil,<br />
G.J.A. Walker, WHO Action Programme on<br />
Essential Drugs: Stability <strong>of</strong> Oral Oxytocics<br />
in Tropical Climates : Results <strong>of</strong> Simulation<br />
Studies on Oral Ergometrine, Oral<br />
Methylergometrine, Buccal Oxytocin and<br />
Buccal Desamino-Oxytocin, World Health<br />
Organization, Geneva, 1994.<br />
7. M.C. Manning, D.K. Chou, B.M. Murphy,<br />
R.W. Payne, D.S. Katayama, Stability <strong>of</strong><br />
protein pharmaceuticals: an update, Pharm.<br />
Res. 27 (2010) 544-75.<br />
8. A. Hawe, R. Poole, S. Romeijn, P. Kasper,<br />
R. van der Heijden, W. Jiskoot, Towards<br />
heat-stable oxytocin <strong>for</strong>mulations: analysis<br />
<strong>of</strong> degradation kinetics and identification<br />
<strong>of</strong> degradation products, Pharm. Res. 26<br />
(2009) 1679-88.<br />
9. S.C. Gill, P.H. von Hippel, Calculation <strong>of</strong><br />
protein extinction coefficients from amino<br />
acid sequence data, Anal. Biochem. 182<br />
(1989) 319-26.<br />
10. International Conference on Harmonisation;<br />
Stability Data Package <strong>for</strong> Registration<br />
Applications in Climatic Zones III and IV;<br />
Stability Testing <strong>of</strong> New Drug Substances<br />
and Products; availability. Notice, Fed.<br />
Regist. 68 (2003) 65717-8.<br />
11. W. Grimm, Extension <strong>of</strong> the International<br />
Conference on Harmonization Tripartite<br />
Guideline <strong>for</strong> Stability Testing <strong>of</strong> New Drug<br />
Substances and Products to countries <strong>of</strong><br />
climatic zones III and IV, Drug Dev. Ind.<br />
Pharm. 24 (1998) 313-25.<br />
12. H.V. Hogerzeil, G.J.A. Walker, M.J. De Goeje,<br />
Stability <strong>of</strong> injectable ocytocics in tropical<br />
climates, World Health Organization,<br />
Geneva WHO/DAP/93.6. (1993).<br />
13. L.A. Trissel, Y. Zhang, K. Douglass, E.<br />
Kastango, Extended Stability <strong>of</strong> Oxytocin<br />
in common infusion solution, International<br />
Journal <strong>of</strong> <strong>Pharma</strong>ceutical Compounding 10<br />
(2006) 156-8.<br />
StabIlIty <strong>of</strong> the dIvalent Metal-cItrate-oxytocIn coMplex<br />
3<br />
47
3<br />
48<br />
14. A.B. Joshi, M. Sawai, W.R. Kearney, L.E.<br />
Kirsch, Studies on the mechanism <strong>of</strong> aspartic<br />
acid cleavage and glutamine deamidation in<br />
the acidic degradation <strong>of</strong> glucagon, J. Pharm.<br />
Sci. 94 (2005) 1912-27.<br />
15. H. Yang, R.A. Zubarev, Mass spectrometric<br />
analysis <strong>of</strong> asparagine deamidation and<br />
aspartate isomerization in polypeptides,<br />
Electrophoresis 31 (2010) 1764-72.<br />
16. N.E. Robinson, Protein deamidation, Procl.<br />
Natl. Acad. Sci. 99 (2002) 5283-8.<br />
17. C. Leeuwenburgh, J.E. Rasmussen, F.F. Hsu,<br />
D.M. Mueller, S. Pennathur, J.W. Heinecke,<br />
Mass spectrometric quantification <strong>of</strong><br />
markers <strong>for</strong> protein oxidation by tyrosyl<br />
radical, copper, and hydroxyl radical in low<br />
density lipoprotein isolated from human<br />
atherosclerotic plaques, J. Biol. Chem. 272<br />
(1997) 3520-6.<br />
18. A. Fiser, I. Simon, Predicting the oxidation<br />
state <strong>of</strong> cysteines by multiple sequence<br />
alignment, Bioin<strong>for</strong>matics 16 (2000) 251-6.<br />
19. M. Klingenberg, M. Appel, The uncoupling<br />
protein dimer can <strong>for</strong>m a disulfide cross-link<br />
between the mobile C-terminal SH groups,<br />
Eur J Biochem 180 (1989) 123-31.<br />
20. V.S. Ananthanarayanan, M.P. Belciug,<br />
B.S. Zhorov, Interaction <strong>of</strong> oxytocin with<br />
Ca2+: II. Proton magnetic resonance<br />
and molecular modeling studies <strong>of</strong><br />
con<strong>for</strong>mations <strong>of</strong> the hormone and its Ca2+<br />
complex, Biopolymers 40 (1996) 445-64.<br />
21. D. Liu, A.B. Seuthe, O.T. Ehrler, X. Zhang,<br />
T. Wyttenbach, J.F. Hsu, M.T. Bowers,<br />
Oxytocin-receptor binding: why divalent<br />
metals are essential, J. Am. Chem. Soc. 127<br />
(2005) 2024-5.<br />
22. W. Wang, S. Martin-Moe, C. Pan, L. Musza,<br />
Y.J. Wang, <strong>Stabilization</strong> <strong>of</strong> a polypeptide in<br />
non-aqueous solvents, Int. J. Pharm. 351<br />
(2008) 1-7.
Christina Avanti 1 , Hjalmar P. Permentier 2 , Annie van Dam 2 , Robert Poole 3 ,<br />
Wim Jiskoot 3 , Henderik W. Frijlink 1 , and Wouter L. J. Hinrichs 1<br />
1 Department <strong>of</strong> <strong>Pharma</strong>ceutical Technology & Biopharmacy and<br />
2 Mass Spectrometry Core Facility,<br />
University <strong>of</strong> Groningen, Groningen, The Netherlands<br />
3 Division <strong>of</strong> Drug Delivery Technology, Leiden/Amsterdam Center <strong>for</strong> Drug Research,<br />
Leiden University, Leiden, The Netherlands
a new strategy to stabIlIze oxytocIn<br />
In aqueous solutIons: II. suppressIon<br />
<strong>of</strong> cysteIne-medIated Intermolecular<br />
reactIons by a combInatIon <strong>of</strong><br />
dIvalent metal Ions and cItrate<br />
Mol. <strong>Pharma</strong>ceutics. 2012 ; 9(3): 554-62<br />
4
4<br />
abstract<br />
A series <strong>of</strong> studies have been conducted to develop a heat-stable liquid oxytocin <strong>for</strong>mulation.<br />
Oxytocin degradation products have been identified including citrate adducts <strong>for</strong>med in a<br />
<strong>for</strong>mulation with citrate buffer. In a more recent study we have found that divalent metal<br />
salts in combination with citrate buffer strongly stabilize oxytocin in aqueous solutions<br />
(Avanti, C.; et al. AAPS J. 2011, 13, 284−290). The aim <strong>of</strong> the present investigation was<br />
to identify various degradation products <strong>of</strong> oxytocin in citrate-buffered solution after<br />
thermal stress at a temperature <strong>of</strong> 70°C <strong>for</strong> 5 days and the changes in degradation pattern<br />
in the presence <strong>of</strong> divalent metal ions. Degradation products <strong>of</strong> oxytocin in the citrate<br />
buffer <strong>for</strong>mulation with and without divalent metal ions were analyzed using liquid<br />
chromatography−mass spectrometry/ mass spectrometry (LC−MS/MS). In the presence<br />
<strong>of</strong> divalent metal ions, almost all degradation products, in particular citrate adduct, tri-<br />
and tetrasulfides, and dimers, were greatly reduced in intensity. No significant difference<br />
in the stabilizing effect was found among the divalent metal ions Ca 2+ , Mg 2+ , and Zn 2+ . The<br />
suppressed degradation products all involve the cysteine residues. We there<strong>for</strong>e postulate<br />
that cysteine-mediated intermolecular reactions are suppressed by complex <strong>for</strong>mation <strong>of</strong><br />
the divalent metal ion and citrate with oxytocin, thereby inhibiting the <strong>for</strong>mation <strong>of</strong> citrate<br />
adducts and reactions <strong>of</strong> the cysteine thiol group in oxytocin.
1. IntroductIon<br />
Oxytocin is a neurohypophyseal hormone, which was first discovered by H. H. Dale in<br />
1909 [1,2]. Oxytocin is produced by neurons <strong>of</strong> the posterior lobe <strong>of</strong> the hypophysis and<br />
pulsatively released into the periphery. In clinical practice, oxytocin has been prescribed<br />
primarily <strong>for</strong> labor induction and augmentation, control <strong>of</strong> postpartum hemorrhage and<br />
uterine hypotonicity in the third stage <strong>of</strong> labor [3,4]. Oxytocin is commonly administered<br />
by intravenous infusion [5].<br />
The oxytocin structure was elucidated in 1951 [6-8] and the characterization and<br />
biosynthesis <strong>of</strong> oxytocin were reported in 1953 by du Vigneaud [9]. Oxytocin consists<br />
<strong>of</strong> nine amino acids: cyclo-(Cys 1 -Tyr 2 -Ile 3 -Gln 4 -Asn 5 -Cys 6 )-Pro 7 -Leu 8 -Gly 9 -NH 2 with a<br />
disulfide bridge between Cys residues 1 and [6,10,11]. The primary structure <strong>of</strong> oxytocin is<br />
shown in Figure 1. A major problem <strong>of</strong> the compound is its intrinsic instability in aqueous<br />
<strong>for</strong>mulations [12]. Recently significant attention was focused on ef<strong>for</strong>ts to overcome the<br />
instability <strong>of</strong> oxytocin [13,14]. We have conducted several studies with the aim to develop a<br />
heat-stable oxytocin <strong>for</strong>mulation.<br />
We identified the main degradation products <strong>of</strong> oxytocin stressed at a temperature <strong>of</strong><br />
70°C in various buffers, pH values and storage time [13] as well as citryl oxytocin in citratebuffered<br />
<strong>for</strong>mulations [14]. The degradation reactions and target residues <strong>of</strong> oxytocin are<br />
indicated in Figure 1.<br />
In a recent publication we described that <strong>for</strong>mulations containing divalent metal salts<br />
in combination with citrate buffer strongly stabilize oxytocin in aqueous solutions [15].<br />
However, the role by which divalent metal ions stabilize oxytocin in citrate buffer and<br />
their influence on (inhibition <strong>of</strong>) <strong>for</strong>mation <strong>of</strong> degradation products were not elucidated.<br />
There<strong>for</strong>e, the aim <strong>of</strong> the present study was to identify by LC−MS/MS the degradation<br />
products <strong>of</strong> oxytocin solutions after thermal stress and to investigate which degradation<br />
pathways are suppressed by the <strong>for</strong>mulations containing combinations <strong>of</strong> divalent metal<br />
ions and citrate buffer.<br />
Figure 1. Molecular structure <strong>of</strong> oxytocin with its ring and tail fragment ions <strong>for</strong>med upon<br />
MS/MS fragmentation.<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
53
4<br />
54<br />
2. materIals and methods<br />
2.1 Materials<br />
The following materials were used in this study: oxytocin monoacetate powder (Diosynth,<br />
Oss, The Netherlands), citric acid, calcium chloride (Riedel-de Haen, Seelze, Germany),<br />
magnesium chloride, zinc chloride (Fluka, Steinheim, Germany), sodium hydroxide,<br />
acetonitrile, ammonium acetate (Merck, Darmstadt, Germany), DL-dithiothreitol and<br />
ammonium bicarbonate (Sigma-Aldrich Chemie Gmbh, Steinheim, Germany).<br />
2.2 Formulation and Stability Study<br />
Three independent batches <strong>of</strong> each <strong>of</strong> the four <strong>for</strong>mulations given in Table 1 were prepared.<br />
The concentration <strong>of</strong> the citrate buffer was 10 mM, and the pH was adjusted to 4.5. All metal<br />
ion (M 2+ ) solutions were prepared using their respective chloride salts at concentrations <strong>of</strong><br />
10 mM (MCl 2 ). The solutions were stored <strong>for</strong> 5 days protected from light at either 4 or 70<br />
°C. Prior to analysis, the samples were diluted 10-fold with water.<br />
2.3 Reversed-Phase High Per<strong>for</strong>mance Liquid Chromatography<br />
HPLC was per<strong>for</strong>med using a Shimadzu LC system, consisting <strong>of</strong> LC-20AD gradient pumps<br />
and a SIL-20AC autosampler. Chromatographic separation was achieved on an Alltima C18<br />
column (2.1 × 150 mm 5 μm, Grace Davison Discovery Sciences). The injection volume was<br />
50 μL. Elution was per<strong>for</strong>med by a linear gradient from 5% to 60% B in 30 min, followed<br />
by an increase to 90% B in 1 min, where it was kept 4 min, after which it returned to the<br />
starting conditions. Eluent A was 95% water/5% acetonitrile (AcN), and eluent B was 5%<br />
water/95% AcN, both containing 0.05% v/v acetic acid and 10 mM ammonium acetate. The<br />
flow rate was 0.2 mL/min. The UV signal was recorded at 220 nm.<br />
2.4 Liquid Chromatography−Mass Spectrometry<br />
The HPLC system was coupled to an API 3000 triple-quadrupole mass spectrometer<br />
(Applied Biosystems/MDS Sciex) via a turbo ion spray source. The ionization was per<strong>for</strong>med<br />
by electrospray in the positive mode. Full scan spectra were recorded at a scan rate <strong>of</strong> 2 s<br />
from m/z 500 to 1300 and a step size <strong>of</strong> 0.2 amu with a declustering potential (DP) <strong>of</strong> 40 V<br />
and a focusing potential (FP) <strong>of</strong> 250 V. Product ion scans were acquired in specified time<br />
windows with a DP <strong>of</strong> 60 V, a FP <strong>of</strong> 300 V, a collision energy <strong>of</strong> 40 V and a collision cell exit<br />
Table 1. The Composition <strong>of</strong> Oxytocin Liquid Formulation<br />
<strong>for</strong>mulation a oxytocin metal ion buffer<br />
OCB 0.1 mM citrate 10 mM<br />
OCBCa 0.1 mM Ca2+ 10 mM citrate 10 mM<br />
OCBMg 0.1 mM Mg2+ 10 mM citrate 10 mM<br />
OCBZn 0.1 mM Zn2+ 10 mM citrate 10 mM<br />
a OCB = oxytocin in the absence <strong>of</strong> divalent metal ions in 10 mM citrate-buffered solution at pH 4.5.<br />
OCBCa = oxytocin in the presence <strong>of</strong> Ca 2+ in 10 mM citrate-buffered solution at pH 4.5.<br />
OCBMg = oxytocin in the presence <strong>of</strong> Mg 2+ in 10 mM citrate-buffered solution at pH 4.5.<br />
OCBZn = oxytocin in the presence <strong>of</strong> Zn2+ in 10 mM citratebuffered solution at pH 4.5.
potential <strong>of</strong> 20 V. Data acquisition and processing was per<strong>for</strong>med using Analyst version<br />
1.4.2 and 1.5 s<strong>of</strong>tware (Applied Biosystems/MDS Sciex).<br />
LC−MS/MS were used to identify the oxytocin degradation products observed by<br />
RP-HPLC. The analyses were carried out on purified fractions corresponding to each <strong>of</strong> the<br />
major degradation peaks <strong>for</strong> oxytocin in citrate buffer <strong>for</strong>mulation without divalent metal<br />
salts as well as by LC−MS <strong>for</strong> the <strong>for</strong>mulation with and without divalent metal salts.<br />
Reduction <strong>of</strong> disulfide bonds was per<strong>for</strong>med by adding 10 mM DL-dithiothreitol (DTT)<br />
in 200 mM ammonium carbonate. The samples were analyzed after an incubation time <strong>of</strong> at<br />
least 10 min at room temperature.<br />
3. results<br />
3.1 Degradation Products <strong>of</strong> Oxytocin in Citrate Buffer with and<br />
without Divalent Metal Ions<br />
Assignment <strong>of</strong> degradation products was done based on the m/z value <strong>of</strong> each compound<br />
from the LC−MS data, using known assignments and retention time order from previous<br />
studies [13,14]. MS/MS data were used <strong>for</strong> the more intense peaks to confirm the identification<br />
and to determine in which part <strong>of</strong> the molecule the modification had occurred.<br />
The highest degradation product intensities were found in the <strong>for</strong>mulation without<br />
divalent metal salts stressed at 70°C <strong>for</strong> 5 days. The HPLC pr<strong>of</strong>iles based on ultraviolet (UV)<br />
absorbance at 220 nm and the total ion current (<strong>TI</strong>C) <strong>of</strong> mass spectra with m/z between<br />
900 and 1200 are shown in Figure 2a and Figure 2b, respectively. The UV pr<strong>of</strong>ile shows<br />
11 main peaks and the <strong>TI</strong>C 12 main peaks (see labels in Figure 2b). The LC−MS contour<br />
plot (Figure 2c) reveals that some peaks contain several molecular species, which cannot be<br />
distinguished in the <strong>TI</strong>C, notably peaks 2 and 3 reveal two compounds at 12.0 and 12.1 min,<br />
peaks 10 and 11 reveal two compounds at 16.7 and 16.8 min, and peaks 14 and 15 reveal two<br />
compounds at 18.7 and 18.9 min. All 15 assigned molecular species are presented in Table 2,<br />
and MS/MS product ion spectra are shown in Figures S1−S10 in the Supporting In<strong>for</strong>mation.<br />
The protonated molecular ion <strong>of</strong> unmodified oxytocin was found at a retention time <strong>of</strong><br />
12.6 min. Figure 2d shows its extracted ion chromatogram (XIC) at m/z 1007.6. MS/MS<br />
fragmentation <strong>of</strong> unmodified oxytocin results in fragments <strong>of</strong> m/z 723.4 and 285.2, from<br />
the disulfide-linked ring <strong>of</strong> residues 1−6 and the C-terminal residues 7−9, respectively<br />
(Figure 1 and Figure S4 in the Supporting In<strong>for</strong>mation).<br />
The degradation peaks eluting be<strong>for</strong>e the oxytocin peak (labeled 1 and 2 in Figure 2b)<br />
were identified as amide- and imide-linked N-citryl oxytocin with m/z <strong>of</strong> 1181.8 and<br />
1163.8, respectively.14 The MS/MS spectrum <strong>of</strong> m/z 1181.6 showed that the ring fragment<br />
was shifted from m/z 723.4 to 897.4, but the tail fragment was still observed at m/z 285.2.<br />
It shows that the citrate modification is in the ring fragment and the most likely location is<br />
on the N-terminal amine (see Figure 1 and Figure S2 in the Supporting In<strong>for</strong>mation) [14].<br />
At a retention time <strong>of</strong> 12.0 and 12.1 min (Figure 3b and 3c), there is overlap <strong>of</strong> two<br />
different compounds (peaks 2 and 3), namely a monodeamidated species <strong>of</strong> oxytocin with<br />
a protonated molecular ion at m/z 1008.6 in a very low intensity and dehydrated (imidelinked)<br />
N-citryl oxytocin with the protonated molecular ion at m/z 1163.6 [14]. The MS/<br />
MS spectrum <strong>of</strong> m/z 1008.6 showed a shift <strong>of</strong> the ring fragment from m/z 723.4 to 724.4, but<br />
no shift <strong>of</strong> the tail fragment mass (Figure S3 in the Supporting In<strong>for</strong>mation), which means<br />
that the deamidation occurred at Asn 5 or Gln 4 [13]. The 13 C denoted as peak in Figure 3c is<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
55
4<br />
56<br />
Figure 2. LC−UV trace (a), total ion chromatogram (m/z 900−1200) (b), and LC−MS contour<br />
plot (c) <strong>of</strong> stressed oxytocin in citrate buffered solution with the extracted ion current <strong>of</strong><br />
oxytocin (d) and acetylated oxytocin (e).
Table 2. Summary <strong>of</strong> the Observed Degradation Products <strong>of</strong> Oxytocin after Stressing at 70°C <strong>for</strong> 5 Days<br />
in 10 mM Citrate Buffer<br />
no. t (min) R m/z charge Assignment<br />
1 11.8 1181.6 1+ oxytocin N-citryl amide<br />
2 12.0 1008.6 1+ monodeamidation at Asn5 or Gln4 3 12.1 1163.6 1+ oxytocin N-citryl imide<br />
4 12.6 1007.6 1+ oxytocina 5 13.5 1039.6 1+ oxytocin trisulfide<br />
6 13.9 1195.6 1+ oxytocin trisulfide N-citryl imide<br />
7 14.4 1049.6 1+ N-acetyl oxytocina 8 14.8 1071.6 1+ oxytocin tetrasulfide<br />
9 15.5 975.6 2+ dimer 1<br />
10 16.7 975.6 2+ dimer 2<br />
11 16.8 1008.6 2+ dimer 3<br />
12 17.4 1024.6b 2+ dimer 4<br />
13 17.8 1008.6 2+ dimer 5<br />
14 18.7 1024.6b 2+ dimer 6<br />
15 18.9 1008.6 2+ dimer 7<br />
a No degradation products; compounds were also found in the original oxytocin preparation. b Ammoniated.<br />
the second isotope peak <strong>of</strong> oxytocin (m/z 1007.6), which is also found at m/z 1008.6. The<br />
three other peaks in Figure 3b, at retention times <strong>of</strong> 16.8, 17.8, and 18.9 min, are the dimer<br />
degradation products described in Figure 4c.<br />
A compound at m/z 1049.6 was found at a retention time <strong>of</strong> 14.4 min (Figure 2e) in every<br />
sample including the original oxytocin preparation. The difference <strong>of</strong> +42 Da with respect<br />
to oxytocin suggests acetylation, presumably at the N-terminal amine (see Figure S6 in the<br />
Supporting In<strong>for</strong>mation). This acetylation reaction may have occurred during synthesis<br />
since oxytocin is supplied as the monoacetate salt.<br />
At a retention time <strong>of</strong> 13.5 min a trisulfide degradation product was identified from<br />
the m/z 1039.6 peak. Upon MS/ MS fragmentation, the ring fragment shifted to m/z 755.4,<br />
while the tail fragment mass did not change (see Figure S5 in the Supporting In<strong>for</strong>mation).<br />
The mass difference <strong>of</strong> +32 Da can be assigned to a trisulfide modification in the oxytocin<br />
ring [13]. At a retention time <strong>of</strong> 13.9 min, a product <strong>of</strong> m/z 1195.6 was found, which is<br />
tentatively identified as the N-citryl oxytocin with the trisulfide modification. The <strong>for</strong>mation<br />
<strong>of</strong> a tetrasulfide product (m/z 1071.6) was found at the retention time <strong>of</strong> 14.8 min. Molecular<br />
structure N-citryl oxytocin, trisulfide and tetrasulfide modification produced from the<br />
degradation <strong>of</strong> stressed oxytocin in citrate buffered solution are shown in Figure 5.<br />
Seven peaks <strong>of</strong> dimers were evident as doubly charged ions in MS eluting between 15.5<br />
and 18.9 min (Table 2 and Figure 4). Dimer 1 and 2 had identical masses (m/z 975.6) and<br />
were also observed in a previous study <strong>of</strong> stressed oxytocin solutions [13] where they were<br />
assigned to the doubly protonated <strong>for</strong>m <strong>of</strong> a sulfur-linked dimeric oxytocin species that has<br />
been doubly deamidated and which has lost two sulfur atoms through β-elimination [13].<br />
Identification <strong>of</strong> dimers is difficult due to the overlapping LC peaks and multiple ion <strong>for</strong>ms<br />
(protonated and ammoniated, and doubly charged, Figure 2c). In addition, their MS/MS<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
57
4<br />
58<br />
Figure 3. LC−MS extracted ion currents <strong>for</strong> oxytocin N-citryl amide (a) and N-citryl imide<br />
(b), monodeamidation (c), trisulfide (d), trisulfide N-citryl imide (e), and tetrasulfide products<br />
from the degradation <strong>of</strong> stressed oxytocin in citrate buffered solution.
Figure 4. LC−MS extracted ion currents <strong>for</strong> dimer degradation products.<br />
fragmentation spectra only show the tail fragment <strong>of</strong> m/z 285.2 (Figures S7−10 in the<br />
Supporting In<strong>for</strong>mation) and consequently they are not in<strong>for</strong>mative.<br />
In order to distinguish disulfide linked dimers from other types <strong>of</strong> dimers, we have<br />
done the LC−MS analysis on the most degraded solution after disulfide bond reduction<br />
with DTT. All monomeric products observed were reduced, as summarized in Table 3,<br />
producing oxytocin, oxytocin acetate, and N-citryl amide/imide in the reduced thiol<br />
<strong>for</strong>m (Figures S11−S17 in the Supporting In<strong>for</strong>mation). Tri- and tetrasulfide <strong>for</strong>ms were<br />
also absent after reduction and are expected to have been converted to reduced oxytocin.<br />
The product ion spectrum after DTT reduction <strong>of</strong> m/z 993.6 at a retention time <strong>of</strong> 16.6<br />
min (Figure S18 in the Supporting In<strong>for</strong>mation) showed that it is a monomer which has<br />
undergone β-elimination at Cys 1 and deamidation at Asn 5 or Gln 4 . This is as evidenced by<br />
the presence <strong>of</strong> an unmodified y 4 ion and a b 6 ion (Figure S1 in the Supporting In<strong>for</strong>mation)<br />
mass <strong>of</strong> −33 Da with respect to oxytocin. β-Elimination leads to a loss <strong>of</strong> 34 Da (-H 2 S), and<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
59
4<br />
60<br />
Figure 5. Molecular structures <strong>of</strong> oxytocin N-citryl amide (a), N-citryl imide (b), trisulfide (c) and<br />
tetrasulfide (d) produced from the degradation <strong>of</strong> stressed oxytocin in citrate buffered solution.<br />
deamidation increases the mass by 1 Da. The presence <strong>of</strong> this reduced monomer at high<br />
intensity confirms the assignment <strong>of</strong> dimers 1 and 2, which proves that these dimers are<br />
disulfide linked at the unmodified Cys 6 residues.<br />
In the LC−MS pr<strong>of</strong>ile (Figure 6) it was also found that the LC−MS peaks <strong>of</strong> all dimers<br />
had disappeared after reduction. However there were 2 new peaks present at retention<br />
times <strong>of</strong> 17.9 and 18.1 min with m/z 994.0 and 993.6 respectively, which are doubly charged<br />
ion (Figures S19−S20 in the Supporting In<strong>for</strong>mation). The presence <strong>of</strong> dimers after DTT<br />
reduction, combined with the observation that all original 7 dimer peaks were absent after<br />
reduction, shows that all dimers had at least one disulfide bond and some had an additional<br />
thio-ether bond (which may <strong>for</strong>m after β-elimination and reaction with another Cys<br />
residue). Tyrosine-linked dimers were unlikely to be present, since no monomeric tyrosine<br />
oxidation products were observed.<br />
3.2 Effects <strong>of</strong> Divalent Metal Ions on Oxytocin’s Degradation Pr<strong>of</strong>ile<br />
Addition <strong>of</strong> divalent metal salts in the <strong>for</strong>mulation did not result in any additional degradation<br />
products compared to samples without metal salts. As expected from our previous study [14],<br />
it resulted in a significant reduction <strong>of</strong> most degradation peaks. Figure 7 shows that, <strong>of</strong> all<br />
<strong>for</strong>mulations tested, oxytocin <strong>for</strong>mulated in 10 mM citrate buffer (pH 4.5) without divalent<br />
metal ions (OCB) was most degraded with only approximately 35% oxytocin recovered after<br />
incubation at 70°C <strong>for</strong> 5 days. However, when divalent metal ions were added, the degradation<br />
was decelerated and a recovery <strong>of</strong> about 70% oxytocin was found, irrespective <strong>of</strong> the nature <strong>of</strong><br />
the divalent metal ion, which is in line with our previous study [15].
Table 3. Summary <strong>of</strong> the Observed Degradation Products <strong>of</strong> Oxytocin after Stressing at 70 °C <strong>for</strong> 5 Days in<br />
10 mM Citrate Buffer Followed by Reduction by DTT a<br />
no. t (min) R m/z charge Assignment<br />
a 12.8 1007.4 1+ oxytocin (disulfide)<br />
b 13.1 1200.6b 1+ oxytocin N-citryl amide<br />
c 13.5 1009.6 1+ oxytocin<br />
d 14.6 1182.6b 1+ oxytocin N-citryl imide<br />
e 15.5 1068.8b 1+ N-acetyl oxytocin<br />
f 16.6 993.6b 1+ β-elimination at Cys1 , monodeamidation at Asn5 or Gln4 g 17.9 994.0 2+ dimer<br />
h 18.1 993.6 2+ dimer<br />
a Assigned products are in reduced <strong>for</strong>m. b Ammoniated.<br />
From each stressed <strong>for</strong>mulation <strong>of</strong> oxytocin in citrate buffer in the presence <strong>of</strong> calcium,<br />
magnesium, or zinc ions, we recorded the intensity <strong>of</strong> each degradation product from their<br />
respective extracted MS ion currents. The relative intensities in percentage <strong>of</strong> the peak area<br />
<strong>of</strong> each degradation product in the presence <strong>of</strong> divalent metal ions with respect to that <strong>of</strong><br />
the same product in the absence <strong>of</strong> divalent metal ions are listed in Table 4.<br />
Addition <strong>of</strong> divalent metal ions reduced the peak intensity <strong>of</strong> almost all degradation<br />
products. The <strong>for</strong>mation <strong>of</strong> the N-citryl oxytocin and the tri- and tetrasulfide was reduced<br />
by 20−70%. The <strong>for</strong>mation <strong>of</strong> the dehydrated N-citryl oxytocin and dimmers 1, 2, 4, 5 was<br />
even more strongly reduced, by 90%. No clear reduction <strong>of</strong> the intensity <strong>of</strong> the deamidated<br />
species (m/z 1008.6, retention time 12.1 min) and dimer 3 (m/z 1024.8, retention time<br />
17.4 min) was observed, possibly because <strong>of</strong> their low signal intensity which made peak<br />
integration difficult (Figures 3b and 4b).<br />
The acetylated oxytocin species (peak number 7) appeared to be very stable: no<br />
significant difference in its relative intensity was observed either in the presence or in the<br />
absence <strong>of</strong> divalent metal ions. In addition, there is no significant difference <strong>of</strong> the intensity<br />
<strong>of</strong> this product in unstressed or stressed condition. We conclude that this molecule is not a<br />
degradation product <strong>for</strong>med under stress conditions but, as mentioned above, an impurity<br />
<strong>of</strong> oxytocin <strong>for</strong>med during synthesis.<br />
4. dIscussIon<br />
Several biological and physicochemical methods have been described to monitor<br />
the stability <strong>of</strong> oxytocin. HPLC with UV/vis detection is the most frequently used<br />
physicochemical method to monitor oxytocin stability. Especially when combined with<br />
mass spectrometry (MS) detection, most degradation products can be identified and<br />
quantified [16,17]. Hyphenation <strong>of</strong> the LC unit to MS via an electrospray ionization (ESI)<br />
interface allows sensitive and selective confirmation <strong>of</strong> degradation products by extracting<br />
corresponding ion chromatograms from the recorded total ion current (<strong>TI</strong>C)[14] and<br />
by using MS/MS <strong>for</strong> the identification <strong>of</strong> degradation products. Furthermore, previous<br />
studies have demonstrated that LC−MS can be used to monitor the stability <strong>of</strong> oxytocin in<br />
pharmaceutical dosage <strong>for</strong>ms [13,18].<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
61
4<br />
62<br />
Figure 6. Total ion chromatogram (m/z 900−1200) <strong>of</strong> stressed oxytocin in citrate buffered<br />
solution followed by reduction using DTT.<br />
Figure 7. Recovery <strong>of</strong> oxytocin in the absence (OCB) and presence <strong>of</strong> 10 mM Ca 2+ (OCBCa),<br />
Mg 2+ (OCBMg) and Zn 2+ (OCBZn) in 10 mM citrate-buffered solution at pH 4.5 under<br />
stressed condition at a temperature <strong>of</strong> 70 °C <strong>for</strong> 5 days. Oxytocin recovery determined by<br />
LC−MS. The results are depicted as averages <strong>of</strong> three independent measurements±SD.<br />
The degradation <strong>of</strong> oxytocin in various <strong>for</strong>mulations was analyzed after 5 days<br />
<strong>of</strong> incubation at 70°C. A temperature <strong>of</strong> 70°C was chosen to ensure the <strong>for</strong>mation <strong>of</strong><br />
qualitatively similar degradation products as found by Poole et al., who incubated oxytocin<br />
in citrate buffered solutions at 70°C <strong>for</strong> 2 days [14]. The aim <strong>of</strong> the present study was to<br />
investigate the effect <strong>of</strong> stabilizing divalent metal ions on the degradation pr<strong>of</strong>ile.<br />
In order to observe sufficient amounts <strong>for</strong> characterization <strong>of</strong> all degradation products in<br />
the stabilized <strong>for</strong>mulations, it was necessary to prolong the thermal stress from 2 to 5 days.<br />
The strongly decreased intensity <strong>of</strong> the dimeric products suggests that <strong>for</strong>mulation with<br />
divalent metal ions protects against thiol exchange, hence avoiding dimerization. At low<br />
pH, dimerization occurs via thiol exchange in the disulfide bridge, which progresses via
Table 4. The Effect <strong>of</strong> Divalent Metal Ions on the Intensity <strong>of</strong> Degradation Products <strong>of</strong> Oxytocin Found<br />
after Stressing at 70°C <strong>for</strong> 5 Days in 10 mM Citrate Buffer a<br />
Assignment<br />
relative peak intensity (%)<br />
OCBCa OCBMg OCBZn<br />
oxytocin N-citryl amide 71 ± 19 67 ± 20 52 ± 12<br />
monodeamidation at Asn5 or Gln 4 534 ± 86 b 436 ± 65 b 906 ± 50 b<br />
oxytocin N-citryl imide 58 ± 19 73 ±9 50± 8<br />
oxytocin trisulfide 40 ± 39 37 ±5 30± 28<br />
oxytocin trisulfide N-citryl imide 3 ±4 4±5 1± 1<br />
N-acetyl oxytocin c 107 ±7 89± 13 114 ± 13<br />
oxytocin tetrasulfide 53 ± 64 31 ±5 33± 31<br />
dimer 1 5 ±1 5±4 2± 1<br />
dimer 2 7 ±1 10± 7 3± 2<br />
dimer 3 15 ± 11 14 ±7 6± 4<br />
dimer 4 172 ± 42 b 150 ± 54 b 90 ± 7 b<br />
dimer 5 15 ± 11 15 ±5 6± 5<br />
dimer 6 27 ± 21 17 ±9 17± 16<br />
dimer 7 43 ± 42 b 42 ± 11 b 31 ± 30 b<br />
a The level <strong>of</strong> degradation products is expressed relative to the levels found in the solution without metal ions.<br />
b Weak MS signal. c No degradation product; compound was also found in the original oxytocin preparation.<br />
a sulfonium cation, which is <strong>for</strong>med following protonation <strong>of</strong> the disulfide bridge [19,20].<br />
This could also explain the ability <strong>of</strong> these <strong>for</strong>mulations to inhibit the <strong>for</strong>mation <strong>of</strong> tri/<br />
tetrasulfide oxytocin species.<br />
The combination <strong>of</strong> citrate buffer and divalent metal ions greatly reduces the <strong>for</strong>mation<br />
<strong>of</strong> most dimers after thermal stress. The absence <strong>of</strong> a significant decrease in the intensity<br />
<strong>of</strong> dimer 4 hardly contributes to the total amount <strong>of</strong> dimers because its intensity in the<br />
<strong>for</strong>mulation without divalent metal ions after thermal stress was already low.<br />
In this study, the buffer used was citrate, which is considered to be safe and occurs in<br />
many foods and is also a normal metabolite in the body. Its calcium, potassium and sodium<br />
salts do not constitute a significant hazard to humans [21]. After thermal stress, covalent<br />
citrate-oxytocin adducts were <strong>for</strong>med through a mechanism involving the intermediate<br />
production <strong>of</strong> citrate anhydride, which reacts with the N-terminal amino group from the<br />
cysteine residue [14]. Inhibition <strong>of</strong> the <strong>for</strong>mation <strong>of</strong> N-citryl oxytocin by divalent metal<br />
ions as found in the present study might be due to a differential interaction <strong>of</strong> oxytocin and<br />
divalent metal ions <strong>for</strong> citrate. Wyttenbach et al [22] found that under acidic conditions<br />
(pH 3.0) divalent metal ions bind to the carbonyl groups in the ring structure <strong>of</strong> oxytocin.<br />
The presence <strong>of</strong> doubly charged cations, such as Zn 2+ , Mg 2+ , Ni 2+ , Mn 2+ and Co 2+ , has been<br />
found to be essential in increasing the potency <strong>of</strong> the specific binding <strong>of</strong> oxytocin to its<br />
receptor. There<strong>for</strong>e the complex <strong>for</strong>med might increase the biological activity [23]. Further,<br />
isothermal titration calorimetry data showed that citrate interacts with divalent metal ions,<br />
and this interaction is stronger than that <strong>of</strong> citrate with oxytocin (Tables SI and SII and<br />
Figures S21−S22 in the Supporting In<strong>for</strong>mation). Free citrate ions are probably more reactive<br />
toward the N-terminal amino group from the cysteine residue than divalent metal−citrate<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
63
4<br />
64<br />
adducts. In the presence <strong>of</strong> divalent metal ions, the <strong>for</strong>mation <strong>of</strong> metal−citrate adducts<br />
reduces the concentration <strong>of</strong> free citrate, which reduces the driving <strong>for</strong>ce <strong>for</strong> citrate adduct<br />
<strong>for</strong>mation with oxytocin [15]. A second possibility is that binding <strong>of</strong> a divalent metal ion to<br />
oxytocin may change the position <strong>of</strong> the N-terminal amino group from the cysteine residue<br />
rendering it less accessible <strong>for</strong> binding with citrate.<br />
It can be concluded that citrate has two opposite effects on oxytocin stability. First, as<br />
also shown by Poole et al.,[14] it is reactive itself and can attack the N-terminal amino group<br />
from the cysteine residue to <strong>for</strong>m an adduct. Second, as shown in our previous study, [15]<br />
it protects oxytocin from degradation in the presence <strong>of</strong> divalent metal ions. In this study,<br />
we clearly show that the stabilization is due to the suppression <strong>of</strong> N-citryl oxytocin, tri/<br />
tetrasulfide and dimer <strong>for</strong>mation. Furthermore, all reactions that were suppressed occurred<br />
on Cys 1 , and possibly Cys 6 . Cysteine is susceptible to oxidation and β-elimination, and<br />
degradation <strong>of</strong> oxytocin involving cysteine leads to dimerization, <strong>for</strong>mation <strong>of</strong> tri/<br />
tetrasulfide and β-elimination followed by thio-ether <strong>for</strong>mation.<br />
Divalent metal ions in combination with citrate buffer suppress intermolecular reactions<br />
in the ring structure <strong>of</strong> oxytocin presumably by <strong>for</strong>ming a complex in the region where the<br />
degradation reaction occurs. No significant difference was observed among the three tested<br />
divalent metal ions, Ca 2+ , Mg 2+ , and Zn 2+ , suggesting that divalency is the most important<br />
property <strong>of</strong> the metal contributing to stabilization <strong>of</strong> the oxytocin-metal-citrate cluster.<br />
assocIated content<br />
*Supporting In<strong>for</strong>mation<br />
Figures depicting the MS/MS series <strong>of</strong> reduced oxytocin, numerous product ion spectra,<br />
and calorimetric titration pr<strong>of</strong>iles and tables <strong>of</strong> thermodynamics data.<br />
acknowledgments<br />
The authors want to thank MSD Oss <strong>for</strong> providing oxytocin <strong>for</strong> the study. This study was<br />
per<strong>for</strong>med within the framework <strong>of</strong> the Dutch Top Institute <strong>Pharma</strong> project: number D6−202.<br />
references<br />
1. H.H. Dale, The Action <strong>of</strong> Extracts <strong>of</strong> the<br />
Pituitary Body, Biochem. J. 4 (1909) 427-47.<br />
2. C.M. Karbiwnyk, K.C. Faul, S.B. Turnipseed,<br />
W.C. Andersen, K.E. Miller, Determination <strong>of</strong><br />
oxytocin in a dilute IV solution by LC-MS(n),<br />
J harm Biomed Anal 48 (2008) 672-7.<br />
3. J. Owen, J.C. Hauth, Oxytocin <strong>for</strong> the<br />
induction or augmentation <strong>of</strong> labor, Clin<br />
Obstet Gynecol 35 (1992) 464-75.<br />
4. F.A. Chaibva, R.B. Walker, Development and<br />
validation <strong>of</strong> a stability-indicating analytical<br />
method <strong>for</strong> the quantitation <strong>of</strong> oxytocin<br />
in pharmaceutical dosage <strong>for</strong>ms, J Pharm<br />
Biomed Anal 43 (2007) 179-85.<br />
5. J.W. Gard, J.M. Alexander, R.E. Bawdon,<br />
J.T. Albrecht, Oxytocin preparation stability<br />
in several common obstetric intravenous<br />
solutions, Am. J. Obstet. Gynecol. 186<br />
(2002) 496-8.<br />
6. R.A. Turner, J.G. Pierce, V. du Vigneaud, The<br />
purification and the amino acid content <strong>of</strong><br />
vasopressin preparations, J. Biol. Chem. 191<br />
(1951) 21-8.<br />
7. H. Davoll, R.A. Turner, J.G. Pierce, V. du<br />
Vigneaud, An investigation <strong>of</strong> the free<br />
amino groups on oxytocin and desulfurized<br />
oxytocin preparations, J. Biol. Chem. 193<br />
(1951) 363-70.<br />
8. M. Winkler, W. Rath, A risk-benefit<br />
assessment <strong>of</strong> oxytocics in obstetric practice,<br />
Drug. Saf. 20 (1999) 323-45.
9. V. du Vigneaud, C. Ressler, S. Trippett, The<br />
sequence <strong>of</strong> amino acids in oxytocin, with a<br />
proposal <strong>for</strong> the structure <strong>of</strong> oxytocin, J. Biol.<br />
Chem. 205 (1953) 949-57.<br />
10. V.S. Ananthanarayanan, K.S. Brimble,<br />
Interaction <strong>of</strong> oxytocin with Ca2+: I. CD and<br />
fluorescence spectral characterization and<br />
comparison with vasopressin, Biopolymers<br />
40 (1996) 433-43.<br />
11. G. Gimpl, F. Fahrenholz, The oxytocin<br />
receptor system: structure, function, and<br />
regulation, Physiol. Rev. 81 (2001) 629-83.<br />
12. H.V. Hogerzeil, G.J.A. Walker, M.J. De Goeje,<br />
Stability <strong>of</strong> injectable ocytocics in tropical<br />
climates, World Health Organization,<br />
Geneva WHO/DAP/93.6. (1993).<br />
13. A. Hawe, R. Poole, S. Romeijn, P. Kasper,<br />
R. van der Heijden, W. Jiskoot, Towards<br />
heat-stable oxytocin <strong>for</strong>mulations: analysis<br />
<strong>of</strong> degradation kinetics and identification<br />
<strong>of</strong> degradation products, Pharm. Res. 26<br />
(2009) 1679-88.<br />
14. R.A. Poole, P.T. Kasper, W. Jiskoot, Formation<br />
<strong>of</strong> amide- and imide-linked degradation<br />
products between the peptide drug oxytocin<br />
and citrate in citrate-buffered <strong>for</strong>mulations,<br />
J. Pharm. Sci. 100 (2011) 3018-22.<br />
15. C. Avanti, J.P. Amorij, D. Setyaningsih, A.<br />
Hawe, W. Jiskoot, J. Visser, A. Kedrov, A.J.<br />
Driessen, W.L. Hinrichs, H.W. Frijlink, A<br />
new strategy to stabilize oxytocin in aqueous<br />
solutions: I. The effects <strong>of</strong> divalent metal ions<br />
and citrate buffer, AAPS J. 13 (2011) 284-90.<br />
16. The United States <strong>Pharma</strong>copeial<br />
Convention,Rockville, USP-29 NF-24, MD,<br />
2005. (2006).<br />
17. G.S. Shaw, Synthetic calcium-binding<br />
peptides, Methods Mol Biol 173 (2002) 175-<br />
82.<br />
18. C.W. Huck, V. Pezzei, T. Schmitz, G.K.<br />
Bonn, A. Bernkop-Schnurch, Oral peptide<br />
delivery: are there remarkable effects on<br />
drugs through sulfhydryl conjugation?, J<br />
Drug Target 14 (2006) 117-25.<br />
19. M.C. Manning, D.K. Chou, B.M. Murphy,<br />
R.W. Payne, D.S. Katayama, Stability <strong>of</strong><br />
protein pharmaceuticals: an update, Pharm.<br />
Res. 27 (2010) 544-75.<br />
20. G. Bulaj, Formation <strong>of</strong> disulfide bonds<br />
in proteins and peptides, Biotechnology<br />
Advances 23 (2005) 87-92.<br />
21. Joint FAO/WHO Expert Committee on<br />
Food Additives, Citric acid and its calcium,<br />
potassium and sodium salts 539 (1974).<br />
22. T. Wyttenbach, D. Liu, M.T. Bowers,<br />
Interactions <strong>of</strong> the hormone oxytocin with<br />
divalent metal ions, J. Am. Chem. Soc. 130<br />
(2008) 5993-6000.<br />
23. A.F. Pearlmutter, M.S. Sol<strong>of</strong>f, Characterization<br />
<strong>of</strong> the metal ion requirement <strong>for</strong> oxytocinreceptor<br />
interaction in rat mammary gland<br />
membranes, J Biol Chem 254 (1979) 3899-906.<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
65
4<br />
66<br />
supportIng In<strong>for</strong>matIon<br />
Figure S1. Oxytocin MS/MS series<br />
ggg<br />
ggg<br />
Figure S2. Product ion spectrum <strong>of</strong> m/z 1198.8 at a retention time <strong>of</strong> 11.8 min <strong>of</strong> oxytocin<br />
N-citryl amide
ggg<br />
Figure S3. Product ion spectrum <strong>of</strong> m/z 1007.6 at a retention time <strong>of</strong> 12.0 min <strong>of</strong><br />
monodeamidated species <strong>of</strong> oxytocin<br />
Figure S4. Product ion spectrum <strong>of</strong> m/z 1007.6 at a retention time <strong>of</strong> 12.6 min <strong>of</strong> unmodified<br />
oxytocin<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
67
4<br />
68<br />
ggg<br />
Figure S5. Product ion spectrum <strong>of</strong> m/z 1039.6 at a retention time <strong>of</strong> 13.5 min <strong>of</strong> oxytocin<br />
trisulfide<br />
Figure S6. Product ion spectrum <strong>of</strong> m/z 1066.4 at a retention time <strong>of</strong> 14.4 min <strong>of</strong> oxytocin<br />
acetate (ammoniated)
ggg<br />
Figure S7. Product ion spectrum <strong>of</strong> m/z 993.2 at a retention time <strong>of</strong> 15.5 min <strong>of</strong> oxytocin<br />
dimer 1(ammoniated)<br />
Figure S8. Product ion spectrum <strong>of</strong> m/z 1000.4 at a retention time <strong>of</strong> 16.8 min <strong>of</strong> oxytocin<br />
dimer 3<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
69
4<br />
70<br />
ggg<br />
Figure S9. Product ion spectrum <strong>of</strong> m/z 1000.4 at a retention time <strong>of</strong> 17.8 min <strong>of</strong> oxytocin<br />
dimer 5<br />
Figure S10. Product ion spectrum <strong>of</strong> m/z 1000.4 at a retention time <strong>of</strong> 18.7 min <strong>of</strong> oxytocin<br />
dimer 6
ggg<br />
<strong>TI</strong>C <strong>of</strong> +Q1: from Sample 14 (july 3 DTT) <strong>of</strong> Q1.wiff (Turbo Spray) Max. 1.9e7 cps.<br />
Figure S11. Overlay <strong>of</strong> UV-RP-HPLC traces <strong>of</strong> oxytocin in citrate buffer after storage at 70°C<br />
<strong>for</strong> 5 days detected at a wavelength <strong>of</strong> 220 nm be<strong>for</strong>e DTT reduction (light gray) and after<br />
DTT reduction (dark gray).<br />
2.0e7<br />
1.8e7<br />
1.6e7<br />
1.4e7<br />
1.2e7<br />
1.0e7<br />
8.0e6<br />
6.0e6<br />
4.0e6<br />
2.0e6<br />
0.0<br />
no DTT<br />
with DTT<br />
6 8 10 12 14 16 18 20 22<br />
Time, min<br />
Figure S12. Overlay <strong>of</strong> LC-MS <strong>TI</strong>C traces <strong>of</strong> oxytocin in citrate buffer after storage at 70°C<br />
<strong>for</strong> 5 days be<strong>for</strong>e and after DTT reduction<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
71
4<br />
72<br />
ggg<br />
Figure S13. Product ion spectrum after DTT reduction <strong>of</strong> m/z 1007.4 at a retention time <strong>of</strong><br />
12.9 min <strong>of</strong> oxytocin disulfide<br />
Figure S14. Product ion spectrum after DTT reduction <strong>of</strong> m/z 1200.6 at a retention time <strong>of</strong><br />
13.1 min <strong>of</strong> N-citryl oxytocin reduced
ggg<br />
Figure S15. Product ion spectrum after DTT reduction <strong>of</strong> m/z 1009.4 at a retention time <strong>of</strong><br />
13.5 min <strong>of</strong> oxytocin reduced<br />
Figure S16. Product ion spectrum after DTT reduction <strong>of</strong> m/z 1182.6 at a retention time <strong>of</strong><br />
14.7 min <strong>of</strong> N-citryl oxytocin dehydrated (ammoniated)<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
73
4<br />
74<br />
ggg<br />
Figure S17. Product ion spectrum after DTT reduction <strong>of</strong> m/z 1068.8 at a retention time <strong>of</strong><br />
15.6 min <strong>of</strong> Oxytocin acetate ammoniated<br />
Figure S18. Product ion spectrum after DTT reduction <strong>of</strong> m/z 993.6 at a retention time <strong>of</strong><br />
16.6 min <strong>of</strong> β-elimination and deamidation at Asn 5 or Gln 4
ggg<br />
Figure S19. Product ion spectrum after DTT reduction <strong>of</strong> m/z 994.0 at a retention time <strong>of</strong><br />
18.0 min <strong>of</strong> dimer<br />
Figure S20. Product ion spectrum after DTT reduction <strong>of</strong> m/z 993.6 at a retention time <strong>of</strong><br />
18.1 min <strong>of</strong> dimer<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
4<br />
75
4<br />
76<br />
(a) Ca-OT-water (b) Mg-OT-water (c) Zn-OT-water<br />
Figure S21. Calorimetric titration pr<strong>of</strong>iles <strong>of</strong> aliquots <strong>of</strong> 125 mM divalent metal ions:<br />
(a) Ca 2+ , (b) Mg 2+ , and (c) Zn 2+ into 5 mM oxytocin in water pH 4.5. The heat absorbed per<br />
mol <strong>of</strong> titrant is plotted versus the ratio <strong>of</strong> the total concentration <strong>of</strong> divalent metal ions to<br />
the total concentration <strong>of</strong> oxytocin<br />
(a) Ca-citrate-water (b) Mg- citrate-water (c) Zn-citrate-water<br />
Figure S22. Calorimetric titration pr<strong>of</strong>iles <strong>of</strong> aliquots <strong>of</strong> 125 mM divalent metal ions: (a)<br />
Ca 2+ , (b) Mg 2+ , and (c) Zn 2+ into 10 mM citrate buffer pH 4.5. The heat absorbed per mol<br />
<strong>of</strong> titrant is plotted versus the ratio <strong>of</strong> the total concentration <strong>of</strong> divalent metal ions to the<br />
total concentration <strong>of</strong> oxytocin
Table SI. Thermodynamics <strong>of</strong> divalent metal binding to oxytocin as determined by isothermal titration<br />
calorimetry in the absence <strong>of</strong> buffer.<br />
Metal Phase N (sites) K a (M -1 ) ΔH° (cal mol -1 ) ΔS° (cal/mol/deg)<br />
Ca 2+<br />
1<br />
2<br />
1<br />
0.27<br />
30<br />
300<br />
Mg2+ No binding observed<br />
Zn2+ No binding observed<br />
-4000<br />
-3516<br />
The results are depicted as averages <strong>of</strong> three independent measurements with relative standard deviations<br />
below 10%<br />
Table SII. Thermodynamics <strong>of</strong> divalent metal binding to citrate as determined by isothermal titration<br />
calorimetry in the absence <strong>of</strong> oxytocin.<br />
Metal Phase N (sites) Ka (M -1 ) ΔH° (cal mol -1 ) ΔS° (cal/mol/deg)<br />
Ca 2+<br />
1<br />
2<br />
0.3<br />
1.1<br />
800<br />
38<br />
5700<br />
-1800<br />
Mg2+ 1 0.2 300 5800 29<br />
Zn2+ 1 0.42 540 1600 18<br />
The results are depicted as averages <strong>of</strong> three independent measurements with relative standard deviations<br />
below 10%<br />
SuppreSSIon <strong>of</strong> cySteIne-MedIated InterMolecular reactIonS<br />
-8<br />
1<br />
15<br />
1.5<br />
4<br />
77
Christina Avanti 1 , Wouter L.J. Hinrichs 1 , Angela Casini 2 , Anko C. Eissens 1 ,<br />
Annie van Dam 3 , Alexej Kedrov 4 Arnold J. M. Driessen 4 , Henderik W. Frijlink 1 ,<br />
and Hjalmar P. Permentier 3<br />
1 Department <strong>of</strong> <strong>Pharma</strong>ceutical Technology & Biopharmacy,<br />
2 <strong>Pharma</strong>cokinetics, Toxicology and Targeting, Research Institute <strong>of</strong> <strong>Pharma</strong>cy,<br />
3 Mass Spectrometry Core Facility and<br />
4 Department <strong>of</strong> Molecular Microbiology,<br />
University <strong>of</strong> Groningen, The Netherlands
InsIght Into the stabIlIty <strong>of</strong> the<br />
zInc-aspartate-oxytocIn <strong>for</strong>mulatIon<br />
5
5<br />
abstract<br />
The stability <strong>of</strong> the peptide hormone oxytocin in pharmaceutical <strong>for</strong>mulations during<br />
prolonged storage at high temperatures is strongly dependent on the solution composition.<br />
The aim <strong>of</strong> this study was to investigate the effect <strong>of</strong> divalent metal ions (Ca 2+ , Mg 2+<br />
and Zn 2+ ) on the stability <strong>of</strong> oxytocin in aspartate buffer (pH 4.5) and determine their<br />
interaction with the peptide in aqueous solution. Reversed phase and size exclusion-high<br />
per<strong>for</strong>mance liquid chromatography (RP-HPLC and HP-SEC) measurements indicated<br />
that after 4 weeks <strong>of</strong> storage at 55°C all tested divalent metal ions improved the stability <strong>of</strong><br />
oxytocin in aspartate buffered solutions (pH 4.5). However, the stabilizing effects <strong>of</strong> Zn 2+<br />
were by far superior compared to Ca 2+ and Mg 2+ . Liquid chromatography-tandem mass<br />
spectrometry (LC-MS/MS) showed that the combination <strong>of</strong> aspartate and Zn 2+ in particular<br />
suppressed the <strong>for</strong>mation <strong>of</strong> peptide dimers. As shown by isothermal titration calorimetry,<br />
Zn 2+ interacted with oxytocin in the presence <strong>of</strong> aspartate buffer while Ca 2+ or Mg 2+ did<br />
not. In conclusion, the stability <strong>of</strong> oxytocin in the aspartate buffered-solution is strongly<br />
improved in the presence <strong>of</strong> zinc ions, and the stabilization effect is correlated with the<br />
ability <strong>of</strong> the divalent metal ions in aspartate buffer to interact with oxytocin. The reported<br />
results are discussed in relation to the possible mode <strong>of</strong> interactions between the peptide,<br />
zinc and buffer components leading to the observed stabilization effects.
1. IntroductIon<br />
Pregnant women may face life-threatening blood loss at the time <strong>of</strong> delivery. As stated in<br />
the ICM-FIGO joint statement, the drug <strong>of</strong> choice to prevent bleeding after child-birth<br />
(post-partum hemorrhage) is oxytocin [1]. Oxytocin (Figure 1) is a nonapeptide hormone<br />
that is composed <strong>of</strong> a cyclic sequence <strong>of</strong> Cys 1 -Tyr 2 -Ile 3 -Gln 4 - Asn 5 -Cys 6 with an N-terminal<br />
amino group, and <strong>of</strong> a linear Pro 7 -Leu 8 -Gly 9 with a C-terminal amide [2].<br />
Un<strong>for</strong>tunately, a major problem in practice is that injectable oxytocin <strong>for</strong>mulations are<br />
highly unstable as the storage temperature rises to 30°C or higher [3]. There<strong>for</strong>e, oxytocin<br />
should be stored and transported refrigerated, the so-called cold chain, which is not<br />
always guaranteed especially in rural and tropical areas [4]. At a molecular level oxytocin<br />
can undergo degradation via deamidation, oxidation or thiol exchange. In particular, the<br />
stability <strong>of</strong> the Cys 1 -Cys 6 disulfide bridge with respect to thiol exchange, or to oxidation due<br />
to the presence <strong>of</strong> oxygen, light and/or metal ions, is crucial to avoid oxytocin degradation<br />
and progressive aggregation [5].<br />
Several studies have been conducted to improve the stability <strong>of</strong> oxytocin<br />
<strong>for</strong>mulations [6-8]. Within this frame, we have previously demonstrated that the use <strong>of</strong><br />
combinations <strong>of</strong> divalent metal ions with citrate buffer greatly improve the stability <strong>of</strong><br />
oxytocin in aqueous solution, while divalent metal ions added to non-buffered aqueous<br />
oxytocin <strong>for</strong>mulation have only limited stabilizing effects at 40 or 55°C [7]. In a consecutive<br />
study, we have shown that <strong>for</strong>mation <strong>of</strong> a complex <strong>of</strong> divalent metal ions and citrate with<br />
oxytocin leads to the suppression <strong>of</strong> cysteine-mediated inter- or intramolecular reactions,<br />
thus suppressing tri/tetrasulfide and dimer <strong>for</strong>mation [8].<br />
We have also observed that different type <strong>of</strong> buffers may lead to a different interaction between<br />
divalent metal ions and oxytocin, there<strong>for</strong>e having a different impact on oxytocin stability. For<br />
example, we demonstrated that addition <strong>of</strong> divalent metal ions to oxytocin solutions in either<br />
acetic or citric acid buffers results in different stabilization effects <strong>of</strong> the peptide in aqueous<br />
solution [7]. While Ca 2+ , Mg 2+ and Zn 2+ ions in combination with citrate buffer were successful in<br />
Figure 1. Oxytocin structure.<br />
StabIlIty <strong>of</strong> the dIvalent Metal -aSpartate-oxytocIn coMplex<br />
5<br />
81
5<br />
82<br />
stabilizing oxytocin, no stabilizing effect was observed <strong>for</strong> the combination <strong>of</strong> the same divalent<br />
metal ions and acetate buffer [7]. Acetic acid and citric acid are carboxylic acids with one and<br />
three carboxylate groups, respectively. Thus, the above mentioned results suggested that one<br />
carboxylate group is not sufficient <strong>for</strong> stabilizing the oxytocin-metal-buffer salt cluster.<br />
The aim <strong>of</strong> the present study is to investigate the effects <strong>of</strong> divalent metal ions (Ca 2+ ,<br />
Mg 2+ and Zn 2+ ) on the stability <strong>of</strong> oxytocin in aspartate buffer. Aspartate is a commonly<br />
used buffer in parenteral products approved by the FDA <strong>for</strong> <strong>for</strong>mulation purposes [9]. It<br />
has one amine and two carboxylate groups, which may establish different interactions and<br />
affect differently the peptide stability with respect to citric acid.<br />
It should be noted that both the pH and concentration <strong>of</strong> metal ions in the <strong>for</strong>mulation<br />
were found to be crucial in the effectiveness <strong>of</strong> oxytocin stabilization in citrate buffer [7].<br />
Thus, since the optimum stability <strong>of</strong> oxytocin was reported to be at pH 4.5 [6], the peptide<br />
<strong>for</strong>mulations were maintained at that pH, while various divalent metal ion concentrations<br />
were tested. Afterwards, we have also investigated the effect <strong>of</strong> divalent metal ions on the<br />
degradation pr<strong>of</strong>ile <strong>of</strong> oxytocin in aspartate buffer by liquid chromatography-tandem mass<br />
spectrometry (LC-MS/MS), as well as the thermodynamics <strong>of</strong> the oxytocin-metal-buffer<br />
system by isothermal titration calorimetry (ITC).<br />
2. materIals and methods<br />
2.1. Materials<br />
Oxytocin monoacetate powder (Diosynth. Oss, The Netherlands) was kindly provided by MSD,<br />
Oss, The Netherlands. L-aspartic acid, magnesium chloride, and zinc chloride were purchased<br />
from Fluka, Steinheim, Germany. Calcium chloride was from Riedel-de Haen, Seelze, Germany,<br />
and sodium hydroxide, sodium dihydrogen phosphate dihydrate, acetonitrile (supergradient<br />
grade), as well as <strong>for</strong>mic acid were purchased from Merck, Darmstadt, Germany.<br />
2.2. Methods<br />
2.2.1. Formulation <strong>of</strong> oxytocin solution and stability study<br />
Oxytocin solution was <strong>for</strong>mulated at a concentration <strong>of</strong> 0.1 mg/mL (0.094 mM) in 10 mM<br />
aspartate buffer at pH 4.5 (pH adjusted with sodium hydroxide) with different concentrations<br />
<strong>of</strong> divalent metal ions Ca2+ , Mg2+ and Zn2+ . All divalent metal ion solutions were prepared<br />
using their chloride salts at concentrations <strong>of</strong> 2, 5, 10 and 50 mM. The concentration <strong>of</strong><br />
oxytocin was determined using a UV spectrometer as described previously [7,10]. After<br />
preparation, the solutions were stored in 6R glass type 1 vials <strong>for</strong> 4 weeks at either 4 or<br />
55°C, and protected from light. Some selected <strong>for</strong>mulations were also stored <strong>for</strong> 5 days at<br />
70°C and the samples were diluted 10-fold with water <strong>for</strong> LC-MS/MS analysis. During the<br />
stability study controlled pH levels in the samples were within 0.1 pH units.<br />
It must be noted that oxytocin is commonly <strong>for</strong>mulated at very low concentration, which<br />
is about 1/10 <strong>of</strong> the concentration within this study (RP-HPLC and HP-SEC). The higher<br />
concentration was chosen to have better intensity <strong>of</strong> signal <strong>for</strong> the degradation products<br />
produced by the heat stress.<br />
2.2.2. Reversed-Phase High-Per<strong>for</strong>mance Liquid Chromatography (RP-HPLC)<br />
RP-HPLC was carried out according to the procedure described earlier [6,7]. An Alltima<br />
C-18 RP column with 5 μm particle size, inner diameter <strong>of</strong> 4.6 mm, and length <strong>of</strong> 150 mm
(Alltech, Ridderkerk, Netherlands), a Waters (Millipore) 680 Automated Gradient<br />
Controller, two Waters 510 HPLC pumps, a Waters 717 Plus Autosampler, and a Waters<br />
486 Tunable Absorbance UV Detector were used. Samples <strong>of</strong> 20 μL were injected and<br />
the separation was carried out at a flow rate <strong>of</strong> 1.0 mL/min and UV detection at 220 nm.<br />
Samples were eluted using 15% (v/v) acetonitrile in 65 mM phosphate buffer, pH 5.0 as<br />
solvent A and 60% (v/v) acetonitrile in 65 mM phosphate buffer, pH 5.0 as solvent B. The<br />
acetonitrile concentration was linearly increased from 15% at the beginning, to 20% at 10<br />
min, to 30% at 20 min and finally to 60% at 25 min. The recovery <strong>of</strong> oxytocin is expressed<br />
as the percentage <strong>of</strong> initial amount.<br />
2.2.3. Size Exclusion HPLC (HP-SEC)<br />
HP-SEC was per<strong>for</strong>med according to the method previously reported [6,7]. A Superdex<br />
Peptide 10/300 GL column (GE Healthcare Inc., Brussels, Belgium) was used on an<br />
isocratic HPLC system with a Waters 510 pump, a Waters 717 plus auto sampler, a Waters<br />
474 Scanning Fluorescence Detector and Waters 484 Tunable Absorbance Detector<br />
(Waters, Mil<strong>for</strong>d Massachusetts, USA). Samples <strong>of</strong> 50 μL were injected, and separation was<br />
per<strong>for</strong>med at a flow rate <strong>of</strong> 1 mL/min. Chromatograms were obtained using fluorescence<br />
detection at excitation wavelength <strong>of</strong> 274 nm and emission wavelength <strong>of</strong> 310 nm. The<br />
mobile phase consisted <strong>of</strong> 30% acetonitrile and 70% 0.04 M <strong>for</strong>mic acid. The recovery <strong>of</strong><br />
monomeric oxytocin is expressed as the percentage <strong>of</strong> initial amount.<br />
2.2.4. Liquid Chromatography-Mass Spectrometry/Mass Spectrometry<br />
(LC-MS/ MS)<br />
The LC-MS/MS system was set up according to the method described earlier [6,8]. A<br />
Shimadzu LC system equipped with LC-20AD gradient pumps and a SIL-20AC autosampler<br />
was used. The chromatographic separation was carried out on an Alltima C18 column<br />
(internal diameter 2.1 mm, length 150 mm, particle size 5 µm, Grace Davison Discovery<br />
Sciences). The gradient mobile phase composition was a mixture <strong>of</strong> solvent A consisting <strong>of</strong><br />
95% water / 5% acetonitrile and solvent B consisting <strong>of</strong> 5% water / 95% acetonitrile, both<br />
containing 0.05% (v/v) acetic acid and 10 mM ammonium acetate. Elution was per<strong>for</strong>med<br />
by a linear gradient from 5 to 60% <strong>of</strong> the solvent B in 30 min, followed by an increase to<br />
90% solvent B in 1 min, where it was kept 4 min, after which it returned to the starting<br />
conditions. The flow rate was 0.2 mL/min. The UV signal was recorded at 220 nm. The<br />
injection volume was 50 μL.<br />
The HPLC system was coupled to an API 3000 triple-quadrupole mass spectrometer<br />
(Applied Biosystems/MDS Sciex) via a Turbo Ion Spray source. The ionization was<br />
per<strong>for</strong>med by electrospray in the positive mode. Full scan spectra were recorded at a scan<br />
rate <strong>of</strong> 2 s from m/z 500 to 1300 and a step size <strong>of</strong> 0.2 amu with a Declustering Potential<br />
(DP) <strong>of</strong> 40 V and a Focusing Potential (FP) <strong>of</strong> 250 V. Product ion scans were acquired in<br />
specified time windows with a DP <strong>of</strong> 60 V, a FP <strong>of</strong> 300 V, a Collision Energy <strong>of</strong> 40 V and a<br />
Collision Cell Exit Potential <strong>of</strong> 20 V. Data acquisition and processing was per<strong>for</strong>med using<br />
Analyst version 1.5 s<strong>of</strong>tware (Applied Biosystems/MDS Sciex).<br />
LC-MS/MS was used to identify the oxytocin degradation products observed by<br />
RP-HPLC with UV detection. The analyses were carried out on each <strong>of</strong> the major degradation<br />
peaks <strong>for</strong> oxytocin in aspartate buffer <strong>for</strong>mulation without divalent metal salts as well as by<br />
LC-MS <strong>for</strong> the <strong>for</strong>mulation with divalent metal salts.<br />
StabIlIty <strong>of</strong> the dIvalent Metal -aSpartate-oxytocIn coMplex<br />
5<br />
83
5<br />
84<br />
2.2.5. Isothermal Titration Calorimetry (ITC)<br />
Microcalorimetric titrations <strong>of</strong> divalent metal ions to oxytocin in aspartate buffer were<br />
per<strong>for</strong>med by using a MicroCal ITC 200 microcalorimeter (Northampton, MA 01060 USA)<br />
as described previously [7]. A solution <strong>of</strong> 30 µL <strong>of</strong> 125 mM divalent metal chloride (calcium,<br />
magnesium, or zinc chloride) in 10 mM aspartate buffer, pH 4.5 was placed in the syringe, while<br />
300 µL <strong>of</strong> 5 mM oxytocin in 10 mM <strong>of</strong> aspartate buffer, pH 4.5 was placed in the sample cell.<br />
The reference cell contained 300 µL <strong>of</strong> aspartate buffer. Experiments were per<strong>for</strong>med at 55°C.<br />
The effective heat <strong>of</strong> the peptide-metal ion interaction upon each titration step was corrected<br />
<strong>for</strong> dilution and mixing effects, as measured by titrating the divalent metal ion solution into<br />
the buffer and by titrating the buffer into the oxytocin solution (reference measurement). To<br />
investigate the possibility <strong>of</strong> oxytocin or metal ion binding to the buffer components, control<br />
experiments were per<strong>for</strong>med in water. The heats <strong>of</strong> bimolecular interactions were obtained by<br />
integrating the peak following each injection. All measurements were per<strong>for</strong>med in triplicate.<br />
ITC data were analyzed by using the ITC non-linear curve fitting functions <strong>for</strong> one or two<br />
binding sites from Origin 7.0 s<strong>of</strong>tware (MicroCal S<strong>of</strong>tware, Inc.).<br />
3. results<br />
3.1. The effect <strong>of</strong> divalent metal ions on oxytocin stability in<br />
aspartate buffer solution<br />
To investigate the stability <strong>of</strong> oxytocin in the aspartate buffered solutions in the presences <strong>of</strong><br />
divalent metal ions, oxytocin <strong>for</strong>mulation in 10 mM aspartate buffer (pH 4.5) with various<br />
concentrations <strong>of</strong> divalent metal ions were prepared. Oxytocin is commonly <strong>for</strong>mulated at<br />
a concentration <strong>of</strong> about 0.01 mg/mL, which is 10 times lower than the concentration used<br />
in this study. The higher concentration was chosen to have better intensity <strong>of</strong> degradation<br />
product produced by the heat stress in RP-HPLC and HP-SEC analysis. Oxytocin recovery<br />
and the presence <strong>of</strong> oxytocin monomer were determined <strong>for</strong> all the samples after 4 weeks at<br />
different storage temperatures by RP-HPLC and HP-SEC, respectively, as described in the<br />
experimental section. The obtained results are shown in Figure 2.<br />
After 4 weeks <strong>of</strong> storage at 4°C, oxytocin remained stable in all <strong>for</strong>mulations. Instead,<br />
after 4 weeks <strong>of</strong> storage at 55°C oxytocin stability increased with increasing divalent metal<br />
ions concentration. Ca 2+ and Mg 2+ had similar effects on improving the oxytocin stability:<br />
the recoveries <strong>of</strong> oxytocin, as well as the remaining percentage <strong>of</strong> oxytocin monomer were<br />
increased up to 45% in the presence <strong>of</strong> 50 mM Mg 2+ , and up to 35% in the presence <strong>of</strong> 50<br />
mM Ca 2+ . Notably, Zn 2+ was much more effective in stabilizing oxytocin: at a concentration<br />
<strong>of</strong> 2 mM Zn 2+ increased oxytocin recovery up to 35% (i.e. to a comparable extent as 50<br />
mM Ca 2+ or Mg 2+ ), and almost ca. 75% <strong>of</strong> oxytocin monomer was recovered. Overall, both<br />
RP-HPLC and HP-SEC showed that Zn 2+ has superior stabilizing effects in aspartate buffer<br />
compared to Ca 2+ and Mg 2+<br />
3.2. The effect <strong>of</strong> divalent metal ions on the degradation pr<strong>of</strong>ile <strong>of</strong><br />
oxytocin in aspartate buffer<br />
To assign the degradation products <strong>of</strong> oxytocin in the <strong>for</strong>mulations, oxytocin in10 mM<br />
aspartate buffer (pH 4.5) in the absence and presence <strong>of</strong> 10 mM divalent metal ions was<br />
analyzed by LC-MS/MS after incubation <strong>of</strong> the samples at 70°C <strong>for</strong> 5 days to ensure the
.<br />
Figure 2 Effect <strong>of</strong> Ca 2+ (squares), Mg 2+ (circles), and Zn 2+ Figure 2. Effect <strong>of</strong> Ca<br />
(triangles) concentration on the<br />
recovery <strong>of</strong> oxytocin in 10 mM aspartate buffer, pH 4.5, after 4 weeks <strong>of</strong> storage at either 4°C<br />
(solid symbols) or 55°C (open symbols). A: Oxytocin recovery as determined by RP-HPLC.<br />
B: Oxytocin monomer remaining as determined by HP-SEC. The results are depicted as<br />
averages <strong>of</strong> three independent measurements ± SD<br />
2+ (squares), Mg2+ (circles), and Zn2+ (triangles) concentration on<br />
the recovery <strong>of</strong> oxytocin in 10 mM aspartate buffer, pH 4.5, after 4 weeks <strong>of</strong> storage at<br />
either 4°C (solid symbols) or 55°C (open symbols). A: Oxytocin recovery as determined<br />
by RP-HPLC. B: Oxytocin monomer remaining as determined by HP-SEC. The results are<br />
depicted as averages <strong>of</strong> three independent measurements ± SD<br />
<strong>for</strong>mation <strong>of</strong> sufficient degradation products intensity. MS/MS analysis was applied to<br />
confirm the identification <strong>of</strong> the most intense peaks, and degradation products were assigned<br />
following our previous reports [6,8,11]. Figure 3A shows the total ion-current (<strong>TI</strong>C) <strong>of</strong> mass<br />
spectra in the m/z range between 900 and 1200 <strong>for</strong> <strong>for</strong>mulation <strong>of</strong> oxytocin without divalent<br />
metal ions (solid line), and with 10 mM Zn2+ After 4 weeks <strong>of</strong> storage at 4°C, oxytocin remained stable in all <strong>for</strong>mulations. Instead,<br />
after 4 weeks <strong>of</strong> storage at 55°C oxytocin stability increased with increasing divalent<br />
metal ions concentration. Ca (dashed line). Spectra <strong>of</strong> <strong>for</strong>mulations with<br />
calcium and magnesium ions are reported in the supplementary material available.<br />
Nine peaks are observed in the MS pr<strong>of</strong>ile, which revealed 10 molecular species in the<br />
LC-MS contour plot (Figure 3B). The protonated molecular ion <strong>of</strong> unmodified oxytocin was<br />
found at a retention time <strong>of</strong> 12.8 min with m/z <strong>of</strong> 1007.4. A peak at a retention time <strong>of</strong> 14.5<br />
min with m/z 1066.6 corresponded to oxytocin acetate (ammoniated), but also appeared in<br />
the LC-MS <strong>of</strong> the unstressed oxytocin preparation. Most likely, oxytocin acetate was <strong>for</strong>med<br />
during synthesis as the oxytocin is supplied as a monoacetate salt. Complete assignment <strong>of</strong><br />
the degradation products is listed in Table 1. At retention time <strong>of</strong> 13.7 and 15.0 min tri- and 80<br />
tetrasulfide <strong>for</strong>m degradation products were identified from the m/z 1039.6 and 1071.6<br />
peaks, respectively (see supplementary material <strong>for</strong> the structures <strong>of</strong> tri- and tetrasulfide<br />
oxytocin species). Other peaks (Dimers 1 to 6 in Table 1) are interpreted as various <strong>for</strong>ms <strong>of</strong><br />
disulfide or thioether-linked oxytocin dimers as described previously [8].<br />
2+ and Mg 2+ had similar effects on improving the<br />
oxytocin stability: the recoveries <strong>of</strong> oxytocin, as well as the remaining percentage <strong>of</strong><br />
oxytocin monomer were increased up to 45% in the presence <strong>of</strong> 50 mM Mg 2+ , and up<br />
to 35% in the presence <strong>of</strong> 50 mM Ca 2+ . Notably, Zn 2+ was much more effective in<br />
stabilizing oxytocin: at a concentration <strong>of</strong> 2 mM Zn 2+ increased oxytocin recovery<br />
up to 35% (i.e. to a comparable extent as 50 mM Ca 2+ or Mg 2+ ), and almost ca. 75%<br />
<strong>of</strong> oxytocin monomer was recovered. Overall, both RP-HPLC and HP-SEC showed<br />
StabIlIty <strong>of</strong> the dIvalent Metal -aSpartate-oxytocIn coMplex<br />
5<br />
85
5<br />
86<br />
Figure 3. A: Total ion chromatogram (m/z 900-1200) and B: Contour plot <strong>of</strong> oxytocin and<br />
its degradation products in 10 mM aspartate buffer after 5 days <strong>of</strong> storage at 70°C and pH<br />
4.5 (solid-line) and in the presence <strong>of</strong> 10 mM Zn 2+ (dashed-line).<br />
Addition <strong>of</strong> divalent metal salts in the <strong>for</strong>mulation with the aspartate buffer did not<br />
result in any additional degradation products compared to samples without metal salts (see<br />
dashed line in Figure 3A <strong>for</strong> zinc salt addition). Conversely, as expected from the RP-HPLC<br />
and HP-SEC data, it resulted in a significant reduction <strong>of</strong> the major degradation peaks.<br />
The intensity <strong>of</strong> each degradation product from each stressed <strong>for</strong>mulation <strong>of</strong> oxytocin in<br />
aspartate buffer in combination with Ca 2+ , Mg 2+ , and Zn 2+ was also recorded. Table 1 reports<br />
the relative intensities in percentage <strong>of</strong> the peak area <strong>of</strong> each degradation product in the<br />
presence <strong>of</strong> divalent metal ions with respect to the same <strong>for</strong>mulations without divalent<br />
metal ions.Addition <strong>of</strong> Ca 2+ and Mg 2+ reduced the <strong>for</strong>mation <strong>of</strong> dimer 1 and 2 <strong>of</strong> about 30%,<br />
an effect that was even more marked in the presence <strong>of</strong> Zn 2+ : i.e., 53 and 60% reduction <strong>for</strong><br />
dimer 1 and 2, respectively.<br />
Zinc is the most efficient element in suppressing the total amount <strong>of</strong> the degradation<br />
products <strong>of</strong> oxytocin in aspartate buffer. In fact, at variance with Ca 2+ and Mg 2+ , Zn 2+ also<br />
reduced the <strong>for</strong>mation <strong>of</strong> other dimers: dimer 3 was reduced by 14% and dimer 5 was<br />
reduced by 30%. However, an increase in the <strong>for</strong>mation <strong>of</strong> tri and tetrasulfide species was
Table 1. Effect <strong>of</strong> divalent metal ions on the intensity <strong>of</strong> oxytocin in aspartate buffered solution. Relative<br />
peak intensity is expressed with respect to that <strong>of</strong> the same product in the absence <strong>of</strong> divalent metal ions.<br />
Assignment<br />
retention<br />
time (min) m/z<br />
Relative peak intensity <strong>of</strong> oxytocin a (%)<br />
Ca 2+ Mg 2+ Zn 2+<br />
oxytocin trisulfide 13.7 1039.4 111 ± 4 101 ± 8 146 ± 5<br />
Oxytocin tetrasulfide 15.0 1071.6 135 ± 13 114 ± 10 187 ± 12<br />
dimer 1 15.7 993.2b 68 ± 3 68 ± 2 47 ± 0<br />
dimer 2 16.7 993.2b 69 ± 3 76 ± 2 39 ± 2<br />
dimer 3 16.9 1008.6 115 ± 4 111 ± 2 86 ± 2<br />
dimer 4 17.6 1024.6b 147 ± 19 135 ± 3 127 ± 5<br />
dimer 5 18.0 1008.6 111 ± 2 111 ± 1 71 ± 3<br />
dimer 6 18.9 1024.6b 152 ± 8 144 ± 3 115 ± 18<br />
a Aspartate buffer solution with 10 mM <strong>of</strong> the indicated divalent metal ion<br />
b ammoniated.<br />
Figure 4. Recovery <strong>of</strong> oxytocin in the absence (OAP) and presence <strong>of</strong> 10 mM Ca 2+ (OAPCa),<br />
Mg 2+ (OAPMg) and Zn 2+ (OAPZn) in 10 mM aspartate-buffered solution at pH 4.5 under<br />
stressed condition at a temperature <strong>of</strong> 70 °C <strong>for</strong> 5 days. Oxytocin recovery was determined<br />
by LC-MS. The results are depicted as averages <strong>of</strong> three independent measurements ± SD<br />
also observed. The signal intensity <strong>of</strong> the tetrasulfide <strong>for</strong>m (m/z 1071.6, retention time<br />
15.0 min) is very low compared to that <strong>of</strong> the dimers, there<strong>for</strong>e peak integration is less<br />
accurate. We have no explanation <strong>for</strong> the increased intensity <strong>of</strong> the trisulfide <strong>for</strong>m (m/z<br />
1039.4 retention time 13.7 min) in the presence <strong>of</strong> Zn 2+ .<br />
Oxytocin <strong>for</strong>mulated in aspartate buffer after 5 days at 70°C without divalent metal ions<br />
was the most degraded <strong>of</strong> all <strong>for</strong>mulations tested, and only approximately 20% oxytocin<br />
could be recovered (Figure 4). Interestingly, addition <strong>of</strong> 10 mM Ca 2+ and Mg 2+ did not<br />
markedly improve the peptide recovery. However, when Zn 2+ was added, the degradation<br />
was reduced and a recovery <strong>of</strong> about 45% oxytocin was found.<br />
StabIlIty <strong>of</strong> the dIvalent Metal -aSpartate-oxytocIn coMplex<br />
5<br />
87
3.3. Interaction <strong>of</strong> oxytocin and divalent metal ions in aspartate<br />
buffer<br />
The interaction <strong>of</strong> oxytocin with divalent metal ions was investigated by ITC at a temperature<br />
<strong>of</strong> 55°C. Figure 5 shows the exothermic (DH obs ) events when Ca 2+ or Mg 2+ was titrated into<br />
a solution <strong>of</strong> oxytocin in 10 mM aspartate buffer. However, the effects were very small (not<br />
more than 0.05 kcal/mole <strong>of</strong> injectant) and only slightly different from the corresponding<br />
reference measurements. When Zn2+ was titrated into oxytocin in aspartate buffer, a strong<br />
endothermic binding reaction was observed and the heat effects upon the titration reached<br />
more than 0.5 kcal/mole <strong>for</strong> the first injection. The shape <strong>of</strong> the titration curve <strong>for</strong> the<br />
Zn2+ -oxytocin in aspartate is indicative <strong>for</strong> a binding reaction. Using the analysis model<br />
5<br />
with two distinct types <strong>of</strong> binding sites, the binding constants K at 55°C <strong>for</strong> Zn a 2+ -oxytocin<br />
interaction are 6.2 ± 1.0 × 103 M-1 and 72 ± 17 M-1 Mg<br />
, respectively. Thus, the only system that<br />
shows thermodynamics interactions has the most pronounced effect on oxytocin stability<br />
in <strong>for</strong>mulations.<br />
2+ was titrated into a solution <strong>of</strong> oxytocin in 10 mM aspartate buffer. However, the<br />
effects were very small (not more than 0.05 kcal/mole <strong>of</strong> injectant) and only slightly<br />
different from the corresponding reference measurements. When Zn 2+ was titrated<br />
into oxytocin in aspartate buffer, a strong endothermic binding reaction was observed<br />
and the heat effects upon the titration reached more than 0.5 kcal/mole <strong>for</strong> the first<br />
injection. The shape <strong>of</strong> the titration curve <strong>for</strong> the Zn 2+ -oxytocin in aspartate is<br />
indicative <strong>for</strong> a binding reaction. Using the analysis model with two distinct types <strong>of</strong><br />
binding sites, the binding constants Ka at 55°C <strong>for</strong> Zn 2+ -oxytocin interaction are 6.2 ±<br />
4. dIscussIon<br />
1.0 × 10<br />
In a previous study, we have found that oxytocin can be stabilized by a combination <strong>of</strong> divalent<br />
metal ions and citrate buffer at pH 4.5. Divalent metal ions in combination with citrate<br />
buffer suppressed intermolecular reactions in the ring structure <strong>of</strong> oxytocin presumably by<br />
3 M -1 and 72 ± 17 M -1 , respectively. Thus, the only system that shows<br />
thermodynamics interactions has the most pronounced effect on oxytocin stability in<br />
<strong>for</strong>mulations.<br />
Figure 5 Least squares fit <strong>of</strong> the data from calorimetric titration pr<strong>of</strong>iles <strong>of</strong> aliquots <strong>of</strong> 125<br />
mM divalent metal ions: Ca 2+ (solid square), Mg 2+ (open square), and Zn 2+ Figure 5. Least squares fit <strong>of</strong> the data from calorimetric titration pr<strong>of</strong>iles <strong>of</strong> aliquots <strong>of</strong> 125<br />
mM divalent metal ions: Ca<br />
(open triangle)<br />
into 5 mM oxytocin in 10 mM aspartate buffer pH 4.5. The heat absorbed per mole <strong>of</strong> titrant<br />
is plotted versus the ratio <strong>of</strong> the total concentration <strong>of</strong> divalent metal ions to the total<br />
concentration <strong>of</strong> oxytocin.<br />
2+ (solid square), Mg2+ (open square), and Zn2+ (open triangle)<br />
into 5 mM oxytocin in 10 mM aspartate buffer pH 4.5. The heat absorbed per mole <strong>of</strong><br />
titrant is plotted versus the ratio <strong>of</strong> the total concentration <strong>of</strong> divalent metal ions to the total<br />
concentration <strong>of</strong> oxytocin.<br />
88<br />
4 Discussion
<strong>for</strong>ming a complex in the disulfide bridge region where the degradation reaction occurred.<br />
There were no significant differences observed among the three tested divalent metal ions,<br />
Ca 2+ , Mg 2+ , and Zn 2+ [7]. Our present study indicates that <strong>for</strong>mulations with a combination<br />
<strong>of</strong> divalent metal ions and aspartate buffer behave differently in improving oxytocin<br />
stability in aqueous solution. In fact, we have shown that in aspartate buffer, only addition<br />
<strong>of</strong> Zn 2+ results in a comparable stabilizing effect on oxytocin as citrate with Ca 2+ , Mg 2+ and<br />
Zn 2+ [7]. The LC-MS/MS results show that <strong>for</strong>mation <strong>of</strong> oxytocin dimers is hampered by the<br />
presence <strong>of</strong> Zn 2+ in aspartate, while it is not so affected by Ca 2+ and Mg 2+ . A previous study<br />
by us [8] showed that all the dimers are produced from the thiol exchange in the disulfide<br />
bridge, and there are no additional dimers produced from this <strong>for</strong>mulation. Thus, it appears<br />
that the combination <strong>of</strong> Zn 2+ and aspartate was able to protect the disulfide Cys 1,6 bridge on<br />
oxytocin. In line with these results, ITC data demonstrate that only zinc, among the tested<br />
divalent metal ions, is able to strongly interact with oxytocin in the <strong>for</strong>mulation conditions.<br />
The observed stabilization effects might be ascribed to the <strong>for</strong>mation <strong>of</strong> divalent metal<br />
ion adducts to oxytocin. In fact, several studies reported on the ability <strong>of</strong> Ca 2+ , Mg 2+ and<br />
Zn 2+ to <strong>for</strong>m complexes with oxygen atoms from the carbonyl backbone <strong>of</strong> the peptide, but<br />
the strength <strong>of</strong> the interactions is different [12-14]. The fact that the three metals have a<br />
smaller ionic radius than the oxytocin ring [15,16] suggested that Ca 2+ , Mg 2+ and Zn 2+ are<br />
located within the ring <strong>of</strong> oxytocin, and adduct <strong>for</strong>mation, arising from coordination <strong>of</strong> the<br />
metal ions via the backbone carbonyl oxygens <strong>of</strong> oxytocin, induces a compact structure <strong>of</strong><br />
an oxytocin-metal octahedral complex [12,17,18].<br />
Most importantly, the binding affinity <strong>of</strong> oxytocin <strong>for</strong> Mg 2+ has been reported to be lower than<br />
in the case <strong>of</strong> Ca 2+ and far below Zn 2+ [12,17,18]. Indeed, Zn 2+ ions very effectively coordinate<br />
with oxytocin and strongly affect its con<strong>for</strong>mation in physiological environment [18].<br />
The oxytocin stabilizing effect <strong>of</strong> Zn 2+ with respect to Ca 2+ and Mg 2+ might be due to the<br />
higher stability <strong>of</strong> the oxytocin-Zn 2+ adduct in the reported <strong>for</strong>mulation conditions.<br />
In addition, our results indicate that also the type <strong>of</strong> buffer used in the <strong>for</strong>mulations<br />
plays an important role in the peptide stabilization effects. At a pH <strong>of</strong> 4.5, citrate has one<br />
carboxylate ion (–COO - ) in α position and other two carboxylates that can contribute to<br />
the coordination <strong>of</strong> a metal or to binding to oxytocin itself. At the same pH <strong>of</strong> 4.5, aspartate<br />
has only two carboxylate ions that could act as electron donors towards a metal ion or<br />
oxytocin. This difference <strong>of</strong> available electron donor groups between the two buffers might<br />
influence the oxytocin-metal adduct <strong>for</strong>mation, as well as its stabilization. The lack <strong>of</strong> an<br />
additional electron donor oxygen moiety in aspartate is likely to discriminate the binding <strong>of</strong><br />
Ca 2+ or Mg 2+ with respect to Zn 2+ , as suggested by the ITC results, which did not show any<br />
significant interactions between oxytocin and Ca 2+ or Mg 2+ in aspartate buffer.<br />
5. conclusIon<br />
Our study clearly shows that the stability <strong>of</strong> oxytocin in the aspartate buffered-solution is<br />
strongly improved in the presence <strong>of</strong> zinc ions, and the stabilization effect is correlated with<br />
the strength <strong>of</strong> interaction between oxytocin and divalent metal ions. Further studies using<br />
NMR and molecular modeling are initiated to characterize the mechanisms <strong>of</strong> oxytocin<br />
stabilization at a molecular level. We can conclude that Zn 2+ binding to peptide in aspartate<br />
solution suppress intermolecular degradation reactions near the Cys 1,6 disulfide bridge that<br />
is responsible <strong>for</strong> oxytocin degradation.<br />
StabIlIty <strong>of</strong> the dIvalent Metal -aSpartate-oxytocIn coMplex<br />
5<br />
89
acknowledgement<br />
The authors want to thank MSD, Oss, The Netherlands <strong>for</strong> providing oxytocin <strong>for</strong> the study.<br />
This research was per<strong>for</strong>med within the framework <strong>of</strong> the Dutch Top Institute <strong>Pharma</strong><br />
(project number D6–202) and Rosalind Franklin fellowship funding <strong>for</strong> Angela Casini.<br />
references<br />
1. International Confederation <strong>of</strong> Midwives,<br />
International Federation <strong>of</strong> Gynaecology<br />
and Obstetrics. Joint statement management<br />
<strong>of</strong> the third stage <strong>of</strong> labour to prevent<br />
post-partum haemorrhage, Http://www.<br />
Internationalmidwives. org/Whatwedo/<br />
P ro g r a m m e s / P O P P H I / Po s t Pa r t u m -<br />
Haemorrhage/tabid/339/Default. Aspx 2012<br />
(2012).<br />
2. A. Ohno, N. Kawasaki, K. Fukuhara, H.<br />
Okuda, T. Yamaguchi, Complete NMR<br />
analysis <strong>of</strong> oxytocin in phosphate buffer,<br />
Magn. Reson. Chem. 48 (2010) 168-72.<br />
3. J.W. Gard, J.M. Alexander, R.E. Bawdon,<br />
J.T. Albrecht, Oxytocin preparation stability<br />
in several common obstetric intravenous<br />
solutions, Am. J. Obstet. Gynecol. 186<br />
(2002) 496-8.<br />
4. H.V. Hogerzeil, G.J. Walker, Instability <strong>of</strong><br />
(methyl)ergometrine in tropical climates: an<br />
overview, Eur. J. Obstet. Gynecol. Reprod.<br />
Biol. 69 (1996) 25-9.<br />
5. M.C. Manning, D.K. Chou, B.M. Murphy,<br />
R.W. Payne, D.S. Katayama, Stability <strong>of</strong><br />
protein pharmaceuticals: an update, Pharm.<br />
Res. 27 (2010) 544-75.<br />
6. A. Hawe, R. Poole, S. Romeijn, P. Kasper,<br />
R. van der Heijden, W. Jiskoot, Towards<br />
heat-stable oxytocin <strong>for</strong>mulations: analysis<br />
<strong>of</strong> degradation kinetics and identification<br />
<strong>of</strong> degradation products, Pharm. Res. 26<br />
(2009) 1679-88.<br />
7. C. Avanti, J.P. Amorij, D. Setyaningsih, A.<br />
Hawe, W. Jiskoot, J. Visser, A. Kedrov, A.J.<br />
Driessen, W.L. Hinrichs, H.W. Frijlink, A<br />
new strategy to stabilize oxytocin in aqueous<br />
solutions: I. The effects <strong>of</strong> divalent metal ions<br />
and citrate buffer, AAPS J. 13 (2011) 284-90.<br />
8. C. Avanti, H.P. Permentier, A.V. Dam, R.<br />
Poole, W. Jiskoot, H.W. Frijlink, W.L.J.<br />
Hinrichs, A New Strategy To Stabilize<br />
Oxytocin in Aqueous Solutions: II.<br />
Suppression <strong>of</strong> Cysteine-Mediated<br />
Intermolecular Reactions by a Combination<br />
<strong>of</strong> Divalent Metal Ions and Citrate, Mol.<br />
<strong>Pharma</strong>ceutics 9 (3) (2012) 554-62.<br />
9. P. Jurgens, C. Panteliadis, G. Fondalinski,<br />
Total parenteral nutrition <strong>of</strong> premature<br />
infants: metabolic effects <strong>of</strong> an exogenous<br />
supply <strong>of</strong> L-aspartic acid and L-glutamic<br />
acid, Z. Ernahrungswiss. 21 (1982) 225-45.<br />
10. [10] S.C. Gill, P.H. von Hippel, Calculation<br />
<strong>of</strong> protein extinction coefficients from<br />
amino acid sequence data, Anal. Biochem.<br />
182 (1989) 319-26.<br />
11. R.A. Poole, P.T. Kasper, W. Jiskoot, Formation<br />
<strong>of</strong> amide- and imide-linked degradation<br />
products between the peptide drug oxytocin<br />
and citrate in citrate-buffered <strong>for</strong>mulations,<br />
J. Pharm. Sci. 100 (2011) 3018-22.<br />
12. V.S. Ananthanarayanan, K.S. Brimble,<br />
Interaction <strong>of</strong> oxytocin with Ca2+: I. CD and<br />
fluorescence spectral characterization and<br />
comparison with vasopressin, Biopolymers<br />
40 (1996) 433-43.<br />
13. J.P. Glusker, A.K. Katz, C.W. Bock, Metal ions<br />
in biological systems, The Rigaku Journal 16<br />
(1999) 8-19.<br />
14. H. Einspahr, C.E. Bugg, The Geometry<br />
<strong>of</strong> Calcium-Carboxylate Interactions in<br />
Crystalline Complexes, Acta Cryst. B37<br />
(1981) 1044-52.<br />
15. A.K. Katz, J.P. Glusker, S.A. Beebe, C.W. Bock,<br />
Calcium Ion Coordination: A Comparison<br />
with That <strong>of</strong> Beryllium, Magnesium, and<br />
Zinc, J. Am. Chem. Soc. 118 (1996) 5752-63.<br />
16. R.H. Holm, P. Kennepohl, E.I. Solomon,<br />
Structural and Functional Aspects <strong>of</strong> Metal<br />
Sites in Biology, Chem. Rev. 96 (1996) 2239-<br />
314.<br />
17. V.S. Ananthanarayanan, M.P. Belciug,<br />
B.S. Zhorov, Interaction <strong>of</strong> oxytocin with<br />
Ca2+: II. Proton magnetic resonance<br />
and molecular modeling studies <strong>of</strong><br />
con<strong>for</strong>mations <strong>of</strong> the hormone and its Ca2+<br />
complex, Biopolymers 40 (1996) 445-64.<br />
18. D. Liu, A.B. Seuthe, O.T. Ehrler, X. Zhang,<br />
T. Wyttenbach, J.F. Hsu, M.T. Bowers,<br />
Oxytocin-receptor binding: why divalent<br />
metals are essential, J. Am. Chem. Soc. 127<br />
(2005) 2024-5.
5<br />
92<br />
supportIng In<strong>for</strong>matIon<br />
Figure S1. Total ion chromatogram (m/z 900-1200) <strong>of</strong> oxytocin and its degradation<br />
products in 10 mM aspartate buffer after 5 days <strong>of</strong> storage at 70°C and pH 4. in the<br />
presence <strong>of</strong> 10 mM Ca 2+ .<br />
Figure S2. Total ion chromatogram (m/z 900-1200) <strong>of</strong> oxytocin and its degradation<br />
products in 10 mM aspartate buffer after 5 days <strong>of</strong> storage at 70°C and pH 4. in the<br />
presence <strong>of</strong> 10 mM Ca 2+ .
Figure S3. Structure <strong>of</strong> oxytocin trisulfide<br />
Figure S4. Structure <strong>of</strong> oxytocin tetrasulfide<br />
StabIlIty <strong>of</strong> the dIvalent Metal -aSpartate-oxytocIn coMplex<br />
5<br />
93
Christina Avanti 1,* , Nur Alia Oktaviani 2,* , Wouter L.J. Hinrichs 1 ,<br />
Henderik W. Frijlink 1 , and Frans A.A. Mulder 2,3<br />
# Equally contributed first author<br />
1 Department <strong>of</strong> <strong>Pharma</strong>ceutical Technology and Biopharmacy,<br />
2 Groningen Biomolecular Sciences and Biotechnology Institute,<br />
University <strong>of</strong> Groningen, Groningen, The Netherlands<br />
3 Department <strong>of</strong> Chemistry and Interdisciplinary Nanoscience Center iNANO,<br />
University <strong>of</strong> Aarhus, Aarhus C, Denmark
aspartate buffer and dIvalent<br />
metal Ions affect the oxytocIn<br />
con<strong>for</strong>matIon In aqueous solutIon<br />
and protect It from degradatIon<br />
6
6<br />
abstract<br />
Oxytocin is a peptide drug used to induce labor and prevent bleeding after child birth.<br />
Due to its instability, transport and storage <strong>of</strong> oxytocin <strong>for</strong>mulations under tropical<br />
condition is problematic. In a previous study, we described that the stability <strong>of</strong> oxytocin<br />
in aspartate buffered <strong>for</strong>mulation is improved by the addition <strong>of</strong> divalent metal ions. The<br />
stabilizing effect <strong>of</strong> Zn 2+ was by far superior compared to that <strong>of</strong> Mg 2+ . In addition, it was<br />
found that stabilization correlated well with the ability <strong>of</strong> the divalent metal ions to interact<br />
with oxytocin in aspartate buffer. Furthermore, LC-MS (MS) measurements indicated<br />
that the combination <strong>of</strong> aspartate buffer and Zn 2+ in particular suppressed intermolecular<br />
degradation reactions near the Cys 1,6 disulfide bridge. These results lead to the hypothesis<br />
that in aspartate buffer, Zn 2+ changes the con<strong>for</strong>mation <strong>of</strong> oxytocin in such a way that the<br />
Cys 1,6 disulfide bridge is shielded from its environment thereby suppressing intermolecular<br />
reactions involving this region <strong>of</strong> the molecule.<br />
To verify this hypothesis, in this study the con<strong>for</strong>mation <strong>of</strong> oxytocin in aspartate buffer<br />
in the presence <strong>of</strong> Mg 2+ or Zn 2+ , was investigated by 2D NMR spectroscopy, i.e. NOESY,<br />
TOCSY, 1 H- 13 C HSQC and 1 H- 15 N HSQC. Almost all 1 H, 13 C and 15 N resonances could<br />
be assigned using HSQC spectroscopy <strong>of</strong> oxytocin without 13 C or 15 N enrichment. 1 H- 13 C<br />
and 1 H- 15 N HSQC spectra showed that aspartate buffer alone induces a minor change in<br />
oxytocin in D 2 O with the largest chemical shift changes are observed in Cys 1 . Zn 2+ causes<br />
more extensive changes in oxytocin in aqueous solution than Mg 2+ . Our findings suggest<br />
that the carboxylate group <strong>of</strong> aspartate neutralizes the positive charge <strong>of</strong> the N terminus <strong>of</strong><br />
Cys 1 , allowing the interactions with Zn 2+ to become more favorable. These interactions may<br />
explain the protection <strong>of</strong> the disulfide bridge against intermolecular reactions that lead to<br />
dimerization and inactivation.
1. IntroductIon<br />
Oxytocin is a nonapeptide hormone secreted by the posterior lobe <strong>of</strong> the pituitary gland<br />
which is involved in the control <strong>of</strong> labor and bleeding cessation after child birth [1]. The<br />
peptide consists <strong>of</strong> nine amino acids (Cys-Tyr-Ile-Gln-Asn-Cys-Pro-Leu- Gly) and an<br />
amidated C-terminus [2]. Oxytocin is the preferred drug to prevent postpartum hemorrhage<br />
and commonly <strong>for</strong>mulated in aqueous solution <strong>for</strong> parenteral administration [3]. Instability<br />
<strong>of</strong> oxytocin in aqueous solution under severe conditions, particularly in tropical conditions,<br />
presents a significant challenge <strong>for</strong> pharmaceutical scientists [4]. Oxytocin instability in<br />
aqueous solution has been reported in several studies [5,6] and the degradation strongly<br />
depends on the pH <strong>of</strong> the <strong>for</strong>mulation, with the highest stability reported at pH 4.5 [4].<br />
Several studies have been aimed at the improvement <strong>of</strong> the stability <strong>of</strong> oxytocin in aqueous<br />
solution [4,7]. The most recent finding is the use <strong>of</strong> divalent metal ions in combination with<br />
certain buffers that strongly increases the stability <strong>of</strong> oxytocin in aqueous solution [8,9].<br />
In a previous study [9] we found that Zn 2+ in combination with aspartate buffer strongly<br />
stabilizes oxytocin in aqueous solutions, while Ca 2+ and Mg 2+ only have minor effects. The<br />
stabilization occurred as a result <strong>of</strong> complex <strong>for</strong>mation <strong>of</strong> Zn 2+ ions and aspartate with<br />
oxytocin, which, by protecting the Cys 1,6 disulfide bridge, suppressed dimerization. In line<br />
with those results, isothermal titration calorimetry data demonstrate that only Zn 2+ ions,<br />
among the tested divalent metal ions (Zn 2+ , Ca 2+ , and Mg 2+ ) is able to strongly interact with<br />
oxytocin in the <strong>for</strong>mulation conditions [9]. Those results lead to the hypothesis that Zn 2+<br />
induces a con<strong>for</strong>mational change, thereby stabilizing oxytocin in aspartate buffer. Aspartic<br />
acid is one <strong>of</strong> the non-essential amino acids which normally synthesized in the body. It<br />
consists <strong>of</strong> two carboxylate groups with pKa 1 and pKa 2 <strong>of</strong> 2.1 and 3.9, and one amine group<br />
(pKa 3 <strong>of</strong> 9.8). Aspartate is a commonly used buffer in parenteral products approved by<br />
the FDA <strong>for</strong> <strong>for</strong>mulation purposes [10]. To investigate the con<strong>for</strong>mation <strong>of</strong> oxytocin in<br />
aspartate buffer in the presence <strong>of</strong> divalent metal ions (Zn 2+ and Mg 2+ ), two-dimensional<br />
Nuclear Magnetic Resonance (NMR) spectroscopy was used.<br />
Nuclear Magnetic Resonance spectroscopy is a suitable technique to study the<br />
con<strong>for</strong>mational details <strong>of</strong> proteins or peptides in solutions. Several one-dimensional NMR<br />
studies <strong>of</strong> oxytocin and vasopressin analogs have previously been per<strong>for</strong>med in various<br />
solvents such as dimethyl sulfoxide [13], deuterated dimethylsulfoxide [14,15], deuterated<br />
trifuoroethanol [16,17], and aqueous solutions [14,18,19]. Since resonance overlap is much<br />
reduced in two-dimensional NMR spectra in comparison with one-dimensional NMR, we<br />
used 2D NMR spectroscopy.<br />
Our most recent study [9] showed that Zn 2+ in combination with aspartate buffer<br />
strongly stabilizes oxytocin in aqueous solutions, while Ca 2+ and Mg 2+ only have minor<br />
effects. The stabilization occurred as a result <strong>of</strong> complex <strong>for</strong>mation <strong>of</strong> Zn 2+ ions and<br />
aspartate with oxytocin, which, by protecting the Cys 1,6 disulfide bridge, suppressed<br />
dimerization. In line with those results, ITC data demonstrate that only Zn 2+ ions, among<br />
the tested divalent metal ions (Zn 2+ , Ca 2+ , and Mg 2+ ) is able to strongly interact with<br />
oxytocin in the <strong>for</strong>mulation conditions [9]. Those results lead to the hypothesis that Zn 2+<br />
ions induce a different con<strong>for</strong>mation, thereby stabilizing oxytocin in aspartate buffer. To<br />
investigate the con<strong>for</strong>mation <strong>of</strong> oxytocin in aspartate buffer in the presence <strong>of</strong> divalent<br />
metal ions (Zn 2+ and Mg 2+ ), two-dimensional Nuclear Magnetic Resonance (NMR)<br />
spectroscopy has been used.<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
97
6<br />
98<br />
Nuclear Magnetic Resonance spectroscopy is a suitable technique to study<br />
con<strong>for</strong>mational detail <strong>of</strong> proteins or peptides in solution. Several one-dimensional NMR<br />
studies <strong>of</strong> oxytocin and vasopressin analogs have been previously per<strong>for</strong>med in various<br />
solvents such as dimethyl sulfoxide [13], deuterated dimethylsulfoxide [14,15], deuterated<br />
trifuoroethanol [16,17], and water [14,18,19]. Since resonance overlap is much reduced<br />
in two-dimensional NMR spectra in comparison with one-dimensional NMR, we used<br />
2D NMR spectroscopy.<br />
Oxytocin solutions are commonly <strong>for</strong>mulated in the concentration <strong>of</strong> 5 IE/mL<br />
or approximately 0.01 mM, while the stability studies previously have been done at a<br />
concentration <strong>of</strong> 0.1 mM. The concentration used in this study (10 mM) was higher, since it<br />
enables NMR measurements without 13 C or 15 N enrichment, relying only on the low natural<br />
abundance <strong>of</strong> these isotopes.<br />
A complete NMR analysis <strong>of</strong> oxytocin in phosphate buffer has been reported [20].<br />
However, no reports are available on the 1 H, 13 C, and 15 N resonance assignments <strong>of</strong> oxytocin<br />
in aspartate buffer in the absence and presence <strong>of</strong> divalent metal ions. In this study, we used<br />
2D NMR spectroscopy to investigate the con<strong>for</strong>mation <strong>of</strong> oxytocin in the presence <strong>of</strong> Zn 2+<br />
or Mg 2+ in aspartate buffer at pH 4.5.<br />
2. materIals and methods<br />
2.1 Materials<br />
Oxytocin monoacetate powder (Diosynth. Oss, The Netherlands) was kindly provided<br />
by MSD, Oss, The Netherlands. Deuterium oxide (D 2 O, isotopic purity 99.9 atom % D)<br />
containing 0.75% TSP(3-(trimethylsilyl) propionic-2,2,3,3,-d4 acid, sodium salt) was<br />
purchased from Aldrich, Steinheim, Germany and deuterated L-aspartic acid-2,3,3-d3 was<br />
purchased from Medical Isotope, Inc, NH. TSP was used as an internal standard having a<br />
chemical shift (d) <strong>of</strong> 0.0 ppm. Zinc chloride was purchased from Fluka, Steinheim, Germany.<br />
All reagents used <strong>for</strong> the NMR experiments were <strong>of</strong> analytical grade (purity > 99%), and<br />
were used without further purification.<br />
2.2 Sample preparation<br />
Two different types <strong>of</strong> NMR samples were prepared with the following compositions:<br />
1. 2D 13 C- 1 H HSQC NMR samples<br />
10 mM oxytocin (natural abundance) in 10 mM deuterated aspartate buffer (pH 4.5 = pD 5.1)<br />
in D 2 O containing 0.75% TSP in the absence (OT-AP) and presence <strong>of</strong> 100 mM ZnCl 2<br />
(OT-AP-Zn) or 100 mM MgCl 2 (OT-AP-Mg). Reference solution was oxytocin in D 2 O.<br />
2. 2D 15 N- 1 H HSQC, 1 H- 1 H TOCSY, and 1 H- 1 H NOESY samples.<br />
10 mM oxytocin (natural abundance) in 10 mM aspartate buffer (pH 4.5) in H 2 O<br />
containing 0.75% TSP in the absence and presence <strong>of</strong> 100 mM ZnCl 2 or 100 mM MgCl 2 .<br />
Reference solution was oxytocin in water.<br />
2.3 Sample preparation<br />
Spectra <strong>of</strong> these samples were recorded using a Varian Unity INOVA 600 MHz NMR<br />
spectrometer equipped with pulsed field-gradient probes. The spectra were recorded at<br />
278 K, processed using NMR Pipe [21] and analyzed using Sparky [22]. All in<strong>for</strong>mation<br />
about the NMR measurements is summarized in Table 1.
Table 1. In<strong>for</strong>mation about the NMR experiments<br />
Experiment Correlation<br />
15 1 N and H<br />
separated by one<br />
bond<br />
15 1 N- H HSQC (NH-HN Nε-Hε <strong>for</strong><br />
Gln, Nd-Hd <strong>for</strong> Asn,<br />
N-H <strong>for</strong> amide <strong>of</strong><br />
Gly9 )<br />
13 1 C and H<br />
separated by one<br />
13 1 C- H HSQC bond<br />
(Ha-Ca, Hβ-Cβ,<br />
Hγ-Cγ, etc.)<br />
Correlates all protons<br />
1 1 H- H TOCSY in a J-coupled spin<br />
system<br />
Correlates all protons<br />
1 1 H- H NOESY which are close in<br />
space (
6<br />
100<br />
Figure 1. 2D NMR spectra <strong>of</strong> oxytocin in aspartate buffer: 15 N- 1 H HSQC spectrum which<br />
shows the correlation between amide protons and amide nitrogens in the backbone <strong>of</strong><br />
the peptide. The correlation between amide protons and amide nitrogens in the side<br />
chain <strong>of</strong> Gln 4 , Asn 5 and the C-terminal Gly 9 (which has a carboxy-amide group instead <strong>of</strong> a<br />
regular carboxylate, indicated by N1-H11 and N1-H12)are also indicated (A); 13 C- 1 H HSQC<br />
spectrum shows the correlation between all side chain carbon signals with signals due to<br />
the attached protons. The Tyr 2 Cd and Cε signals are aliased in this spectrum. The real<br />
Tyr 2 Cd and Cε chemical shifts are 132.69 and 117. 95 ppm, respectively (B)<br />
Figure 1 2D NMR spectra <strong>of</strong><br />
aspartate buffer: 15 N- 1 H HSQC spe<br />
shows the correlation between a<br />
and amide nitrogens in the back<br />
peptide. The correlation between a<br />
and amide nitrogens in the side c<br />
Asn 5 and the C-terminal Gly 9 (<br />
carboxy-amide group instead o<br />
carboxylate, indicated by N<br />
N1-H12)are also indicated (A); 13<br />
spectrum shows the correlation bet<br />
chain carbon signals with signal<br />
attached protons. The Tyr 2 Cδ an<br />
are aliased in this spectrum. The<br />
and Cε chemical shifts are 132.69<br />
ppm, respectively (B)
Figure 2. 2D 1 H- 1 H NOESY spectrum <strong>of</strong> oxytocin in aspartate buffer.<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
101
6<br />
102<br />
Figure 3. Short distances observed in the 2D 1 H- 1 H NOESY spectrum <strong>of</strong> oxytocin in<br />
aspartate buffer.<br />
3.3 Chemical shift difference (Dd)<br />
The chemical shift differences induced by the divalent metal ion upon complexation with<br />
oxytocin in aspartate buffer were analyzed to get in<strong>for</strong>mation about which residues are<br />
involved in binding <strong>of</strong> the divalent metal ions to oxytocin and to learn about the extent <strong>of</strong><br />
the perturbation caused by these ions.<br />
3.3.1 Influence <strong>of</strong> aspartate buffer on the con<strong>for</strong>mation <strong>of</strong> oxytocin in water.<br />
Aspartate buffer is one <strong>of</strong> the buffers known to stabilize oxytocin if divalent metal ions are<br />
present in the liquid <strong>for</strong>mulation. To investigate the influence <strong>of</strong> aspartate buffer on the<br />
con<strong>for</strong>mation <strong>of</strong> oxytocin, the chemical shift (d) <strong>of</strong> Ca and Ha backbone resonances <strong>of</strong><br />
oxytocin in deuterated aspartate buffer was compared with the chemical shift <strong>of</strong> Ca and Ha<br />
resonances <strong>of</strong> oxytocin in D 2 O. The difference between those chemical shifts was expressed<br />
as chemical shift difference (Dd) in ppm.<br />
Figure 4A shows that aspartate buffer induced a minor change in oxytocin in D 2 O. The<br />
largest chemical shift changes were observed in the Ca and Hα resonances <strong>of</strong> Cys 1 . Ca’s<br />
<strong>of</strong> Cys 1 and Cys 6 were more shielded and Ca’s <strong>of</strong> Tyr 2 and Ile 3 were more deshielded in<br />
the presence <strong>of</strong> aspartate buffer. Ca’s <strong>of</strong> Gln 4 , Asn 5 , Pro 7 and Leu 8 were not affected by the<br />
presence <strong>of</strong> aspartate buffer. As shown by black bars in Figure 4A, the Ha’s <strong>of</strong> Cys 1 , Ile 3
Figure 4. Chemical shift difference (Dd) <strong>of</strong> Ca (light grey bars) and Ha (black bars) <strong>of</strong> A.<br />
oxytocin in deuterated aspartate buffer (pD 5.1) and B. oxytocin in D 2O in the presence<br />
<strong>of</strong> Zn 2+ relative to the chemical shifts <strong>of</strong> the same spins measured in D 2O. Oxytocin in<br />
deuterated aspartate buffer (pD 5.1) in the presence <strong>of</strong> C. Zn 2+ and D. Mg 2+ relative to the<br />
chemical shifts <strong>of</strong> the same spins measured in deuterated aspartate buffer pD 5, analyzed<br />
by 2D 13 C- 1 H HSQC spectroscopy.<br />
and Cys 6 are also affected. The opposite effects <strong>of</strong> aspartate on the Ca and Ha chemical<br />
shifts <strong>of</strong> Cys 1 may be due to an electrostatic interaction between the positive charge <strong>of</strong> the<br />
N terminus <strong>of</strong> Cys 1 and the negative charge <strong>of</strong> the carboxylate group <strong>of</strong> the aspartate at<br />
pD 5.1. In order to test this hypothesis, we recorded 2D 13 C- 1 H HSQC <strong>of</strong> oxytocin in the<br />
presence and absence <strong>of</strong> aspartate buffer at pD 2.0. The same chemical shifts were found<br />
in the two spectra (see Supporting in<strong>for</strong>mation). This result shows that at a very low pH,<br />
where the carboxylate groups <strong>of</strong> aspartate are totally protonated, there is no effect at Cys 1 ,<br />
apparently because the electrostatic interaction between the aspartate and the N terminus<br />
<strong>of</strong> Cys 1 has disappeared.<br />
3.3.2 Influence <strong>of</strong> zinc ions on the con<strong>for</strong>mation <strong>of</strong> oxytocin in water.<br />
Figure 4B shows that Ca and Ha <strong>of</strong> the amino acid residues <strong>of</strong> oxytocin other than Leu 8<br />
are shifted in the presence <strong>of</strong> Zn 2+ . The largest chemical shift changes are observed in Ca <strong>of</strong><br />
Tyr 2 . In Tyr 2 , Gln 4 , Cys 6 , Pro 7 , and Gly 9 the Ca spins are more deshielded in the presence <strong>of</strong><br />
Zn 2+ . In contrast, in Ile 3 and Asn 5 the Ca spins are more shielded in the presence <strong>of</strong> Zn 2+ .<br />
The presence <strong>of</strong> Zn 2+ does not cause changes in the Ca and Ha chemical shifts <strong>of</strong> Leu 8 .<br />
The largest Dd observed <strong>for</strong> Ha is in Cys 1 . Interestingly, similar to the effect <strong>of</strong> aspartate<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
103
6<br />
104<br />
buffer on oxytocin, in many residues the chemical shift change <strong>of</strong> Ca induced by Zn 2+ is <strong>of</strong><br />
opposite sign compared to that <strong>of</strong> Ha.<br />
3.3.3 Influence <strong>of</strong> divalent metal ions on the con<strong>for</strong>mation <strong>of</strong> oxytocin in<br />
aspartate buffer<br />
As depicted in Figure 4C, the effects <strong>of</strong> Zn 2+ on chemical shift <strong>of</strong> oxytocin in aspartate<br />
are very similar to the effects observed in D 2 O (Figure 4B), suggesting that a similar<br />
con<strong>for</strong>mational change is induced in both circumstances. A small chemical shift change<br />
was observed only in the Ca <strong>of</strong> Cys 1 .<br />
The effects <strong>of</strong> Mg 2+ on the chemical shifts <strong>of</strong> Ca and Ha resonances <strong>of</strong> oxytocin in<br />
aspartate buffer are much smaller than those <strong>of</strong> Zn 2+ ions. Strikingly, also in the case <strong>of</strong><br />
Mg 2+ , the effects on Ca resonances are opposite to the effects on the Ha resonance <strong>of</strong> the<br />
same residue, except <strong>for</strong> Cys 6 and Leu 8 . The effects <strong>of</strong> Zn 2+ and Mg 2+ ions on the Ca and Ha<br />
chemical shifts <strong>of</strong> oxytocin in the presence <strong>of</strong> aspartate are displayed in Figure 6.<br />
Figure 5. Overlay <strong>of</strong> Ca and Ha signals from 13C-1H HSQC spectra <strong>of</strong> oxytocin in deuterated<br />
aspartate buffer without (red) and with Mg2+ (green) (A) or Zn2+ (blue) (B). The Pro7 Cd-Hd<br />
correlation is also shown in these spectra.<br />
Figure 5 Overlay <strong>of</strong> Cα and Hα signals from 13 C- 1 H HSQC spectra <strong>of</strong> oxytocin in<br />
deuterated aspartate buffer without (red) and with Mg 2+ (green) (A) or Zn 2+ (blue) (B).<br />
The Pro 7 Cδ-Hδ correlation is also shown in these spectra.<br />
4 Discussion<br />
The results <strong>of</strong> this study indicate that Zn 2+ and aspartate buffer changes the
4 dIscussIon<br />
The results <strong>of</strong> this study indicate that Zn 2+ and aspartate buffer changes the con<strong>for</strong>mation<br />
<strong>of</strong> oxytocin. These con<strong>for</strong>mational changes most likely contribute to the stabilization <strong>of</strong><br />
oxytocin in aqueous solution. Our study presents nearly complete NMR assignment<br />
<strong>of</strong> oxytocin in aspartate buffer in the presence and absence <strong>of</strong> divalent metal ions, Zn 2+<br />
and Mg 2+ . 2D 1 H- 1 H NOESY spectra <strong>of</strong> oxytocin in aspartate buffer clearly demonstrate<br />
that the residue pairs Ile 3 -Asn 5 and Ile 3 -Cys 6 are in near proximity. These NOEs are also<br />
present in a 2D 1 H- 1 H NOESY spectrum <strong>of</strong> oxytocin in water. Molecular dynamic studies<br />
by Wyttenbach et al. [11] suggested an interaction between Ha <strong>of</strong> Tyr 2 and the Gly 9 amide<br />
group <strong>of</strong> oxytocin in water, but we could not find evidence <strong>for</strong> this interaction.<br />
Aspartate buffer induces a minor change in the NMR spectrum oxytocin in D 2 O with the<br />
largest chemical shift changes are observed in Cys 1 . We suggest that there is an electrostatic<br />
interaction between the positively charged N terminus <strong>of</strong> Cys 1 and the negatively charged<br />
carboxylate groups <strong>of</strong> aspartate. This hypothesis is supported by the fact that in pD 2, when<br />
aspartate is no longer negatively charged, no changes in chemical shifts are induced by aspartate.<br />
Zn 2+ has similar effects on oxytocin in aspartate buffer and in D 2 O, suggesting that a<br />
similar con<strong>for</strong>mational change is induced in both circumstances. In our previous study [7],<br />
Zn 2+ was shown to improve oxytocin stability only slightly in water, but much more in the<br />
presence <strong>of</strong> aspartate. The small effect <strong>of</strong> Zn 2+ on the chemical shift <strong>of</strong> the Ca <strong>of</strong> Cys 1 , which<br />
was observed only in the presence <strong>of</strong> aspartate, may be relevant in this respect. We suggest<br />
that the carboxylate group <strong>of</strong> aspartate neutralizes the positive charge <strong>of</strong> the N terminus<br />
<strong>of</strong> Cys 1 , allowing the interactions with Zn 2+ to become more favorable. The arrangement<br />
<strong>of</strong> carbonyl/carboxyl groups around the Zn 2+ might also play a role and may explain<br />
the observed chemical-shift changes and the protection <strong>of</strong> the disulfide bridge against<br />
intermolecular reactions that lead to dimerization and inactivation.<br />
The observation <strong>of</strong> chemical-shift changes in Ca and Ha resonances suggest that small<br />
changes in the backbone dihedral angles occur to accommodate Zn 2+ . However drastic<br />
changes in the overall structure <strong>of</strong> the molecule are not likely, since no drastic changes<br />
were observed in the pattern <strong>of</strong> NOEs upon addition <strong>of</strong> Zn 2+ (see supporting in<strong>for</strong>mation).<br />
Increased propensities <strong>for</strong> the <strong>for</strong>mation <strong>of</strong> helical or extended secondary-structure<br />
elements are not likely, because these would be accompanied by correlated changes in Ca<br />
and Ha chemical shifts [29], while in our study we found these changes to be mostly anticorrelated.<br />
On the other hand, the presence <strong>of</strong> a positively charged Zn 2+ ion per se will cause<br />
polarization <strong>of</strong> nearby chemical bonds and thus may explain the observed anti-correlation<br />
in the Ca and Ha chemical-shift changes.<br />
The effects <strong>of</strong> Mg 2+ on the chemical shifts <strong>of</strong> Ca and Ha resonances <strong>of</strong> oxytocin in<br />
aspartate buffer are much smaller than those <strong>of</strong> Zn 2+ . This result is in agreement with our<br />
finding from isothermal titration calorimetry that Mg 2+ show only very weak heat effects<br />
when added to solutions <strong>of</strong> oxytocin [9]. The small chemical-shift changes in the backbone<br />
<strong>of</strong> residues Cys 1 , Tyr 2 and Ile 3 in the presence <strong>of</strong> Mg 2+ in aspartate buffer may be due to<br />
a cation-pi interaction between Mg 2+ and the aromatic side chain <strong>of</strong> Tyr 2 . However, the<br />
interactions between Mg 2+ and oxytocin in aspartate buffer are too weak to cause a significant<br />
shift in the con<strong>for</strong>mational equilibrium that is essential <strong>for</strong> protection against inactivation.<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
105
6<br />
106<br />
5 conclusIon<br />
In conclusion, our NMR studies revealed that Zn 2+ cause a shift in the con<strong>for</strong>mational<br />
equilibrium <strong>of</strong> oxytocin in aqueous solution. Zn 2+ causes changes in the chemical shifts <strong>of</strong><br />
almost all residues <strong>of</strong> oxytocin while Mg 2+ only induces very little chemical-shift changes<br />
in some residues. In contrast an analysis <strong>of</strong> the NOESY spectra in the presence and absence<br />
<strong>of</strong> Zn 2+ showed that the same NOEs are present under both circumstances, although with<br />
slightly different intensities. The exact nature <strong>of</strong> the changes induced by Zn 2+ is not yet<br />
clear. Small changes in the backbone dihedral angles to accommodate the Zn 2+ may explain<br />
the chemical shift changes, but drastic changes in the overall structure <strong>of</strong> the molecule<br />
are not likely, since no drastic changes in the pattern <strong>of</strong> NOEs could be observed upon<br />
addition <strong>of</strong> Zn 2+ . Alternatively, the presence <strong>of</strong> a positively charged zinc ion per se will<br />
cause polarization <strong>of</strong> nearby chemical bonds and thus may explain the observed chemical<br />
shift changes. A definitive explanation <strong>of</strong> these observations must await a more detailed,<br />
dynamic description <strong>of</strong> the peptide in the absence and presence <strong>of</strong> Zn 2+ , possibly from MD<br />
simulations steered by distance- and angle-restraints derived from the NMR spectra.<br />
acknowledgments<br />
This study was per<strong>for</strong>med within the framework <strong>of</strong> the Dutch Top Institute <strong>Pharma</strong> project:<br />
number D6–202.<br />
The authors want to thank MSD Oss <strong>for</strong> providing oxytocin <strong>for</strong> the study and Ruud M.<br />
Scheek at the Groningen Biomolecular Sciences and Biotechnology Institute <strong>for</strong> helpful<br />
discussions.<br />
references<br />
1. K.L. Maughan, S.W. Heim, S.S. Galazka,<br />
Preventing postpartum hemorrhage:<br />
managing the third stage <strong>of</strong> labor, Am. Fam.<br />
Physician 73 (2006) 1025-8.<br />
2. du Vigneaud V, Ressler C,Trippett S., The<br />
sequence <strong>of</strong> amino acids in oxytocin, with a<br />
proposal <strong>for</strong> the structure <strong>of</strong> oxytocin, J. Biol.<br />
Chem 205 (1953) 949-57.<br />
3. J.W. Gard, J.M. Alexander, R.E. Bawdon,<br />
J.T. Albrecht, Oxytocin preparation stability<br />
in several common obstetric intravenous<br />
solutions, Am. J. Obstet. Gynecol. 186<br />
(2002) 496-8.<br />
4. A. Hawe, R. Poole, S. Romeijn, P. Kasper, R.<br />
van der Heijden, W. Jiskoot, Towards heatstable<br />
oxytocin <strong>for</strong>mulations: analysis <strong>of</strong><br />
degradation kinetics and identification <strong>of</strong><br />
degradation products, Pharm Res 26 (2009)<br />
1679-88.<br />
5. H.V. Hogerzeil, G.J.A. Walker, M.J. De Goeje,<br />
Stability <strong>of</strong> injectable ocytocics in tropical<br />
climates, World Health Organization,<br />
Geneva WHO/DAP/93.6 (1993).<br />
6. Trissel LA, Zhang Y, Douglass K,Kastango<br />
E., Extended Stability <strong>of</strong> Oxytocin in<br />
common infusion solution, International<br />
Journal <strong>of</strong> <strong>Pharma</strong>ceutical Compounding 10<br />
(2006) 156-8.<br />
7. C. Avanti, J.P. Amorij, D. Setyaningsih, A.<br />
Hawe, W. Jiskoot, J. Visser, A. Kedrov, A.J.<br />
Driessen, W.L. Hinrichs, H.W. Frijlink, A<br />
new strategy to stabilize oxytocin in aqueous<br />
solutions: I. The effects <strong>of</strong> divalent metal ions<br />
and citrate buffer, AAPS J. 13 (2011) 284-90.<br />
8. C. Avanti, H.P. Permentier, A.V. Dam, R.<br />
Poole, W. Jiskoot, H.W. Frijlink, W.L.J.<br />
Hinrichs, A New Strategy To Stabilize<br />
Oxytocin in Aqueous Solutions: II.<br />
Suppression <strong>of</strong> Cysteine-Mediated<br />
Intermolecular Reactions by a Combination<br />
<strong>of</strong> Divalent Metal Ions and Citrate, Mol.<br />
<strong>Pharma</strong>ceutics 9 (3) (2012) 554-62.<br />
9. C. Avanti, W.L.J. Hinrichs, A. Casini, A.C.<br />
Eissens, A.V. Dam, A. Kedrov, A.J.M.<br />
Driessen, H.W. Frijlink, H.P. Permentier,<br />
Insight into the stability <strong>of</strong> the zincaspartate-oxytocin<br />
<strong>for</strong>mulation, manuscript<br />
in preparation.<br />
10. P. Jurgens, C. Panteliadis, G. Fondalinski,<br />
Total parenteral nutrition <strong>of</strong> premature
infants: metabolic effects <strong>of</strong> an exogenous<br />
supply <strong>of</strong> L-aspartic acid and L-glutamic<br />
acid, Z. Ernahrungswiss. 21 (1982) 225-45.<br />
11. T. Wyttenbach, D. Liu, M.T. Bowers,<br />
Interactions <strong>of</strong> the hormone oxytocin with<br />
divalent metal ions, J. Am. Chem. Soc. 130<br />
(2008) 5993-6000.<br />
12. A.F. Pearlmutter, M.S. Sol<strong>of</strong>f, Characterization<br />
<strong>of</strong> the metal ion requirement <strong>for</strong> oxytocinreceptor<br />
interaction in rat mammary gland<br />
membranes, J. Biol. Chem. 254 (1979) 3899-<br />
906.<br />
13. E. Sikorska, M.J. Slusarz, B. Lammek,<br />
Con<strong>for</strong>mational studies <strong>of</strong> vasopressin<br />
analogues modified with Nmethylphenylalanine<br />
enantiomers in<br />
dimethyl sulfoxide solution, Biopolymers 82<br />
(2006) 603-14.<br />
14. J.D. Glickson, R. Rowan, T.P. Pitner, J. Dadok,<br />
A.A. Bothner-By, R. Walter, 1H nuclear<br />
magnetic resonance double resonance<br />
study <strong>of</strong> oxytocin in aqueous solution,<br />
Biochemistry 15 (1976) 1111-9.<br />
15. L.F. Johnson, I.L. Schwartz, R. Walter,<br />
Oxytocin and neurohypophyseal peptides<br />
spectral assignment and con<strong>for</strong>mational<br />
analysis by 220 M1Hz nuclear magnetic<br />
resonance, Proc. Nat. Acad. Sci 64 (1969)<br />
1269-75.<br />
16. M. Rholam, P. Cohen, N. Brakch, L. Paolillo,<br />
A. Scatturin, C. Di Bello, Evidence <strong>for</strong><br />
beta-turn structure in model peptides<br />
reproducing pro-ocytocin/neurophysin<br />
proteolytic processing site, Biochem,<br />
Biophys, Res, Commun, 168 (1990) 1066-73.<br />
17. V.S. Ananthanarayanan, M.P. Belciug, B.S.<br />
Zhorov, Interaction <strong>of</strong> oxytocin with Ca 2+ : II.<br />
Proton magnetic resonance and molecular<br />
modeling studies <strong>of</strong> con<strong>for</strong>mations <strong>of</strong> the<br />
hormone and its Ca 2+ complex, Biopolymers<br />
40 (1996) 445-64.<br />
18. E. Sikorska, S. Rodziewicz-Motowidlo,<br />
Con<strong>for</strong>mational studies <strong>of</strong> vasopressin<br />
and mesotocin using NMR spectroscopy<br />
and molecular modelling methods. Part I:<br />
Studies in water, J. Pept. Sci. 14 (2008) 76-84.<br />
19. I.C. Smith, R. Deslauriers, H. Saito, R.<br />
Walter, C. Garrigou-Lagrange, H. McGregor,<br />
D. Sarantakis, Carbon-13 NMR studies <strong>of</strong><br />
peptide hormones and their components,<br />
Ann. N. Y. Acad. Sci. 222 (1973) 597-627.<br />
20. A. Ohno, N. Kawasaki, K. Fukuhara, H.<br />
Okuda, T. Yamaguchi, Complete NMR<br />
analysis <strong>of</strong> oxytocin in phosphate buffer,<br />
Magn. Reson. Chem. 48 (2010) 168-172.<br />
21. F. Delaglio, S. Grzesiek, G.W. Vuister,<br />
G. Zhu, J. Pfeifer, A. Bax, NMR pipe - a<br />
multidimensional spectral processing<br />
system based on unix pipes., J Biomol NMR<br />
6 (1995) 277-293.<br />
22. T.D. Goddard, D.G. Kneller, SPARKY 3,<br />
University <strong>of</strong> Cali<strong>for</strong>nia, San Francisco,<br />
2003.<br />
23. L.E. Kay, P. Keifer, T. Saarinen, Pure Absorption<br />
Gradient Enhanced Heteronuclear Single<br />
Quantum Spectroscopy with Improved<br />
Sensitivity, J. Am. Chem. Soc. 114 (1992)<br />
10633-5.<br />
24. F.A. Mulder, R. Otten, R.M. Scheek, Origin<br />
and removal <strong>of</strong> mixed-phase artifacts in<br />
gradient sensitivity enhanced heteronuclear<br />
single quantum correlation spectra, J.<br />
Biomol. NMR 51 (2011) 199-207.<br />
25. K.B. John, D. Plant, R.E. Hurd, Improved<br />
Proton-Detected Heteronuclear Correlation<br />
Using Gradient-Enhanced z and zz filters, J.<br />
Magn. Reson. A101 113-7.<br />
26. D.G. Davis, A. Bax, Assignment <strong>of</strong> Complex<br />
1H NMR spectra via Two-Dimensional<br />
Homonuclear Hartman-Hahn Spectroscopy,<br />
J. Am. Chem. Soc. 107 (1985) 2820-1.<br />
27. J. Boyd, G.R. Moore, G. Williams,<br />
Correlation <strong>of</strong> proton chemical shifts in<br />
proteins using two-dimensional exchange<br />
correlated spectroscopy, J. Magn. Reson. 58<br />
(1984) 511-6.<br />
28. C. Cavanagh, W.J. Fairbrother, A.G. Palmer<br />
III, N.J. Skelton (Eds.), Protein NMR<br />
Spectroscopy, Second Edition: Principles<br />
and Practice, 2nd ed., Elsevier Inc., United<br />
State <strong>of</strong> America, 2006.<br />
29. D.S. Wishart, Interpreting protein chemical<br />
shift data, Prog Nucl Magn Reson Spectrosc<br />
58 (2011) 62-87.<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
107
6<br />
108<br />
supportIng In<strong>for</strong>matIon<br />
Table S1. Resonance list <strong>of</strong> oxytocin in aspartate buffer<br />
Group Atom Nuc Shift SDev Assignments<br />
C1 Ca 13 C 54.963 0 1<br />
C1 Cβ 13 C 42.593 0.009 2<br />
C1 Ha 1 H 4.225 0.013 2<br />
C1 Hβ2 1 H 3.279 0 1<br />
C1 Hβ3 1 H 3.477 0 1<br />
Y2 Ca 13 C 58.13 0 1<br />
Y2 Cβ 13 C 38.843 0.009 2<br />
Y2 Cd 13 C 132.686 0 1<br />
Y2 Cε 13 C 117.951 0 1<br />
Y2 H 1 H 9.118 0.009 10<br />
Y2 Ha 1 H 4.781 0 1<br />
Y2 Hβ2 1 H 3.002 0.005 4<br />
Y2 Hβ3 1 H 3.2 0.007 4<br />
Y2 Hd 1 H 7.247 0.017 5<br />
Y2 Hε 1 H 6.875 0.006 6<br />
Y2 N 15 N 123.84 0 1<br />
I3 Ca 13 C 62.55 0 1<br />
I3 Cβ 13 C 38.694 0 1<br />
I3 Cd 13 C 13.49 0 1<br />
I3 Cγ 13 C 27.28 0.021 2<br />
I3 Cγ1 13 C 17.67 0 1<br />
I3 H 1 H 8.103 0.052 16<br />
I3 Ha 1 H 4.073 0.011 3<br />
I3 Hβ 1 H 1.94 0.002 3<br />
I3 Hd 1 H 0.888 0.009 4<br />
I3 Hγ1 1 H 0.901 0.004 2<br />
I3 Hγ2 1 H 1.011 0.004 2<br />
I3 Hγ3 1 H 1.237 0.003 4<br />
I3 N 15 N 120.224 0 1<br />
Q4 Ca 13 C 57.793 0 1<br />
Q4 Cβ 13 C 28.472 0 1<br />
Q4 Cγ 13 C 33.688 0 2<br />
Q4 H 1 H 8.368 0.016 12<br />
Q4 Ha 1 H 4.116 0.005 3<br />
Q4 Hβ 1 H 2.079 0.004 3<br />
Q4 Hε21 1 H 6.989 0.013 6<br />
Q4 Hε22 1 H 7.705 0.012 8<br />
continued next page
Group Atom Nuc Shift SDev Assignments<br />
Q4 HG2 1 H 2.41 0.003 2<br />
Q4 HG3 1 H 2.426 0.003 2<br />
Q4 N 15 N 120.399 0 1<br />
Q4 Nε 15 N 112.68 0.001 2<br />
N5 Ca 13 C 53.044 0 1<br />
N5 Cβ 13 C 38.413 0 1<br />
N5 H 1 H 8.455 0.006 13<br />
N5 Ha 1 H 4.767 0.022 5<br />
N5 Hβ 1 H 2.878 0.003 2<br />
N5 Hd21 1 H 7.068 0.01 8<br />
N5 Hd22 1 H 7.759 0.009 7<br />
N5 N 15 N 116.465 0 1<br />
N5 Nd 15 N 112.933 0.004 2<br />
C6 Ca 13 C 54.019 0 1<br />
C6 Cβ 13 C 40.798 0.004 2<br />
C6 H 1 H 8.332 0.006 11<br />
C6 Ha 1 H 4.914 0.022 2<br />
C6 Hβ2 1 H 3.005 0.064 3<br />
C6 Hb3 1 H 3.245 0.003 2<br />
C6 N 15 N 119.907 0 1<br />
P7 Ca 13 C 63.203 0 1<br />
P7 Cβ 13 C 32.035 0.002 2<br />
P7 Cd 13 C 50.569 0.001 2<br />
P7 Cγ 13 C 27.467 0 1<br />
P7 Ha 1 H 4.46 0.004 2<br />
P7 Hβ2 1 H 1.941 0 1<br />
P7 Hβ3 1 H 2.314 0 2<br />
P7 Hd2 1 H 3.744 0.011 2<br />
P7 Hd3 1 H 3.774 0 1<br />
P7 Hγ 1 H 2.049 0 1<br />
L8 Ca 13 C 55.346 0 1<br />
L8 Cβ 13 C 41.873 0.005 2<br />
L8 Cd1 13 C 24.903 0 1<br />
L8 Cd2 13 C 23.345 0 1<br />
L8 Cγ 13 C 27.046 0 1<br />
L8 H 1 H 8.705 0.012 11<br />
L8 Ha 1 H 4.313 0.004 3<br />
L8 Hβ2 1 H 1.619 0.005 2<br />
L8 Hβ3 1 H 1.713 0 1<br />
L8 Hd1 1 H 0.950 0.004 3<br />
L8 Hd2 1 H 0.911 0.014 2<br />
L8 Hγ 1 H 1.699 0.004 3<br />
L8 N 15 N 122.909 0 1<br />
continued next page<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
109
6<br />
110<br />
Group Atom Nuc Shift SDev Assignments<br />
G9 Ca 13 C 44.757 0.006 2<br />
G9 H 1 H 8.589 0.007 7<br />
G9 H11 1 H 7.51 0.005 6<br />
G9 H12 1 H 7.215 0.004 10<br />
G9 Ha1 1 H 3.875 0.004 2<br />
G9 Ha2 1 H 3.957 0.01 3<br />
G9 N 15 N 111.181 0 1<br />
G9 N1 15 N 107.471 0.002 2<br />
Table S2. Resonance list <strong>of</strong> oxytocin in the presence <strong>of</strong> Zn 2+ in aspartate buffer<br />
Group Atom Nuc Shift SDev Assignments<br />
C1 Ca 13 C 55.064 0 1<br />
C1 Cβ 13 C 43.085 0.011 2<br />
C1 Ha 1 H 4.317 0.044 2<br />
C1 Hβ2 1 H 3.295 0.008 2<br />
C1 Hβ3 1 H 3.554 0.057 2<br />
Y2 Ca 13 C 58.649 0 1<br />
Y2 Cβ 13 C 38.689 0.005 2<br />
Y2 Cd 13 C 132.905 0 1<br />
Y2 Cε 13 C 118.028 0 1<br />
Y2 Ha 1 H 4.748 0.029 2<br />
Y2 Hβ2 1 H 3.022 0.021 4<br />
Y2 Hβ3 1 H 3.189 0.003 4<br />
Y2 Hd 1 H 7.235 0.011 7<br />
Y2 Hε 1 H 6.872 0.006 6<br />
Y2 N 15 N 123.932 0 1<br />
I3 Ca 13 C 61.979 0 1<br />
I3 Cβ 13 C 38.842 0 1<br />
I3 Cd 13 C 13.663 0 1<br />
I3 Cγ 13 C 27.233 0.001 2<br />
I3 Cγ1 13 C 17.826 0 1<br />
I3 H 1 H 8.131 0.009 15<br />
I3 Ha 1 H 4.098 0.025 3<br />
I3 Hβ 1 H 1.945 0.011 3<br />
I3 Hd 1 H 0.884 0.011 4<br />
I3 Hγ1 1 H 0.894 0.002 3<br />
I3 Hγ2 1 H 1.042 0.021 2<br />
I3 Hγ3 1 H 1.234 0.006 4<br />
I3 N 15 N 120.225 0 1<br />
Q4 Ca 13 C 57.986 0 1<br />
continued next page
Group Atom Nuc Shift SDev Assignments<br />
Q4 Cβ 13 C 28.54 0 1<br />
Q4 Cγ 13 C 33.781 0 2<br />
Q4 H 1 H 8.374 0.027 13<br />
Q4 Ha 1 H 4.096 0.017 3<br />
Q4 Hβ 1 H 2.07 0.005 3<br />
Q4 Hε21 1 H 6.995 0.017 6<br />
Q4 Hε22 1 H 7.704 0.014 7<br />
Q4 Hγ2 1 H 2.399 0.009 2<br />
Q4 Hγ3 1 H 2.423 0.005 2<br />
Q4 N 15 N 120.391 0 1<br />
Q4 Nε 15 N 112.749 0.006 2<br />
N5 Ca 13 C 52.976 0 1<br />
N5 Cβ 13 C 38.542 0 1<br />
N5 H 1 H 8.456 0.006 13<br />
N5 Ha 1 H 4.758 0.012 4<br />
N5 Hβ 1 H 2.88 0.003 3<br />
N5 Hd21 1 H 7.068 0.009 6<br />
N5 N 15 N 116.474 0 1<br />
N5 Nd 15 N 113.046 0 2<br />
C6 Ca 13 C 54.282 0 1<br />
C6 Cβ 13 C 40.909 0.001 2<br />
C6 H 1 H 8.33 0.005 11<br />
C6 Ha 1 H 4.878 0.022 2<br />
C6 Hβ2 1 H 2.97 0.01 2<br />
C6 Hβ3 1 H 3.264 0.004 2<br />
C6 N 15 N 119.856 0 1<br />
P7 Ca 13 C 63.595 0 1<br />
P7 Cβ 13 C 32.115 0 2<br />
P7 Cd 13 C 50.779 0.003 2<br />
P7 Cγ 13 C 27.628 0 1<br />
P7 Ha 1 H 4.449 0.004 2<br />
P7 Hβ2 1 H 1.936 0 1<br />
P7 Hβ3 1 H 2.308 0 2<br />
P7 Hd2 1 H 3.738 0.017 2<br />
P7 Hd3 1 H 3.77 0 1<br />
P7 Hγ 1 H 2.036 0 1<br />
L8 Ca 13 C 55.351 0 1<br />
L8 Cβ 13 C 41.604 0.003 2<br />
L8 Cd1 13 C 25.116 0 1<br />
L8 Cd2 13 C 23.383 0 1<br />
L8 Cγ 13 C 27.151 0 1<br />
L8 H 1 H 8.699 0.005 10<br />
L8 Ha 1 H 4.308 0.007 3<br />
continued next page<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
111
6<br />
112<br />
Group Atom Nuc Shift SDev Assignments<br />
L8 Hβ2 1 H 1.611 0.01 3<br />
L8 Hβ3 1 H 1.768 0 1<br />
L8 Hd1 1 H 0.941 0.032 3<br />
L8 Hd2 1 H 0.879 0.01 2<br />
L8 Hγ 1 H 1.688 0.022 3<br />
L8 N 15 N 122.855 0 1<br />
G9 Ca 13 C 44.726 0.002 2<br />
G9 H 1 H 8.572 0.006 7<br />
G9 H11 1 H 7.51 0.004 5<br />
G9 H12 1 H 7.217 0.003 9<br />
G9 Ha1 1 H 3.877 0.011 2<br />
G9 Ha2 1 H 3.94 0 1<br />
G9 N 15 N 111.157 0 1<br />
G9 N1 15 N 107.5 0.016 2<br />
Table S3. Resonance list <strong>of</strong> oxytocin in the presence <strong>of</strong> Mg 2+ in aspartate buffer<br />
Group Atom Nuc Shift SDev Assignments<br />
C1 Ca 13 C 54.873 0 1<br />
C1 Cβ 13 C 42.513 0 2<br />
C1 Ha 1 H 4.287 0.015 2<br />
C1 Hβ2 1 H 3.295 0.011 2<br />
C1 Hβ3 1 H 3.51 0.003 2<br />
Y2 Ca 13 C 58.233 0 1<br />
Y2 Cβ 13 C 38.786 0.002 2<br />
Y2 Cd 13 C 132.706 0 1<br />
Y2 Cε 13 C 117.98 0 1<br />
Y2 H 1 H 9.133 0.005 12<br />
Y2 Ha 1 H 4.771 0.003 2<br />
Y2 Hβ2 1 H 3.012 0.008 3<br />
Y2 Hβ3 1 H 3.184 0.004 4<br />
Y2 Hd 1 H 7.233 0.009 4<br />
Y2 Hε 1 H 6.872 0.007 7<br />
Y2 N 15 N 123.963 0 1<br />
I3 Ca 13 C 62.434 0 1<br />
I3 Cβ 13 C 38.714 0 1<br />
I3 Cd 13 C 13.585 0.084 2<br />
I3 Cγ 13 C 27.312 0 2<br />
I3 Cγ1 13 C 17.739 0.076 2<br />
I3 H 1 H 8.1 0.063 13<br />
continued next page
Group Atom Nuc Shift SDev Assignments<br />
I3 Ha 1 H 4.074 0.008 3<br />
I3 Hβ 1 H 1.938 0.003 3<br />
I3 Hd 1 H 0.877 0.007 7<br />
I3 Hγ1 1 H 0.894 0.005 3<br />
I3 Hγ2 1 H 1.018 0.008 2<br />
I3 Hγ3 1 H 1.23 0.006 4<br />
I3 N 15 N 120.247 0 1<br />
Q4 Ca 13 C 57.874 0 1<br />
Q4 Cβ 13 C 28.457 0 1<br />
Q4 Cγ 13 C 33.668 0 2<br />
Q4 H 1 H 8.359 0.013 10<br />
Q4 Ha 1 H 4.108 0.002 2<br />
Q4 Hβ 1 H 2.069 0.006 3<br />
Q4 Hε21 1 H 6.974 0 2<br />
Q4 Hε22 1 H 7.696 0.01 6<br />
Q4 Hγ2 1 H 2.404 0.003 2<br />
Q4 Hγ3 1 H 2.435 0 1<br />
Q4 N 15 N 120.407 0 1<br />
N5 Ca 13 C 53.012 0 1<br />
N5 Cβ 13 C 38.448 0 1<br />
N5 H 1 H 8.452 0.007 10<br />
N5 Ha 1 H 4.756 0.01 4<br />
N5 Hβ 1 H 2.881 0.005 3<br />
N5 Hd21 1 H 7.065 0.009 4<br />
N5 Hd22 1 H 7.763 0 5<br />
N5 N 15 N 116.458 0 1<br />
N5 Nd 15 N 113.06 0.008 2<br />
C6 Ca 13 C 54.063 0 1<br />
C6 Cβ 13 C 40.784 0.007 2<br />
C6 H 1 H 8.325 0.004 7<br />
C6 Ha 1 H 4.897 0.004 2<br />
C6 Hβ2 1 H 2.96 0.006 2<br />
C6 Hβ3 1 H 3.262 0.008 2<br />
C6 N 15 N 119.877 0 1<br />
P7 Ca 13 C 63.284 0 1<br />
P7 Cβ 13 C 32.052 0.002 2<br />
P7 Cd 13 C 50.61 0 2<br />
P7 Cγ 13 C 27.497 0 1<br />
P7 Ha 1 H 4.453 0.006 2<br />
P7 Hβ2 1 H 1.938 0 1<br />
P7 Hβ3 1 H 2.312 0 1<br />
P7 Hd2 1 H 3.745 0.004 2<br />
P7 Hd3 1 H 3.749 0 1<br />
continued next page<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
113
6<br />
114<br />
Group Atom Nuc Shift SDev Assignments<br />
P7 Hγ 1 H 2.044 0 1<br />
L8 Ca 13 C 55.332 0 1<br />
L8 Cβ 13 C 41.813 0.003 2<br />
L8 Cd1 13 C 24.926 0 1<br />
L8 Cd2 13 C 23.365 0 1<br />
L8 Cγ 13 C 27.07 0 1<br />
L8 H 1 H 8.693 0.003 9<br />
L8 Ha 1 H 4.309 0.006 3<br />
L8 Hβ2 1 H 1.622 0.007 2<br />
L8 Hβ3 1 H 1.725 0 1<br />
L8 Hd1 1 H 0.932 0.029 3<br />
L8 Hd2 1 H 0.904 0.002 2<br />
L8 Hγ 1 H 1.695 0.008 2<br />
L8 N 15 N 122.836 0 1<br />
G9 Ca 13 C 44.806 0.001 2<br />
G9 H 1 H 8.564 0.001 5<br />
G9 H12 1 H 7.214 0.003 9<br />
G9 Ha1 1 H 3.87 0.006 2<br />
G9 Ha2 1 H 3.942 0 1<br />
G9 N 15 N 111.19 0 1<br />
G9 N1 15 N 107.554 0.009 2<br />
Table S4. Resonance List <strong>of</strong> oxytocin in water<br />
Group Atom Nuc Shift SDev Assignments<br />
C1 Ca 13 C 55.228 0 1<br />
C1 Cβ 13 C 42.996 0.004 2<br />
C1 Ha 1 H 4.199 0.051 3<br />
C1 Hβ2 1 H 3.22 0 1<br />
C1 Hβ3 1 H 3.392 0 1<br />
Y2 Ca 13 C 58.032 0 1<br />
Y2 Cβ 13 C 38.834 0 2<br />
Y2 Cd 13 C 132.686 0 1<br />
Y2 Cε 13 C 117.951 0 1<br />
Y2 H 1 H 9.125 0.004 11<br />
Y2 Ha 1 H 4.78 0 1<br />
Y2 Hβ2 1 H 3.002 0.006 4<br />
Y2 Hβ3 1 H 3.21 0.039 5<br />
Y2 Hd 1 H 7.244 0.015 5<br />
Y2 Hε 1 H 6.876 0.005 6<br />
Y2 N 15 N 123.788 0 1<br />
continued next page
Group Atom Nuc Shift SDev Assignments<br />
I3 Ca 13 C 62.415 0 1<br />
I3 Cβ 13 C 38.757 0 1<br />
I3 Cd 13 C 13.49 0 2<br />
I3 Cγ 13 C 27.259 0.002 2<br />
I3 Cγ1 13 C 17.67 0 2<br />
I3 H 1 H 8.109 0.055 16<br />
I3 Ha 1 H 4.072 0.012 3<br />
I3 Hβ 1 H 1.941 0.006 3<br />
I3 Hd 1 H 0.885 0.008 5<br />
I3 Hγ1 1 H 0.902 0.005 3<br />
I3 Hγ2 1 H 1.01 0.015 2<br />
I3 Hγ3 1 H 1.228 0.007 4<br />
I3 N 15 N 120.158 0 1<br />
Q4 Ca 13 C 57.793 0 1<br />
Q4 Cγ 13 C 33.683 0.002 2<br />
Q4 H 1 H 8.371 0.023 12<br />
Q4 Ha 1 H 4.114 0.006 3<br />
Q4 Hβ 1 H 2.078 0.005 3<br />
Q4 Hε21 1 H 6.989 0.014 6<br />
Q4 Hε22 1 H 7.704 0.009 8<br />
Q4 HG2 1 H 2.413 0.002 2<br />
Q4 HG3 1 H 2.423 0.009 2<br />
Q4 N 15 N 120.41 0 1<br />
Q4 Nε 15 N 112.664 0.005 2<br />
N5 Ca 13 C 53.044 0 1<br />
N5 Cβ 13 C 38.356 0 1<br />
N5 H 1 H 8.458 0.006 13<br />
N5 Ha 1 H 4.763 0.016 5<br />
N5 Hβ 1 H 2.877 0.003 2<br />
N5 Hd22 1 H 7.756 0.01 7<br />
N5 N 15 N 116.463 0 1<br />
N5 Nd 15 N 112.86 0.001 2<br />
C6 Ca 13 C 54.093 0 1<br />
C6 Cβ 13 C 40.903 0.007 2<br />
C6 H 1 H 8.337 0.004 11<br />
C6 Ha 1 H 4.895 0.015 2<br />
C6 Hβ2 1 H 3.02 0.088 3<br />
C6 Hb3 1 H 3.244 0.007 2<br />
C6 N 15 N 119.936 0 1<br />
P7 Ca 13 C 63.203 0 1<br />
P7 Cβ 13 C 32.034 0.002 2<br />
P7 Cd 13 C 50.57 0.002 2<br />
P7 Cγ 13 C 27.467 0 1<br />
continued next page<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
115
6<br />
116<br />
Group Atom Nuc Shift SDev Assignments<br />
P7 Ha 1 H 4.46 0.003 2<br />
P7 Hβ2 1 H 1.94 0 1<br />
P7 Hβ3 1 H 2.312 0.002 2<br />
P7 Hd2 1 H 3.744 0.011 2<br />
P7 Hd3 1 H 3.774 0 1<br />
P7 Hγ 1 H 2.049 0 1<br />
L8 Ca 13 C 55.346 0 1<br />
L8 Cβ 13 C 41.873 0.005 2<br />
L8 Cd1 13 C 24.879 0 1<br />
L8 Cd2 13 C 23.383 0 1<br />
L8 Cγ 13 C 27.046 0 1<br />
L8 Ha 1 H 4.311 0.005 3<br />
L8 Hβ2 1 H 1.657 0.042 2<br />
L8 Hβ3 1 H 1.713 0 1<br />
L8 Hd1 1 H 0.958 0.005 3<br />
L8 Hd2 1 H 0.911 0 1<br />
L8 Hγ 1 H 1.696 0.003 3<br />
L8 N 15 N 122.962 0 1<br />
G9 Ca 13 C 44.763 0 2<br />
G9 H 1 H 8.592 0.013 8<br />
G9 H11 1 H 7.515 0.003 6<br />
G9 H12 1 H 7.213 0.005 10<br />
G9 Ha1 1 H 3.872 0.005 2<br />
G9 Ha2 1 H 3.955 0.009 3<br />
G9 N 15 N 111.18 0 1<br />
G9 N1 15 N 107.426 0.005 2
Figure S1. Overlay <strong>of</strong> Ca and Ha<br />
signals from 13 C- 1 H HSQC spectra<br />
<strong>of</strong> oxytocin in D 2 O pD 2 (red) and<br />
oxytocin in deuterated aspartate<br />
buffer pD 2 (green).<br />
13 C (ppm)<br />
5.0<br />
4.5<br />
4.5<br />
4.0<br />
4.0<br />
1 H (ppm)<br />
3.5<br />
45 45<br />
50 P7Cδ-Hδ 50<br />
N5Cα-Hα<br />
C6Cα-Hα<br />
C1Cα-Hα<br />
55 55<br />
Y2Cα-Hα<br />
L8Cα-Hα<br />
60 60<br />
5.0<br />
G9Cα-Hα<br />
Q4Cα-Hα<br />
I3Cα-Hα<br />
P7Cα-Hα<br />
3.5<br />
con<strong>for</strong>MatIon <strong>of</strong> dIvalent Metal-aSpartate-oxytocIn coMplex<br />
6<br />
117
6<br />
118<br />
1 H (ppm)<br />
0<br />
2<br />
4<br />
6<br />
8<br />
10<br />
9.5<br />
Y2Hβ3-H<br />
Y2Hβ2-H<br />
C1Hβ2-Y2H<br />
C1Hβ3-Y2H<br />
L8Hγ-H<br />
I3Hβ-Q4H<br />
L8Hδ-N5H<br />
I3Hβ-H<br />
L8Hδ-H<br />
Q4Hβ-H<br />
Q4Hβ-N5H<br />
Q4Hγ2-H<br />
P7Hβ3-L8H<br />
L8Hα-H<br />
C1Hα-Y2H<br />
P7Hα-L8H<br />
Y2Hα-H<br />
I3Hδ-N5H<br />
I3Hγ1-Q4H<br />
I3Hδ-H<br />
I3Hγ2-H<br />
I3Hγ3-N5H I3Hγ3-H<br />
I3Hγ3-Q4H<br />
L8Hβ2-H<br />
L8Hβ2-G9H<br />
N5Hβ3-C6H<br />
N5Hβ-H<br />
N5Hβ2-Hδ22<br />
C6Hβ2-H<br />
Y2Hβ2-Hδ<br />
C6Hβ3-H Y2Hβ2-I3H<br />
P7Hδ2-C6H<br />
G9Hα1-H<br />
Q4Hα-N5H<br />
I3Hα-H<br />
Q4Hα-C6H<br />
I3Hα-Q4H<br />
L8Hα-G9H<br />
N5Hα-H N5Hα-Q4H<br />
C6Hα-H<br />
Y2Hβ3-I3H<br />
Y2Hβ3-Hδ<br />
N5Hδ21-Hδ22<br />
Y2Hε-Hδ<br />
Q4Hε21-Hε22<br />
G9H12-H11 G9H12-H12<br />
Y2Hε-Hε<br />
N5Hδ21-Q4Hε22<br />
Y2Hδ-Hδ<br />
Q4Hε21-Hε21<br />
Q4Hε22-Hε22<br />
N5Hδ21-Hδ21<br />
N5Hδ22-Hδ22<br />
Y2Hδ-Y2Hε<br />
I3H-Y2H I3H-C6H<br />
I3H-Q4H I3H-H<br />
Q4H-H<br />
Q4H-N5H<br />
G9H-H<br />
L8H-H<br />
C6H-H<br />
G9H11-H12<br />
G9H11-H11<br />
N5Hδ22-Hδ21<br />
Q4Hε21-Hε22<br />
Q4H-I3H<br />
Q4Hε22-Hε21<br />
Y2H-H<br />
N5H-H N5H-C6H<br />
Y2H-I3H<br />
Y2H-Hδ<br />
9.0<br />
8.5<br />
8.0<br />
7.5<br />
1 H (ppm)<br />
I3Hδ-Y2Hδ I3Hγ1-Y2Hε<br />
Y2Hα-Hδ<br />
N5Hβ-Hδ21<br />
7.0<br />
6.5<br />
Figure S2. 2D 1 H- 1 H NOESY spectrum <strong>of</strong> oxytocin in the presence <strong>of</strong> Zn 2+ in aspartate<br />
buffer.
Christina Avanti 1,*, Vinay Saluja 1,2,* , Erwin L.P. van Streun 1 , Imma Boyten 1 ,<br />
Henderik W. Frijlink 1 , Wouter L.J. Hinrichs 1<br />
*Equally contributed first author<br />
1 Department <strong>of</strong> <strong>Pharma</strong>ceutical Technology &Biopharmacy,<br />
University <strong>of</strong> Groningen, Groningen, The Netherlands<br />
2 Vaccinology, National Institute <strong>for</strong> Public Health and the Environment,<br />
Bilthoven, The Netherlands
extremolytes: are there unIversal<br />
stabIlIzers <strong>for</strong> proteIns In aqueous<br />
solutIon?<br />
7
7<br />
abstract<br />
The purpose <strong>of</strong> this study was to investigate the ability <strong>of</strong> extremolytes to stabilize the model<br />
proteins lysozyme and insulin in aqueous solutions. The effects <strong>of</strong> the extremolytes, betaine,<br />
hydroxyectoine, trehalose, ectoine, and firoin on the stability <strong>of</strong> lysozyme were determined by<br />
Nile red Fluorescence Spectroscopy and a bioactivity assay. Insulin stability was determined<br />
by RP-HPLC and HP-SEC. The effects <strong>of</strong> extremolytes on the unfolding temperature <strong>of</strong><br />
the proteins were analyzed using a thermal shift assay <strong>for</strong> lysozyme and liquid differential<br />
scanning microcalorimetry <strong>for</strong> insulin. The interaction between extremolytes and protein<br />
was studied by isothermal titration calorimetry (ITC). During storage at 70ºC <strong>for</strong> 10 min,<br />
firoin protected lysozyme against inactivation better than the other extremolytes. During<br />
storage at 55ºC <strong>for</strong> 4 weeks, firoin also acted as a stabilizer, however, betaine, hydroxyectoine,<br />
trehalose and ectoine, destabilized lysozyme. These findings surprisingly indicate that<br />
some extremolytes can stabilize proteins under certain stress conditions but destabilize<br />
the same proteins under other stress conditions. The increased stability caused by firoin<br />
is explained by the observed increased unfolding temperature <strong>of</strong> lysozyme in the presence<br />
<strong>of</strong> firoin. After storage at 40ºC <strong>for</strong> 4 weeks, trehalose and ectoine protected insulin against<br />
degradation, while betaine, hydroxyectoine and in particular firoin seemed to destabilize<br />
insulin. Furthermore, it was shown that firoin sharply decreased the unfolding temperature<br />
<strong>of</strong> insulin. The interaction <strong>of</strong> firoin with lysozyme and ectoine or trehalose with insulin<br />
was negligible as determined by ITC. Thus, firoin appears to be an excellent stabilizer <strong>for</strong><br />
lysozyme but strongly destabilizes insulin, whereas ectoine showed the opposite behaviour.<br />
This study clearly shows that there is not one extremolyte that can acts as a universal<br />
stabilizer <strong>for</strong> proteins in aqueous solution.
1. IntroductIon<br />
Protein instability is one <strong>of</strong> the main issues in the administration <strong>of</strong> therapeutic protein<br />
based medicines, in particular in aqueous <strong>for</strong>mulations. Protein stability can be achieved if<br />
there is a balance between intramolecular interaction <strong>of</strong> protein functional groups and their<br />
interaction with solvent environment [1]. A number <strong>of</strong> experimental studies have been<br />
done to overcome protein instability. One <strong>of</strong> the most promising results is the discovery <strong>of</strong><br />
extremolytes, small organic osmolytes found in extremophiles.<br />
Extremophiles are microorganisms which are capable <strong>of</strong> surviving under extreme<br />
conditions, such as high or low temperatures, extreme pressure and high salt concentrations.<br />
Extremolytes are accumulated in response to these extreme conditions and minimize the<br />
denaturation <strong>of</strong> the biological macromolecules and proteins [2]. The polyols derivatives<br />
ectoine and hydroxyectoine are the first extremolytes that are currently produced in a large<br />
scale and are already being used as protein stabilizers. Furthermore, also other molecules<br />
such as betaine[3] various amino acids, carbohydrates such as trehalose[4]and the mannose<br />
derivative firoin[5,6] have been found in extremophyles and have been identified as stabilizers<br />
in these species. Figure 1 shows the molecular structures <strong>of</strong> some <strong>of</strong> these extremolytes.<br />
Extremolytes stabilize proteins by <strong>for</strong>ming solute hydrate clusters [7] that are<br />
preferentially excluded from the hydrate shell <strong>of</strong> the protein [8]. Exclusion occurs as a result<br />
<strong>of</strong> the repulsive interactions between extremolytes and the backbone <strong>of</strong> the proteins [9] and<br />
generates accumulation <strong>of</strong> water near protein domains. Thus, proteins turn into a more<br />
compact tertiary structure with a reduced surface area.<br />
According to this mechanism, in the presence <strong>of</strong> extremolytes the native state <strong>of</strong> proteins<br />
is in the lower free energy state than in the unfolded state. Proteins lose their con<strong>for</strong>mational<br />
Figure 1. Molecular structure <strong>of</strong> the extremolytes (a) mannosylglycerate (firoin), (b) betaine,<br />
(c) ectoine, (d) hydroxyectoine, and (e) and trehalose<br />
extreMolyteS aS proteInS StabIlIzerS<br />
7<br />
123
7<br />
124<br />
Figure 2. Mode <strong>of</strong> action extremolytes, adapted from Lentzen G., 2006 (2)<br />
entropy with the greater entropic loss in the unfolded state (S u ), leading to an overall shift<br />
in equilibrium towards the native state [10]. The entropy difference between the two states<br />
increases, thus the folded protein (S f ) is less likely to unfold as ilustrated in Figure 3 [11].<br />
The aim <strong>of</strong> this study was to investigate whether known extremolytes such as betaine,<br />
hydroxyectoine, trehalose, ectoine, and firoin are able to stabilize the model proteins<br />
lysozyme and insulin at elevated temperatures and whether they can be used as universal<br />
stabilizers <strong>for</strong> proteins in aqueous solutions.<br />
2. materIals and methods<br />
2.1 Materials<br />
The following materials were used in this study: Hen egg-white lysozyme (Sigma-<br />
Aldrich, Steinheim, Germany), Recombinant human insulin, USP (provided by MSD,<br />
Oss, The Netherlands), ultrapure trehalose(Cargill, Krefeld, Germany), ultrapure betaine<br />
(Fluka Biochemika, Buchs, Switzerland), firoin(Biotop, Berlin-Brandenburg, Germany),<br />
ultrapure ectoine and hydroxyectoine(Biomol, Hamburg, Germany), citric acid (Riedel-de<br />
Haen, Seelze, Germany), sodium hydroxide, acetonitrile (Merck, Darmstadt, Germany),<br />
dimetylsulfoxide(DMSO), Nile red and M. Lysodeickticus(Sigma-Aldrich, Steinheim,<br />
Germany), PBS buffer (66 mM phosphate, pH 6.2), SYPRO orange protein gel stain,<br />
(invitrogenEugene, Oregon, USA), sodium sulphate, concentrated phosphoric acid,<br />
ethanolamine, hydrochloric acid (Merck, Darmstad, Germany), phosphate buffered saline<br />
(Fluka Analytical, Steinheim, Germany).<br />
2.2 The effect <strong>of</strong> extremolytes on the stability <strong>of</strong> lysozyme<br />
2.2.1 Heat shock stability study<br />
Lysozyme solutions at a concentration <strong>of</strong> 100 μg/ml were prepared in 10 mM citrate buffer<br />
(pH 5.0) with and without extremolytes at a concentration <strong>of</strong> 0.5 M [2,5]. Lysozyme solutions<br />
were incubated at 70ºC <strong>for</strong> 10 minutes. The effect <strong>of</strong> extremolytes on the stability <strong>of</strong> lysozyme<br />
was determined by Nile red fluorescence spectroscopy and by measuring its bioactivity.
Figure 3. Proposed mechanism <strong>of</strong><br />
stabilization <strong>of</strong> extremolytes, adapted<br />
from Arakawa, 2006 (12)<br />
2.2.2 Accelerated stability studies<br />
Lysozyme solutions at a concentration <strong>of</strong> 100 μg/ml were prepared in 10 mM citrate buffer<br />
(pH 5.0) with and without 0.5 M stabilizers. Lysozyme solutions were incubated at 55ºC.<br />
The effect <strong>of</strong> extremolytes on the stability <strong>of</strong> lysozyme was determined by measuring its<br />
bioactivity. Bioassays were per<strong>for</strong>med every week <strong>for</strong> 4 weeks.<br />
2.2.3 Nile red Fluorescence Spectroscopy<br />
Fluorescence studies were per<strong>for</strong>med on a SLM-Aminco AB2 Spectr<strong>of</strong>luorometer at<br />
the temperature <strong>of</strong> 25°C using 5 mm cubical quartz cuvette (Hellma GmbH). Prior<br />
to measurement, 5 μl <strong>of</strong> 20 μg/ml Nile red solution in DMSO was added to a 1 ml <strong>of</strong><br />
100 μg/ml lysozyme solution, to obtain a Nile red : lysozyme weight ratio <strong>of</strong> 1:1000. An<br />
excitation wavelength <strong>of</strong> 550 nm was used, with a band pass <strong>of</strong> 4.0 nm <strong>for</strong> the excitation<br />
monochromator and an emission wavelength <strong>of</strong> 610 nm with a band pass <strong>of</strong> 4.0 nm <strong>for</strong> the<br />
emission monochromator. Data were recorded at 1 nm intervals over the range 560–700 nm<br />
with a scanning speed <strong>of</strong> 100 nm/min. Spectra were corrected <strong>for</strong> background signal caused<br />
by buffer and stabilizers [12].<br />
2.2.4 Bioassay<br />
The bioactivity <strong>of</strong> lysozyme was determined by measuring the rate <strong>of</strong> lysis <strong>of</strong> Micrococcus<br />
lysodeikticus by using a method as described by Shugar[13] with some modifications.<br />
Briefly, 1.3 ml <strong>of</strong> a 200 µg/ml <strong>of</strong> Micrococcus lysodeikticus suspension in sterile PBS buffer<br />
(66 mM phosphate, pH 6.2) was mixed with 10 µl <strong>of</strong> the lysozyme solutions in plastic disposable<br />
cuvettes. Immediately after mixing, the cuvette was placed in a UV/VIS spectrophotometer<br />
and the absorbance was measured at a wavelength <strong>of</strong> 450 nm. Subsequently, the change <strong>of</strong> the<br />
absorbance was recorded over time. Remaining activity <strong>of</strong> the stressed samples was obtained<br />
by comparing them with the unstressed samples. Statistical analyses were per<strong>for</strong>med using<br />
Student’s t test with p < 0.05 as the minimal level <strong>of</strong> significance. The results are presented as<br />
mean ± standard error mean unless indicated otherwise.<br />
2.3 The effect <strong>of</strong> extremolytes on the stability <strong>of</strong> insulin<br />
Insulin was <strong>for</strong>mulated at a concentration <strong>of</strong> 20 IE/ml in 2 mM PBS buffer at pH 7.4 with<br />
and without extremolytes at a concentration <strong>of</strong> 0.5 M. After preparation, the solutions were<br />
stored at either 4°C or 40°C <strong>for</strong> 4 weeks, and protected from light. The pH <strong>of</strong> the samples<br />
was measured regularly and was found to be remained at 7.4 ± 0.1 during the stability study.<br />
extreMolyteS aS proteInS StabIlIzerS<br />
7<br />
125
7<br />
126<br />
2.3.1 Reversed-Phase High-Per<strong>for</strong>mance Liquid Chromatography (RP-HPLC)<br />
The recovery <strong>of</strong> insulin was determined by RP-HPLC as described in the USP. A Chrompack<br />
C18 column with 5 µm particle size, inner diameter <strong>of</strong> 4.6 mm and length <strong>of</strong> 250 mm, a<br />
Waters 510 HPLC pump, a Waters 717 Autosampler and a Waters 486 tunable absorbance<br />
UV detector were used. Samples <strong>of</strong> 10 µl were injected and the separation was carried<br />
out at a flow rate <strong>of</strong> 1.0ml/min and UV detection at 214 nm. Samples were eluted using<br />
27% acetonitrile in buffer pH 2.3. The buffer <strong>for</strong> the mobile phase was prepared by dissolving<br />
71.0 g <strong>of</strong> sodium sulphate in 2100 ml <strong>of</strong> Milli Q water. After adding 6.75 ml <strong>of</strong> phosphoric<br />
acid, the pH was adjusted to 2.3 with ethanolamine or 0.5 N phosphoric acid. The total<br />
volume was completed to 2500 mL and filtered through a 0.45µm filter. Integration <strong>of</strong> the<br />
chromatograms was done with Chromeleon s<strong>of</strong>tware. The runtime was 30.5 minutes [14].<br />
2.3.2 High-Per<strong>for</strong>mance Liquid-Size Exclusion Chromatography (HP-SEC)<br />
The monomeric fraction <strong>of</strong> insulin was assessed by HP-SEC[15]. HP-SEC was carried<br />
out using a Waters Insulin HMWP column, inner diameter <strong>of</strong> 7.8 mm, length 200 mm,<br />
Waters 510 HPLC pump, Waters 717 plus autosampler, Waters 486 tunable absorbance<br />
UV detector (Waters, Mil<strong>for</strong>d Massachusetts, USA)was used. Samples <strong>of</strong> 20 µl were<br />
injected, and separation was per<strong>for</strong>med at a flow rate <strong>of</strong> 0.5 ml/min. Peaks were detected<br />
by UV absorption at 276 nm. Samples were eluted using mobile phase consisted <strong>of</strong> 40% <strong>of</strong><br />
L-arginine 1 mg/ml solution, 50% <strong>of</strong> acetonitrile and 10% glacial acetic acid. The runtime<br />
was 30.5 minutes.<br />
2.4 The effect <strong>of</strong> extremolytes on protein unfolding temperature (T m )<br />
2.4.1 Thermal shift assay<br />
The unfolding temperature <strong>of</strong> lysozyme in solutions was analyzed by a thermal shift assay<br />
using a real-time PCR machine [16]. Solutions <strong>of</strong> 17.5 μl <strong>of</strong> 1.0 mg/ml lysozyme with<br />
or without stabilizers (1 M) and 7.5 μl <strong>of</strong> 300 fold diluted SYPRO Orange solution as a<br />
molecular probes were added to the wells <strong>of</strong> a 96-well thin-wall PCR plate (Bio-Rad).<br />
The plates were sealed with optical-quality sealing tape (Bio-Rad Laboratories BV,<br />
Veenendaal, The Netherlands), inserted into a real-time PCR machine (iCycler, Bio-Rad<br />
Laboratories BV, Veenendaal, The Netherlands) and heated from 20 to 90°C with a 0.2°C<br />
increaseper 20 s. The fluorescence changes <strong>of</strong> the SYPRO Orange probe in the wells <strong>of</strong><br />
the plate were monitored simultaneously with a fluorescence detector (MyIQ single-colour<br />
RT-PCR detection system, Bio-Rad) at excitation and emission wavelengths <strong>of</strong> 490 and<br />
575 nm, respectively. The midpoint <strong>of</strong> the transition was taken as the T m .<br />
2.4.2 Liquid Differential Scanning calorimetry<br />
The T m <strong>of</strong> insulin in solutions in a concentration <strong>of</strong> 1.5 mg/mL with or without<br />
extremolytes (0.1 M) was analyzed by differential scanning microcalorimetry. Data collection<br />
was per<strong>for</strong>med using a VP-DSC differential scanning microcalorimeter (MicroCal, LLC,<br />
Northampton, MA) [17,18]. All insulin scans were per<strong>for</strong>med with 2 mM PBS buffer pH 7.4<br />
in the reference cell from 25 to 110°C at a scan rate <strong>of</strong> 90°C/hour and an excess pressure <strong>of</strong><br />
25-26 Psi. All samples and references were degassed immediately be<strong>for</strong>e use. A buffer-buffer<br />
reference scan was subtracted from each sample scan prior to concentration normalization.<br />
Data analysis was carried out using Origin 7.0 (OriginLab, Northampton, MA).
2.5 Interaction <strong>of</strong> extremolytes with protein determined by<br />
Isothermal Titration Calorimetry (ITC)<br />
Microcalorimetric titrations <strong>of</strong> extremolytes to lysozyme in citrate buffer pH 5.0 and<br />
insulin in PBS buffer were per<strong>for</strong>med by using a MicroCal ITC 200 microcalorimeter<br />
(Northampton, MA 01060 USA). A solution <strong>of</strong> 20 mM <strong>of</strong> some selected extremolytes in<br />
10 mM citrate buffer, pH 5.0 was placed in the syringe, while 300 µl <strong>of</strong> 1 mM lysozyme in<br />
10 mM <strong>of</strong> citrate buffer, pH 5.0 was placed in the sample cell. The reference cell contained<br />
300 µl <strong>of</strong> citrate buffer. Experiments were per<strong>for</strong>med at 10, 25, and 55°C. The initial delay<br />
time was 60 s. The reference power and the filter were set to 5 µcal/s and 2 s respectively.<br />
A typical titration experiment consisted <strong>of</strong> 20 injections <strong>of</strong> 2 µl extremolytes solutions<br />
with duration <strong>of</strong> 4 s and the time interval between two consecutive injections was set to<br />
180 s. During the experiments, the sample solution was continuously stirred at 1000 rpm.<br />
The effective heat <strong>of</strong> the protein-extremolytes interaction upon each titration step was<br />
corrected <strong>for</strong> dilution and mixing effects, as measured by titrating the extremolytes<br />
solution into the buffer and by titrating the buffer into the protein solution. The heats <strong>of</strong><br />
bimolecular interactions were obtained by integrating the peak following each injection.<br />
All measurements were per<strong>for</strong>med in triplicate. ITC data were analyzed by using the ITC<br />
non-linear curve fitting functions <strong>for</strong> one or two binding sites from MicroCal Origin 7.0<br />
s<strong>of</strong>tware (MicroCal S<strong>of</strong>tware, Inc.) [19].<br />
3. results and dIscussIon<br />
3.1 The effect <strong>of</strong> extremolytes on the stability <strong>of</strong> proteins<br />
In order to get a clear and unambiguous picture <strong>of</strong> the effect <strong>of</strong> the different extremolytes<br />
on the stability <strong>of</strong> the two proteins it was decided to test the stability effects on each <strong>of</strong><br />
the proteins with two different methods. For the lysozyme the Nile Red Fluorescence<br />
Spectroscopy and the activity assay were used, whereas <strong>for</strong> the insulin the RP-HPLC and<br />
HP-SEC were combined. In all studies the extremolytes concentration <strong>of</strong> 0,5 M was used.<br />
3.1.1 Stability <strong>of</strong> lysozyme as measured by Nile Red Fluorescence Spectroscopy<br />
Nile red fluorescence can be employed to probe changes in protein con<strong>for</strong>mations that are<br />
related to the <strong>for</strong>mation <strong>of</strong> hydrophobic surfaces, such as during aggregation or protein<br />
unfolding because <strong>of</strong> its sensitivity to the polarity <strong>of</strong> its environment [20,21].<br />
The effect <strong>of</strong> a heat shock on lysozyme without extremolytes in citrate buffer solution<br />
pH 5.0 is shown in Figure 4A. A huge increase in the fluorescence intensity <strong>of</strong> Nile red was<br />
observed when lysozyme without extremolytes as stressed at 70°C <strong>for</strong> 10 minutes indicating<br />
substantial denaturation <strong>of</strong> lysozyme. Apparently, heating lysozyme without extremolytes<br />
<strong>for</strong> 10 minutes at 70 ºC caused collapse <strong>of</strong> its secondary and/or tertiary structure. Stressed<br />
lysozyme solutions in the presence <strong>of</strong> betaine(Figure 4B) or hydroxyectoine (Figure 4C)<br />
showed similar Nile red fluorescence spectra. However, when trehalosewas added, we<br />
observed a slight difference in the Nile red fluorescence spectra <strong>of</strong> the lysozyme sample<br />
be<strong>for</strong>e and after heat shock. Figure 4D shows that there was a minor shift in the maximum<br />
peak after stress, however, the huge increase in the fluorescence intensities <strong>of</strong> Nile red<br />
as found <strong>for</strong> lysozyme <strong>for</strong>mulations without extremolytes or in the presence betaine or<br />
hydroxyectoine was not observed. This indicates that trehalose was able to inhibit substantial<br />
denaturation <strong>of</strong> lysozyme. The stabilization effect <strong>of</strong> trehalose might be due to the fact that<br />
extreMolyteS aS proteInS StabIlIzerS<br />
7<br />
127
7<br />
128<br />
Figure 4. The effect <strong>of</strong> extremolytes on the tryptophanyl fluorescence <strong>of</strong> the lysozymenile<br />
red complex after stressing lysozyme at 70°C <strong>for</strong> 10 minutes in citrate buffer pH 5.0.<br />
Fluorescence spectra recorded <strong>of</strong> lysozyme unstressed (solid line) and stressed (dotted line)<br />
<strong>of</strong> lysozyme A: without extremolytes, and with B: betaine, C: hydroxyectoine, D: trehalose,<br />
E: ectoine, and F: firoin
trehalose has a propensity to depress the <strong>for</strong>mation <strong>of</strong> aggregates and chemical reactions <strong>of</strong><br />
lysozyme by inducing α-helical structures and some tertiary structures [22].<br />
Also the Nile red fluorescence spectrum <strong>of</strong> the lysozyme solution <strong>for</strong>mulated with ectoine<br />
showed a minor shift in the maximum peak after stress (Figure 4E). Furthermore, also<br />
significantly higher fluorescence intensity was observed in the stressed lysozyme samples<br />
containing ectoine than those containing trehalose. However, this increased intensity was<br />
much smaller than that found <strong>for</strong> <strong>for</strong>mulations without extremolytes or in the presence<br />
betaine or hydroxyectoine. There<strong>for</strong>e, these measurements indicate that ectoine protected<br />
lysozyme against denaturation, however, to a somewhat lesser extent than trehalose.<br />
In conclusion, trehalose and ectoineare able to partially prevent the denaturation <strong>of</strong> lysozyme,.<br />
In contrast, when firoin was added to lysozyme solution, as the Nile red fluorescence spectra <strong>of</strong><br />
unstressed and stressed lysozyme solution were fully identical (Figure 4 F) indicating that firoin<br />
completely inhibited the denaturation <strong>of</strong> lysozyme during heat shock.<br />
3.1.2 Stability <strong>of</strong> lysozyme determined by using Bioassay<br />
In order to further evaluate the stabilizing effects <strong>of</strong> extremolytes during a heat shock at a<br />
temperature <strong>of</strong> 70ºC <strong>for</strong> 10 minutes and during storage at a temperature <strong>of</strong> 55ºC <strong>for</strong> 4 weeks,<br />
the biological activity <strong>of</strong> lysozyme was determined. After heat shock the ability <strong>of</strong> lysozyme<br />
to inactivate the bacterium M. Lysodeickticus decreased dramatically.<br />
Figure 5 shows that lysozyme only maintained about 20-40% <strong>of</strong> its original activity.<br />
When betaine, trehalose, and ectoine were added, however, lysozyme maintained about<br />
70% <strong>of</strong> its original activity. There was no significant difference observed between the<br />
stabilizing effects <strong>of</strong> hydroxyectoine, trehalose, and ectoine but it is difficult to draw a<br />
conclusion from these data since the level <strong>of</strong> significance was low. It seems, however, that<br />
ectoine stabilized lysozyme better than trehalose and hydroxyectoine, also hydroxyectoine<br />
was not significantly different from control.<br />
Figure 5. The effect <strong>of</strong> lysozyme on M. Lysodeickticus (bioactivity) after stressed at 70°C<br />
<strong>for</strong> 10 minutes and the effect <strong>of</strong> extremolytes on the bioactivity <strong>of</strong> lysozyme. * is the level<br />
<strong>of</strong> significance<br />
extreMolyteS aS proteInS StabIlIzerS<br />
7<br />
129
7<br />
130<br />
The bioactivity assay was also used to monitor the effects <strong>of</strong> extremolytes on the stability<br />
<strong>of</strong> lysozyme solution during incubation <strong>for</strong> 4 weeks at 55°C (accelerated stability study).<br />
Figure 6 shows that the addition <strong>of</strong> the extremolytes betaine, hydroxyectoine, trehalose<br />
and ectoine destabilized lysozyme. These results are in contrast to the heat shock experiments<br />
where these extremolytes lead to minor or substantial stabilizing effects. On the other hand,<br />
the bioactivity assay indicated that firoin stabilized lysozyme during storage at a 55°C <strong>for</strong> 4<br />
weeks, which is completely in line with the results <strong>of</strong> the heat shock experiments.<br />
3.1.3 Stability <strong>of</strong> insulin as measured by RP-HPLC and HP-SEC<br />
Figure 7A shows the results <strong>of</strong> insulin recovery analyzed by RP-HPLC method after<br />
incubation at 4°C and 40°C <strong>for</strong> 4 weeks. In PBS buffer alone approximately 95% insulin<br />
was recovered after 4 weeks at 4°C. The same result was obtained when hydroxyectoine,<br />
trehalose or ectoine was added to the solution. When incubated at 40°C <strong>for</strong> 4 weeks<br />
approximately 30% insulin was recovered from the PBS solution.. Addition <strong>of</strong> firoin, betaine<br />
and hydroxyectoine showed a destabilizing effect as these samples showed no insulin<br />
recovery at all after 4 weeks at 40°C. However, ectoine and trehalose showed significant<br />
stabilizing effects, resulting in an insulin recovery <strong>of</strong> 62% and 72%, respectively<br />
Figure 7B shows the effects <strong>of</strong> extremolytes on the amount <strong>of</strong> insulin monomer remaining<br />
after incubation at 4°C and 40°C <strong>for</strong> 4 weeks as measured by HP-SEC. At 4°C, about 90%<br />
insulin remained as monomer in the absence or presence <strong>of</strong> extremolytes, although firoin<br />
resulted in a lower extent <strong>of</strong> monomer remaining. The results after incubation at 40°C <strong>for</strong><br />
4 weeks are in line with the RP-HPLC measurements. In the absence <strong>of</strong> extremolytes, insulin<br />
showed substantial degradation resulting in a remaining amount <strong>of</strong> insulin monomer <strong>of</strong><br />
about 35%. The HP-SEC results confirm that betaine, hydroxyectoine and firoin destabilized<br />
insulin. However, ectoine and trehalose strongly stabilized insulin, resulting in a recovery<br />
<strong>of</strong> 80% and 88% insulin monomer, respectively.<br />
Figure 6. The effect <strong>of</strong> extremolytes on the bioactivity <strong>of</strong> lysozyme during 4 weeks storage<br />
at 55 °C
Figure 7. The effect <strong>of</strong> stabilizers on the concentration <strong>of</strong> insulin in a liquid <strong>for</strong>mulation<br />
after storage at 4°C (light grey bars) or 40°C (dark grey bars) <strong>for</strong> 4 weeks determined by A.<br />
RP-HPLC, B HP-SEC<br />
In conclusion, the RP-HPLC and HP-SEC measurements indicate that trehalose and<br />
ectoine are able to protect insulin against chemical and physical degradation in liquid<br />
<strong>for</strong>mulations, whereas betaine or hydroxyectoine and in particular firoin destabilized insulin.<br />
3.2 The effect <strong>of</strong> extremolytes on protein unfolding temperature (T m )<br />
The unfolding temperature (T m ) <strong>of</strong> lysozyme in solution was analyzed by a thermal shift<br />
assay using a real-time PCR machine. The T m <strong>for</strong> lysozyme was found to be 82°C. While,<br />
the addition <strong>of</strong> firoin, resulted in a much higher increase in T m i.e. 9°C.<br />
These results may explain the improved stability <strong>of</strong> lysozyme in the presence <strong>of</strong> firoin.<br />
The firoin effect may be due to the increase <strong>of</strong> the melting temperature <strong>of</strong> lysozyme by 9ºC.<br />
It may be that a rise in T m <strong>of</strong> 4ºC is not enough to protect lysozyme <strong>for</strong> a longer period<br />
<strong>of</strong> time at 55ºC, but that a rise <strong>of</strong> 9-ºC is enough to stabilize lysozyme <strong>for</strong> at least a week<br />
at 55ºC. Santoro et al showed that up to 8.2 molar <strong>of</strong> extremolytes, including betaine,<br />
were able to increase the T m <strong>of</strong> lysozyme to about 23ºC [23]. It is possible that higher<br />
extremolytes concentrations are required to improve the stability <strong>of</strong> lysozyme during<br />
storage at a high temperature.<br />
Figure 9 shows the T m <strong>of</strong> insulin in the absence and presence <strong>of</strong> extremolytes as<br />
determined by liquid differential calorimetry. A slight but insignificant increase <strong>of</strong> the<br />
T m <strong>of</strong> insulin was found when ectoine was added to the solution. Addition <strong>of</strong> trehalose<br />
and betaine had no effect on the T m <strong>of</strong> insulin, whereas hydroxyectoine and in particular<br />
firoin even decrease the T m <strong>of</strong> insulin. The decreased T m found <strong>for</strong> firoin may explain the<br />
destabilizing effect <strong>of</strong> this extremolyte on insulin.<br />
extreMolyteS aS proteInS StabIlIzerS<br />
7<br />
131
7<br />
132<br />
Figure 8. The effect <strong>of</strong> stabilizers on the unfolding temperature (T m ) <strong>of</strong> lysozyme. Unfolding<br />
temperatures was measured as transition midpoint analyzed by thermal shift assay<br />
using RT-PCR machine <strong>of</strong> lysozyme on the concentration <strong>of</strong> 1.0 mg/ml with A: betaine,<br />
B: hydroxyectoine, C: trehalose, D: ectoine, and E: firoin<br />
Figure 9. The effect <strong>of</strong> stabilizers on the unfolding temperature (T m ) <strong>of</strong> insulin. Unfolding<br />
temperatures was measured as transition midpoint analysed by liquid differential<br />
calorimetry <strong>of</strong> insulin on the concentration <strong>of</strong> 1.5 mg/ml with A: betaine, B: hydroxyectoine,<br />
C: trehalose, D: ectoine, and E: firoin<br />
3.3 Interaction <strong>of</strong> proteins with extremolytes<br />
The interaction <strong>of</strong> extremolytes with proteins was analyzed using Isothermal Titration<br />
Calorimetry at different temperatures <strong>for</strong> the most stable <strong>for</strong>mulations.<br />
Figure 10 shows the similarity in the magnitude <strong>of</strong> the reaction and dilution enthalpy <strong>for</strong><br />
the titration <strong>of</strong> lysozyme by firoin and the titration <strong>of</strong> insulin by ectoine or trehalose at the<br />
temperatures <strong>of</strong> 10, 25, or 55°C. The results indicate that in aqueous solution extremolytes<br />
do not show any significant interaction with the proteins. The stabilizing effects are rather<br />
due to modification <strong>of</strong> the water properties. This is in line with the preferential hydration<br />
theory that the repulsion between the amide backbone <strong>of</strong> the protein and the extremolytes<br />
is due to the influence <strong>of</strong> the extremolytes on the water structure and there<strong>for</strong>e do not<br />
interact directly with the proteins [2].
Figure 10. The effect <strong>of</strong> firoin (A) on the heat capacity <strong>of</strong> lysozyme, and the effect <strong>of</strong> ectoine<br />
(B) and trehalose (C) to the heat capacity <strong>of</strong> insulin determined by ITC<br />
4 conclusIon<br />
This study clearly shows that there are no extremolytes that can be used as a universal stabilizer<br />
<strong>for</strong> all proteins in aqueous solution. We even found that certain extremolytes, (firoin) can act<br />
as a stabilizer <strong>for</strong> a particular protein (lysozyme), but destabilizes another protein (insulin).<br />
Even worse, certain extremolytes (betaine) can stabilize a protein (lysozyme) under certain<br />
conditions (heat shock) but destabilize the same protein under other stress conditions<br />
(accelerated stability conditions). This implies not only that <strong>for</strong> each protein the stabilizing<br />
effects <strong>of</strong> different extremolytes should be screened but also that the envisaged storage<br />
conditions should be taken into account. Furthermore, <strong>for</strong> screening extremolytes <strong>for</strong> the<br />
stabilization <strong>of</strong> proteins, measuring the T m <strong>of</strong> the protein can be useful to predict stabilizing<br />
effects but only if the extremolyte induces a substantial change in the T m .<br />
acknowledgment<br />
This study was per<strong>for</strong>med within the framework <strong>of</strong> the Dutch Top Institute <strong>Pharma</strong><br />
(projectnumber D6−202).<br />
The authors want to thank G.K. Schuurman-Wolters, A. Kedrov and Pr<strong>of</strong>. A.J.M.<br />
Driessen at the Groningen Institute Biomolecular Sciences & Biotechnology Groningen <strong>for</strong><br />
helpful discussions on L-DSC and ITC.<br />
extreMolyteS aS proteInS StabIlIzerS<br />
7<br />
133
7<br />
134<br />
references<br />
1. G. Nemethy, Orientation <strong>of</strong> amino acid side<br />
chains: intraprotein and solvent interactions,<br />
Methods. Enzymol. 127 (1986) 183-96.<br />
2. G. Lentzen, T. Schwarz, Extremolytes:<br />
Natural compounds from extremophiles<br />
<strong>for</strong> versatile applications, Appl. Microbiol.<br />
Biotechnol. 72 (2006) 623-34.<br />
3. S. Knapp, R. Ladenstein, E.A. Galinski,<br />
Extrinsic protein stabilization by the<br />
naturally occurring osmolytes betahydroxyectoine<br />
and betaine, Extremophiles<br />
3 (1999) 191-8.<br />
4. A. Hedoux, J.F. Willart, L. Paccou, Y. Guinet,<br />
F. Affouard, A. Lerbret, M. Descamps,<br />
Thermostabilization mechanism <strong>of</strong> bovine<br />
serum albumin by trehalose, J. Phys. Chem.<br />
B 113 (2009) 6119-26.<br />
5. T.Q. Faria, S. Knapp, R. Ladenstein,<br />
A.L. Macanita, H. Santos, Protein<br />
stabilisation by compatible solutes: effect<br />
<strong>of</strong> mannosylglycerate on unfolding<br />
thermodynamics and activity <strong>of</strong> ribonuclease<br />
A, ChemBioChem 4 (2003) 734-41.<br />
6. N. Borges, A. Ramos, N.D. Raven, R.J.<br />
Sharp, H. Santos, Comparative study<br />
<strong>of</strong> the thermostabilizing properties <strong>of</strong><br />
mannosylglycerate and other compatible<br />
solutes on model enzymes, Extremophiles 6<br />
(2002) 209-16.<br />
7. E.A. Galinski, M. Stein, B. Amendt, M.<br />
Kinder, The kosmotropic (structure<strong>for</strong>ming)<br />
effect <strong>of</strong> compensatory solutes,<br />
Comp. Biochem. Physiol. 117A (1997)<br />
357-65.<br />
8. T. Arakawa, S.N. Timasheff, The stabilization<br />
<strong>of</strong> proteins by osmolytes, Biophys. J. 47<br />
(1985) 411-4.<br />
9. D.W. Bolen, G.D. Rose, Structure and<br />
Energetics <strong>of</strong> the Hydrogen-Bonded<br />
Backbone in Protein Folding, Annu. Rev.<br />
Biochem. 77 (2008) 339-62.<br />
10. A. Linhananta, G. Amadei, T. Miao,<br />
Computer simulation study <strong>of</strong> folding<br />
thermodynamics and kinetics <strong>of</strong> proteins<br />
in osmolytes and denaturants, Journal <strong>of</strong><br />
Physics: Conference Series 341 (2012).<br />
11. T. Arakawa, D. Ejima, Y. Kita, K. Tsumoto,<br />
Small molecule pharmacological chaperones:<br />
From thermodynamic stabilization to<br />
pharmaceutical drugs, Biochim. Biophys.<br />
Acta 1764 (2006) 1677-87.<br />
12. M. Sutter, S. Oliveira, N.N. Sanders, B.<br />
Lucas, A. van Hoek, M.A. Hink, A.J. Visser,<br />
S.C. De Smedt, W.E. Hennink, W. Jiskoot,<br />
Sensitive spectroscopic detection <strong>of</strong> large<br />
and denatured protein aggregates in solution<br />
by use <strong>of</strong> the fluorescent dye Nile red, J.<br />
Fluoresc. 17 (2007) 181-92.<br />
13. D. Shugar, The measurement <strong>of</strong> lysozyme<br />
activity and the ultra-violet inactivation <strong>of</strong><br />
lysozyme, Biochim. Biophys. Acta 8 (1952)<br />
302-9.<br />
14. X. Xu, Y. Fu, H. Hu, Y. Duan, Z. Zhang,<br />
Quantitative determination <strong>of</strong> insulin<br />
entrapment efficiency in triblockcopolymeric<br />
nanoparticles by high-per<strong>for</strong>mance liquid<br />
chromatography, J. Pharm. Biomed. Anal. 41<br />
(2006) 266-73.<br />
15. C.M. Yu, C.Y. Chin, E.I. Franses, N.H.<br />
Wang, In situ probing <strong>of</strong> insulin aggregation<br />
in chromatography effluents with<br />
spectroturbidimetry, J. Colloid. Interface<br />
Sci. 299 (2006) 733-9.<br />
16. U.B. Ericsson, B.M. Hallberg, G.T. Detitta,<br />
N. Dekker, P. Nordlund, Therm<strong>of</strong>luor-based<br />
high-throughput stability optimization<br />
<strong>of</strong> proteins <strong>for</strong> structural studies, Anal.<br />
Biochem. 357 (2006) 289-98.<br />
17. V.V. Plotnikov, J.M. Brandts, L.N. Lin, J.F.<br />
Brandts, A new ultrasensitive scanning<br />
calorimeter, Anal. Biochem. 250 (1997)<br />
237-44.<br />
18. K. Huus, S. Havelund, H.B. Olsen, M. van de<br />
Weert, S. Frokjaer, Thermal dissociation and<br />
unfolding <strong>of</strong> insulin, Biochemistry 44 (2005)<br />
11171-7.<br />
19. C. H<strong>of</strong>fmann, A. Blume, I. Miller, P. Garidel,<br />
Insights into protein-polysorbate interactions<br />
analysed by means <strong>of</strong> isothermal titration<br />
and differential scanning calorimetry, Eur.<br />
Biophys. J. 38 (2009) 557-68.<br />
20. A. Hawe, M. Sutter, W. Jiskoot, Extrinsic<br />
fluorescent dyes as tools <strong>for</strong> protein<br />
characterization, Pharm. Res. 25 (2008)<br />
1487-99.<br />
21. D.L. Sackett, J. Wolff, Nile red as a polaritysensitive<br />
fluorescent probe <strong>of</strong> hydrophobic<br />
protein surfaces, Anal. Biochem. 167 (1987)<br />
228-34.<br />
22. T. Ueda, M. Nagata, T. Imoto, Aggregation<br />
and chemical reaction in hen lysozyme<br />
caused by heating at pH 6 are depressed by<br />
osmolytes, sucrose and trehalose, J. Biochem<br />
130 (2001) 491-6.<br />
23. M.M. Santoro, Y. Liu, S.M. Khan, L.X. Hou,<br />
D.W. Bolen, Increased thermal stability<br />
<strong>of</strong> proteins in the presence <strong>of</strong> naturally<br />
occurring osmolytes, Biochemistry 31<br />
(1992) 5278-83.
summary, concludIng remarks<br />
and global perspectIve
&<br />
138<br />
summary<br />
The general objective <strong>of</strong> this thesis is to develop liquid <strong>for</strong>mulations <strong>for</strong> polypeptide based<br />
medicines that are stable under tropical condition.<br />
Chapter 2 <strong>of</strong> this thesis discusses peptide instabilities in aqueous solutions and strategies<br />
to improve peptide stability in parenteral <strong>for</strong>mulations. The most prevalent degradation<br />
pathways at accelerated temperature and certain pH are hydrolysis including deamidation,<br />
isomerization and racemization, oxidation including light and metal-induced oxidation,<br />
β-elimination, disulfide exchange, dimerization and further aggregation. Adequate<br />
knowledge about amino acid residues involved in peptide degradation can be used to<br />
develop strategies to improve peptide stability in aqueous solutions. The most common<br />
approaches to protect peptide degradation in aqueous <strong>for</strong>mulations are <strong>for</strong>mulating peptide<br />
at an optimum pH, the proper choice <strong>of</strong> buffers, antioxidants, chelating agents, removal <strong>of</strong><br />
oxygen, protection from light, the use <strong>of</strong> metal ions, organic solvents, or surfactants.<br />
The focus <strong>of</strong> this thesis is on the stabilization <strong>of</strong> oxytocin (a nonapeptide) in <strong>for</strong>mulations<br />
<strong>for</strong> injection. The major approach has been to investigate the stabilization <strong>of</strong> this peptide<br />
by divalent metal ions in combination with a certain buffer. The effect <strong>of</strong> monovalent and<br />
divalent metal ions in combination with buffers at a pH <strong>of</strong> 4.5 on the stability <strong>of</strong> oxytocin<br />
in aqueous solutions is described in Chapter 3. The chloride salts <strong>of</strong> monovalent metal<br />
ions (Na + and K + ) and divalent metal ions (Ca 2+ , Mg 2+ , and Zn 2+ ) in combination with<br />
citrate or acetate buffer were investigated. Utilizing RP-HPLC and HP-SEC the effect <strong>of</strong><br />
combinations <strong>of</strong> buffers and metal ions on the stability <strong>of</strong> aqueous oxytocin solutions after<br />
4 weeks <strong>of</strong> storage at either 4°C or 55°C was quantified. Addition <strong>of</strong> the monovalent ions,<br />
Na + and K + , to acetate- or citrate-buffered solutions did not increase oxytocin stability, nor<br />
did the addition <strong>of</strong> divalent metal ions to acetate buffered solutions. However, the stability<br />
<strong>of</strong> oxytocin in aqueous <strong>for</strong>mulations was improved in the presence <strong>of</strong> 5 and 10 mM citrate<br />
buffer in combination with at least 2 mM CaCl 2 , MgCl 2 , or ZnCl 2 and depended on the<br />
divalent metal ion concentration. Isothermal titration calorimetric (ITC) measurements<br />
were predictive <strong>for</strong> the stabilization effects observed during the stability study. Formulations<br />
in citrate buffer that had an improved stability displayed a strong interaction between<br />
oxytocin and Ca 2+ , Mg 2+ , or Zn 2+ , whereas <strong>for</strong>mulations in acetate buffer did not. In<br />
conclusion, our study showed that divalent metal ions in combination with citrate buffer<br />
strongly improved the stability <strong>of</strong> oxytocin in aqueous solutions.<br />
In chapter 4, various degradation products <strong>of</strong> oxytocin in citrate-buffered solution<br />
after thermal stress at a temperature <strong>of</strong> 70°C <strong>for</strong> 5 days were identified. The changes in<br />
degradation pattern caused by the presence <strong>of</strong> divalent metal ions were also determined.<br />
Degradation products <strong>of</strong> oxytocin in the citrate buffered <strong>for</strong>mulations with and without<br />
divalent metal ions were analyzed using liquid chromatography−mass spectrometry/mass<br />
spectrometry (LC−MS/MS). In the presence <strong>of</strong> divalent metal ions, almost all degradation<br />
products, in particular citrate adduct, tri- and tetrasulfide, and dimers, were greatly reduced<br />
in intensity. Formulations containing divalent metal ions in combination with citrate buffer<br />
suppressed the <strong>for</strong>mation <strong>of</strong> degradation products N-citryl oxytocin, tri/tetrasulfide and<br />
dimers. The suppression occurred on the disulfide bridge between Cys 1 and Cys 6 . Cysteine is<br />
susceptible to oxidation and β-elimination, and degradation <strong>of</strong> oxytocin involving cysteine<br />
leads to dimerization, <strong>for</strong>mation <strong>of</strong> tri/tetrasulfide and β-elimination followed by thioether<br />
<strong>for</strong>mation. No significant difference in the stabilizing effects was found among the<br />
divalent metal ions Ca 2+ , Mg 2+ , and Zn 2+ , suggesting that divalency is the most important
property <strong>of</strong> the metal contributing to stabilization <strong>of</strong> the oxytocin-metal-citrate cluster. The<br />
suppressed degradation pathways all involve the cysteine residues. We there<strong>for</strong>e postulate<br />
that cysteine-mediated intermolecular reactions are suppressed by complex <strong>for</strong>mation <strong>of</strong><br />
the divalent metal ion and citrate with oxytocin, thereby inhibiting the <strong>for</strong>mation <strong>of</strong> citrate<br />
adducts and reactions <strong>of</strong> the cysteine thiol group in oxytocin. Citrate has two opposite<br />
effects on oxytocin stability. First, it is reactive itself and can attack the N-terminal amino<br />
group from the cysteine residue to <strong>for</strong>m an adduct. Secondly, it protects oxytocin from<br />
degradation in the presence <strong>of</strong> divalent metal ions.<br />
In chapter 5, the effect <strong>of</strong> divalent metal ions (Ca 2+ , Mg 2+ and Zn 2+ ) on the stability <strong>of</strong><br />
oxytocin in aspartate buffer (pH 4.5) is described and their interaction with the peptide<br />
in aqueous solution was determined. RP-HPLC and HP-SEC results indicated that after 4<br />
weeks <strong>of</strong> storage at 55°C all tested divalent metal ions improved the stability <strong>of</strong> oxytocin in<br />
aspartate buffered solutions. However, the stabilizing effects <strong>of</strong> Zn 2+ were by far superior<br />
compared to both Ca 2+ and Mg 2+ . LC-MS/MS results showed that the combination <strong>of</strong><br />
aspartate and Zn 2+ in particular suppressed the <strong>for</strong>mation <strong>of</strong> peptide dimers. As shown by<br />
ITC, Zn 2+ interacted with oxytocin in the presence <strong>of</strong> aspartate buffer while Ca 2+ or Mg 2+ did<br />
not. In conclusion, the stability <strong>of</strong> oxytocin in the aspartate buffered-solution is strongly<br />
improved by the presence <strong>of</strong> zinc ions, and the stabilization effect is correlated with the<br />
ability <strong>of</strong> the divalent metal ions in aspartate buffer to interact with oxytocin. Aspartate<br />
presumably sequesters the zinc ion inside the ring structure <strong>of</strong> oxytocin protecting the<br />
intramolecular disulfide bridge from reactions leading to degradation.<br />
In chapter 6, the con<strong>for</strong>mation <strong>of</strong> oxytocin in aspartate buffer in the presence <strong>of</strong> Mg 2+<br />
or Zn 2+ , was investigated by 2D NMR spectroscopy, i.e. NOESY, TOCSY, 1 H- 13 C HSQC<br />
and 1 H- 15 N HSQC. Almost all 1 H, 13 C and 15 N resonance could be assigned using HSQC<br />
spectroscopy <strong>of</strong> oxytocin with neither 13 C nor 15 N enrichment. 1 H- 13 C and 1 H- 15 N HSQC<br />
spectra showed that Zn 2+ caused changes in the chemical shifts <strong>of</strong> almost all amino acid<br />
residues <strong>of</strong> oxytocin, whereas Mg 2+ only induces minor chemical shift changes in several but<br />
not all amino acid residues. On the other hand, NOESY spectra exhibited almost the same<br />
NOEs in the presence and absence <strong>of</strong> Zn 2+ . Our findings indicate the carboxylate group <strong>of</strong><br />
aspartate neutralizes the positive charge <strong>of</strong> the N terminus <strong>of</strong> Cys 1 , allowing the interactions<br />
with Zn 2+ to become more favorable. These interactions may explain the protection <strong>of</strong> the<br />
disulfide bridge against intermolecular reactions that lead to dimerization and inactivation.<br />
Chapter 7, describes a study on the ability <strong>of</strong> various extremolytes to stabilize model<br />
proteins (lysozyme and insulin) in aqueous solution. The effects <strong>of</strong> the extremolytes,<br />
betaine, hydroxyectoine, trehalose, ectoine, and firoin on the stability <strong>of</strong> lysozyme were<br />
determined by Nile red Fluorescence Spectroscopy and a bioactivity assay. Insulin stability<br />
was determined by RP-HPLC and HP-SEC. The effects <strong>of</strong> extremolytes on the unfolding<br />
temperature <strong>of</strong> the proteins were analyzed using a thermal shift assay <strong>for</strong> lysozyme and liquid<br />
differential scanning microcalorimetry <strong>for</strong> insulin. The interaction between extremolytes<br />
and proteins was studied by ITC. During storage at 70°C <strong>for</strong> 10 min, firoin protected<br />
lysozyme against inactivation better than the other extremolytes. During storage at 55°C<br />
<strong>for</strong> 4 weeks, firoin also acted as a stabilizer, however, betaine, hydroxyectoine, trehalose<br />
and ectoine, destabilized lysozyme. These findings indicate that some extremolytes can<br />
stabilize proteins under certain stress conditions but destabilize the same proteins under<br />
other stress conditions. The increased stability caused by firoin is explained by the observed<br />
increased unfolding temperature <strong>of</strong> lysozyme in the presence <strong>of</strong> firoin. After storage at 40°C<br />
SuMMary, concludInG reMarkS and Global perSpectIve<br />
&<br />
139
&<br />
140<br />
<strong>for</strong> 4 weeks, trehalose and ectoine protected insulin against degradation, while betaine,<br />
hydroxyectoine and in particular firoin were found to destabilize insulin. Furthermore,<br />
it was shown that firoin sharply decreased the unfolding temperature <strong>of</strong> insulin. The<br />
interaction <strong>of</strong> firoin with lysozyme and ectoine or trehalose with insulin was negligible<br />
as determined by ITC. Thus, firoin appears to be an excellent stabilizer <strong>for</strong> lysozyme but<br />
strongly destabilizes insulin, whereas ectoine showed the opposite behaviour. This study<br />
clearly shows that there is not one extremolyte that can acts as a universal stabilizer <strong>for</strong><br />
proteins in aqueous solution. There<strong>for</strong>e, to develop stable protein <strong>for</strong>mulations using<br />
extremolytes, one should consider the envisaged storage conditions besides screening the<br />
stabilizing effects <strong>of</strong> different extremolytes. Furthermore, <strong>for</strong> screening extremolytes <strong>for</strong> the<br />
stabilization <strong>of</strong> proteins, measuring the T m <strong>of</strong> the protein can be useful to predict stabilizing<br />
effects but only if the extremolyte induces a substantial change in the T m .<br />
concludIng remarks<br />
In this thesis an attempt is made to develop a heat-stable liquid oxytocin <strong>for</strong>mulation. Several<br />
<strong>for</strong>mulations containing specific combinations <strong>of</strong> divalent metal ions with buffers were<br />
found to stabilize oxytocin in aqueous solution. Different analytical tools such as LC-MS<br />
(MS), ITC and NMR were utilized to gain in<strong>for</strong>mation on the mechanism <strong>of</strong> stabilization<br />
at a molecular level. However, it is still difficult to correlate the results with the mechanism<br />
by which buffers and divalent metal ions may contribute to the stabilization <strong>of</strong> oxytocin.<br />
Although the overall picture was rather complex and mechanisms varied between the<br />
different ions and buffers, it is clear that stabilization <strong>of</strong> the Cys 1,6 disulfide bridge in the<br />
oxytocin molecule is <strong>of</strong> paramount importance <strong>for</strong> stabilization <strong>of</strong> the peptide in aqueous<br />
solution. This was best illustrated in an NMR study showing that the carboxylate group <strong>of</strong><br />
aspartate neutralizes the positive charge <strong>of</strong> the N terminus <strong>of</strong> Cys 1 , allowing the interactions<br />
with Zn 2+ to become more favorable. These interactions may explain the protection <strong>of</strong> the<br />
disulfide bridge against intermolecular reactions that lead to dimerization and inactivation.<br />
Zn 2+ causes a shift in the con<strong>for</strong>mational equilibrium <strong>of</strong> oxytocin, which is evident from<br />
the observed changes in the chemical shifts <strong>of</strong> almost all resonances, but the exact nature<br />
<strong>of</strong> these changes is not clear yet. A final explanation <strong>of</strong> these observations must await a<br />
more detailed, dynamic description possibly from molecular dynamic simulations steered<br />
by distance- and angle-restraints derived from the NMR spectra.<br />
Molecular dynamic simulations could be an interesting approach to provide insight on<br />
a molecular level in the stabilization mechanisms <strong>of</strong> the different divalent metal ions and<br />
buffers. With this type <strong>of</strong> studies it may be elucidated which con<strong>for</strong>mational changes are<br />
caused by the different metal ions and how these con<strong>for</strong>mational changes are affected by the<br />
buffers in the solution. When furthermore a relation between the oxytocin con<strong>for</strong>mation and<br />
its stability can be established, it may be feasible to define an optimum combination <strong>of</strong> metal<br />
ion and buffer that keeps the oxytocin in the most stable con<strong>for</strong>mation. Another approach<br />
that may finally result in a stable oxytocin <strong>for</strong>mulation is to use the knowledge gained in this<br />
thesis in a more or less guided extensive trial-and-error development <strong>for</strong>mulation study.<br />
Cleavage <strong>of</strong> the disulfide bridge was found to be the major degradation pathway <strong>for</strong><br />
oxytocin, and certain carboxylate buffers were found to support complex <strong>for</strong>mation with<br />
certain divalent metal ions, which lead to stabilization. Especially the effects <strong>of</strong> different<br />
carboxylate buffers in combination with different divalent metal ions may be an interesting
topic in such a study. Many different carboxylate buffers exist and only three were tested in<br />
the work described in this thesis. The complexity <strong>of</strong> the role <strong>of</strong> the buffer is nicely illustrated<br />
by the dual role that was found <strong>for</strong> citrate, being both a stabilizer and an adduct <strong>for</strong>mer,<br />
the differences found <strong>for</strong> the different ions in aspartate buffer and the absence <strong>of</strong> these<br />
differences in citrate or acetate buffer. It is unknown whether other carboxylate buffers may<br />
have better stabilizing properties on oxytocin and how their effect can be further improved<br />
with different metal ions. In this respect it is interesting to mention here that a recent study<br />
in our group revealed that malonate buffer had a higher stabilizing effect than aspartate<br />
or citrate buffer. At this moment a real time stability study with a <strong>for</strong>mulation containing<br />
this buffer and zinc ions is ongoing. The six months results that were obtained so far are<br />
promising and this <strong>for</strong>mulation may be suitable <strong>for</strong> use in tropical developing countries.<br />
global perspectIve<br />
As described in the introduction <strong>of</strong> this thesis the starting point <strong>for</strong> our research was the<br />
insufficient access <strong>of</strong> mothers in the tropical developing countries to oxytocin, which leads<br />
to a yearly death rate <strong>of</strong> approximately 150.000 mothers. The availability <strong>of</strong> a suitable,<br />
af<strong>for</strong>dable, liquid oxytocin <strong>for</strong>mulation would potentially prevent their death. When starting<br />
the research we assumed that the fact that none <strong>of</strong> the current oxytocin <strong>for</strong>mulations could<br />
be stored outside the cold chain was a major reason <strong>for</strong> this problem and that a heat stable<br />
<strong>for</strong>mulation would prevent these deaths. This starting point and the results <strong>of</strong> our studies<br />
raise two major questions:<br />
1. Was a liquid heat stable oxytocin <strong>for</strong>mulation developed in this study?<br />
2. What is the best way <strong>for</strong>ward to reduce maternal death due to oxytocin insufficient<br />
availability in tropical regions, i.e. sub-Saharan Africa?<br />
Un<strong>for</strong>tunately, thus far we were not able to find a liquid-oxytocin <strong>for</strong>mulation which<br />
meets the standard requirement <strong>of</strong> the pharmacopoeia i.e. storage stability <strong>for</strong> at least one<br />
year at 40°C or two years at 30°C. However, the <strong>for</strong>mulations, which involve zinc ions and<br />
malonate buffers, did demonstrate promising results. But as long as real time stability data <strong>for</strong><br />
a period <strong>of</strong> at least one to two years are not available, definite conclusions cannot be drawn.<br />
From a technical point <strong>of</strong> view several adequate solutions to solve the immediate need<br />
<strong>for</strong> a heat stable oxytocin <strong>for</strong>mulation are available:<br />
1. In those places where a refrigerator is not available, storage <strong>of</strong> the current oxytocin<br />
<strong>for</strong>mulation at ambient conditions is feasible, as long as the expired material is simply<br />
replaced every six months.<br />
2. Introduce refrigerators on solar energy, in those places where they are currently not<br />
available.<br />
3. A stable freeze dried oxytocin <strong>for</strong>mulation (to be reconstituted be<strong>for</strong>e use) has been<br />
developed by us in collaboration with MSD and could be introduced immediately.<br />
4. The development <strong>of</strong> a stable spray dried oxytocin <strong>for</strong>mulation <strong>for</strong> pulmonary<br />
administration.<br />
5. The introduction <strong>of</strong> the developed <strong>for</strong>mulations involving zinc ions and malonate buffer<br />
is most likely the best solution. However, assessment on the real time stability <strong>for</strong> at least<br />
another six months (until November 2012) is required be<strong>for</strong>e drawing a definite conclusion.<br />
SuMMary, concludInG reMarkS and Global perSpectIve<br />
&<br />
141
&<br />
142<br />
When these five solutions are evaluated the first two solutions require funding to replace<br />
the oxytocin stocks every half a year and to install energy power systems.<br />
With regard to the third option, a fast introduction would not only depend on the<br />
willingness to accept the related increase in process costs and costs <strong>of</strong> goods. Much more<br />
important will be the willingness <strong>of</strong> the responsible authorities to accept such a product on<br />
their markets, as a replacement <strong>for</strong> the current product, without a full registration dossier.<br />
If such a dossier is to be prepared the costs will rise tremendously and they will surely be<br />
too high <strong>for</strong> any <strong>of</strong> the relevant countries in this project. It should in this respect also be<br />
realized that the proposed replacement from a scientific point <strong>of</strong> view would not have any<br />
relevant associated safety risk since it only replaces one parenteral <strong>for</strong>mulation <strong>for</strong> another<br />
parenteral <strong>for</strong>mulation.<br />
The fourth solution involves an alternative dosage <strong>for</strong>m. A pulmonary dry powder<br />
inhalation system may be a good alternative to the parenteral administration from a<br />
scientific point <strong>of</strong> view. But <strong>for</strong> this kind <strong>of</strong> alternatives there is a justified request <strong>for</strong> a full<br />
registration dossier showing equivalence in efficacy and safety with the current injection<br />
product. The development studies needed to fill such a dossier would take years and the<br />
costs would be unacceptably high. Based on these considerations any alternative that goes<br />
beyond a solution <strong>for</strong> injection (or a product that is to be reconstituted to such a solution)<br />
would require too much development time at the cost <strong>of</strong> too many unnecessary loss <strong>of</strong> life.<br />
The fifth solution may potentially be the most ideal solution, but also <strong>for</strong> this solution<br />
authorities should be willing to limit their requirements regarding the registration dossier.<br />
Finally, any solutions clearly require nothing more than good will and funding.<br />
Oxytocin, as we described, <strong>of</strong>fers the most apparent and economical solution, where the<br />
money should be spent on rather than <strong>for</strong> the costs <strong>of</strong> the circus <strong>of</strong> politicians and <strong>of</strong>ficials,<br />
who travel around the globe to discuss “the problem”.
samenvattIng, conclusIes,<br />
aanbevelIngen,<br />
en mondIaal perspectIef
&<br />
146<br />
samenvattIng<br />
De algemene doelstelling van dit proefschrift (beschreven in ho<strong>of</strong>dstuk 1) is om<br />
vloeibare farmaceutische <strong>for</strong>muleringen te ontwikkelen voor op polypeptides gebaseerde<br />
geneesmiddelen die stabiel zijn onder tropische omstandigheden.<br />
Ho<strong>of</strong>dstuk 2 bespreekt de instabiliteit van peptides in waterige oplossingen en<br />
mogelijkheden om de stabiliteit van peptides in parenterale <strong>for</strong>muleringen te verbeteren.<br />
Hydrolyse, oxidatie en aggregatie zijn de meest voorkomende degradatieprocessen die<br />
in versneld houdbaarheidsonderzoek (verhoogde temperatuur) met peptides worden<br />
waargenomen. Hydrolyse kan leiden tot deamidatie, isomerisatie en racemerisatie.<br />
Oxdatie is vaak geïnduceerd en gekatalyseerd door licht en sporen metaalionen en kan<br />
leiden tot β-eliminatie, disulfide uitwisseling en dimerisatie. Gedegen kennis over de<br />
aminozuurresiduen die betrokken zijn bij de degradatieprocessen van peptides kan worden<br />
gebruikt om strategieën te ontwikkelen om de stabiliteit van peptides in waterige oplossingen<br />
te verbeteren. Er zijn verschillende mogelijkheden om de stabiliteit van peptides in waterige<br />
<strong>for</strong>muleringen te verbeteren. Hiertoe behoren onder andere het bepalen van de optimale<br />
pH, verwijdering van zuurst<strong>of</strong>, bescherming tegen licht en de optimale toepassing van<br />
hulpst<strong>of</strong>fen in de <strong>for</strong>mulering zoals buffers, anti-oxidanten, chelatiemiddelen, metaalionen,<br />
organisch oplosmiddelen, en oppervlakte-actieve st<strong>of</strong>fen.<br />
De focus van dit proefschrift ligt op de stabilisatie van oxytocine (een nonapeptide) in<br />
een <strong>for</strong>mulering voor injectie. Met name onderzochten we het stabiliseren van oxytocine<br />
door het combineren van verschillende metaalionen en buffers.<br />
In ho<strong>of</strong>dstuk 3 beschrijven we het effect van monovalente en divalente metaalionen in<br />
combinatie met buffers bij pH 4,5 op de stabiliteit van oxytocine in waterige oplossingen. De<br />
chloridezouten van monovalente (Na + en K + ) en divalente metaalionen (Ca 2+ , Mg 2+ , en Zn 2+ )<br />
in combinatie met citraat- <strong>of</strong> acetaatbuffer werden onderzocht. Het effect van combinaties<br />
van buffer met metaalionen op de stabiliteit van waterige oxytocine-oplossingen na 4 weken<br />
bewaren bij 4°C <strong>of</strong> 55°C is gekwantificeerd met behulp van twee analytische technieken:<br />
Reversed-Phase High Per<strong>for</strong>mance Liquid Chromatography (RP-HPLC) en High-Per<strong>for</strong>mance<br />
Size exclussion Chromatography (HP-SEC). Het toevoegen van de monovalente metaalionen,<br />
Na + en K + aan acetaat- <strong>of</strong> citraat-gebufferde oplossingen en het toevoegen van de divalente<br />
metaalionen Ca 2+ , Mg 2+ , en Zn 2+ aan acetaat- gebufferde oplossingen leidde niet tot een<br />
verbetering van stabiliteit van oxytocine. De stabiliteit in waterige <strong>for</strong>muleringen werd<br />
echter wel sterk verbeterd door in een 5 <strong>of</strong> 10 mM citraatbuffer in combinatie met ten minste<br />
2 mM CaCl 2 , MgCl 2 , or ZnCl 2 en was afhankelijk van de concentratie divalente metaalionen.<br />
Isothermal titration calorimetry (ITC) bleek een techniek te zijn met voorspellende waarde<br />
voor de waargenomen effecten. In <strong>for</strong>muleringen in citraatbuffer met een sterk verbeterde<br />
stabiliteit, werd een sterke interactie gevonden tussen oxytocine en Ca 2+ , Mg 2+ , <strong>of</strong> Zn 2+ . In<br />
<strong>for</strong>muleringen in acetaat buffer vonden we dit niet. Samenvattend kunnen we zeggen dat<br />
de combinatie van divalente metaalionen met citraatbuffer de stabiliteit van oxytocine in<br />
waterige oplossingen sterk verbeterde.<br />
Ho<strong>of</strong>dstuk 4 beschrijft een onderzoek waarin de degradatieproducten van<br />
oxytocine<strong>for</strong>muleringen in citraatbuffer werden geïdentificeerd in de aan- en afwezigheid<br />
van divalente metaalionen, na 5 dagen blootstelling aan een temperatuur van 70°C. De<br />
degradatieproducten werden geanalyseerd met behulp van vloeist<strong>of</strong>chromatografie-massa<br />
spectrometrie/massa-spectrometrie (LC-MS/MS). In de aanwezigheid van divalente<br />
metaalionen werd de vorming van bijna alle degradatieproducten, in het bijzonder van
citraatadduct, tri-en tetrasulfide en dimeren, sterk verminderd. Dit werd vooral veroorzaakt<br />
door stabilisatie van de disulfidebrug tussen de aminozuren Cys 1 en Cys 6 . Cysteine is gevoelig<br />
voor oxidatie en β-eliminatie, en bij de degradatiereacties van oxytocine waarbij cysteine<br />
betrokken was leidde dit tot de vorming van dimeren, tri- en tetrasulfide en thio-ether. We<br />
vonden geen substantieel verschil tussen het stabiliserende effect van de divalente metaalionen,<br />
Ca 2+ , Mg 2+ , en Zn 2+ . Dit suggereert dat divalentie de meest belangrijke eigenschap van het<br />
metaalion is die leidt tot de stabilisatie van het oxytocine-metaal-citraatcluster. Daarom<br />
veronderstellen we dat cysteine-gemedieerde intermoleculaire reacties worden geremd door<br />
de vorming van een complex tussen de divalente metaalionen, citraat en oxytocine. Citraat<br />
heeft twee tegenovergestelde effecten op de stabiliteit van ocytocine. Enerzijds kan het met de<br />
N-terminale aminogroep van cysteineresidue reageren waardoor een adduct wordt gevormd.<br />
Anderzijds kan citraat in aanwezigheid van divalente metaalionen oxytocine juist stabiliseren.<br />
In Ho<strong>of</strong>dstuk 5 wordt het effect van divalent metaalionen (Ca 2+ , Mg 2+ en Zn 2+ ) op de<br />
stabiliteit van oxytocine in aspartaatbuffer (pH 4,5) beschreven. Met behulp van RP-HPLC<br />
en HP-SEC metingen toonden we aan dat na 4 weken bewaren bij 55°C de combinatie<br />
van alle drie de geteste divalente metaalionen en aspartaatbuffer, leidde tot een verbeterde<br />
stabiliteit van oxytocine. De stabiliserende werking van Zn 2+ was veel beter dan van Ca 2+ <strong>of</strong><br />
Mg 2+ . LC-MS/MS resultaten toonden aan dat de combinatie van aspartaat en Zn 2+ vooral de<br />
vorming van dimeren tegenging. Uit ITC metingen bleek dat Zn 2+ in de aanwezigheid van<br />
aspartaatbuffer interacties aangaat met oxytocine terwijl dit met Ca 2+ <strong>of</strong> Mg 2+ niet het geval<br />
was. Samenvattend kan worden gezegd dat de stabiliteit van oxytocine in de met aspartaat<br />
gebufferde oplossing sterk kan worden verbeterd door de aanwezigheid van zinkionen,<br />
en dat dit effect correleert met het vermogen tot interactie tussen de metaalionen en<br />
het oxytocine. Waarschijnlijk wordt de disulfidebrug gestabiliseerd doordat aspartaat<br />
complexering van zinkionen in de ringstructuur van oxytocine mogelijk maakt.<br />
In Ho<strong>of</strong>dstuk 6 werd de con<strong>for</strong>matie van oxytocine in aspartaatbuffer in aanwezigheid<br />
van Mg 2+ <strong>of</strong> Zn 2+ , onderzocht met behulp van een aantal 2D-NMR technieken: NOESY,<br />
TOCSY, 1 H- 13 C HSQC en 1 H- 15 N HSQC. In het HSQC-spectra van oxytocine dat niet was<br />
verrijkt met 13 C <strong>of</strong> 15 N konden bijna alle 1 H, 13 C and 15 N resonanties worden toegewezen.<br />
Uit deze spectra bleek dat Zn 2+ veranderingen van de chemical shifts van vrijwel alle<br />
negen aminozuren van oxytocine veroorzaakt, terwijl Mg 2+ slechts geringe veranderingen<br />
in chemical shifts in sommige aminozuren induceert. Anderzijds vertoonden de NOESYspectra<br />
nagenoeg dezelfde NOEs in de aan- en afwezigheid van Zn 2+ . Dit duidt erop aan<br />
dat de carboxylaatgroep van aspartaat de positieve lading van de N-terminus van Cys 1<br />
neutraliseert, waardoor de interactie met Zn 2+ gunstiger wordt. Deze interacties kunnen de<br />
bescherming van de disulfide brug in oxytocine verklaren.<br />
In Ho<strong>of</strong>dstuk 7 bestuderen we de stabilisatie van twee modeleiwitten (lysozyme en<br />
insuline) in waterige oplossing met behulp van verschillende extremolieten. Het effect van<br />
de extremolieten, betaine hydroxyectoine, trehalose, ectoine en firoine op de stabiliteit van<br />
lysozyme werd bepaald met behulp van Nile Red fluorescentiespectroscopie en het bepalen<br />
van de biologische activiteit van het enzym. De stabiliteit van insuline werd bepaald met behulp<br />
van RP-HPLC en HP-SEC. Het effect van extremolieten op de ontvouwingstemperatuur (Tm)<br />
van de eiwitten werd geanalyseerd met behulp van een thermal shift assay voor lysozyme<br />
en liquid differential scanning microcalorimetry voor insuline. Tijdens incubatie bij 70°C<br />
gedurende 10 minuten werd lysozyme beter gestabiliseerd door fiorine dan door de andere<br />
geteste extremolieten. Ook tijdens de bewaren bij 55°C gedurende 4 weken, stabiliseerde<br />
SaMenvattInG, concluSIeS, en MondIaal perSpectIef<br />
&<br />
147
&<br />
148<br />
fiorine het eiwit. Betaine, hydroxyectoine, trehalose en ectoine bleken lysozyme onder deze<br />
condities echter te destabiliseren. Deze uitkomsten geven aan dat sommige extremolieten<br />
eiwitten kunnen stabiliseren onder bepaalde stress-omstandigheden, terwijl dezelfde eiwitten<br />
juist worden gedestabiliseerd onder andere stress-omstandigheden. De verbeterde stabiliteit<br />
door firoine kan worden verklaard door de waargenomen verhoogde T m van lysozyme in de<br />
aanwezigheid van firoine. Na 4 weken bewaren bij 40°C bleek insulin te worden gestabiliseerd<br />
door trehalose en ectoine, terwijl het eiwit juist gedestabiliseerd werd door betaine,<br />
hydroxyectoine en met name door firoine. Verder vonden we dat de Tm van insuline sterk<br />
afnam in de aanwezigheid van firoine. De interactie van firoine met lysozyme en ectoine<br />
<strong>of</strong> de interactie van trehalose met insuline zoals bepaald met ITC was verwaarloosbaar.<br />
Samenvattend kunnen we zeggen dat firoine een uitstekende stabilisator is voor lysozyme, maar<br />
daarintegen een destabilisator is voor insuline, terwijl ectoine het tegenovergestelde gedrag<br />
vertoont. Uit onze studie blijkt duidelijk dat er geen extremoliet is die kan fungeren als een<br />
universele stabilisator voor eiwitten in waterige oplossing. Om een stabiele eiwit<strong>for</strong>mulering<br />
met extremolieten te ontwikkelen, moet daarom rekening worden gehouden met de beoogde<br />
bewaarcondities. Voor het screenen van extremolieten op hun vermogen tot stabilisatie van<br />
eiwitten kan het nuttig om de T m van de eiwitoplossing met en zonder extremoliet te bepalen,<br />
maar alleen als de extremoliet een substantiële verandering in de T m veroorzaakt.<br />
conclusIes en aanbevelIngen<br />
Dit proefschrift beschrijft onderzoek dat gericht is op het ontwikkelen van een thermostabiele<br />
waterige oxytocine<strong>for</strong>mulering. Diverse combinaties van divalente metaalionen met buffers<br />
bleken oxytocine in waterig milieuw te kunnen stabiliseren. Verschillende analysemethoden,<br />
zoals LC-MS (MS), ITC en NMR, werden toegepast om het mechanisme van stabilisatie op<br />
moleculair niveau op te helderen. Het is echter nog lastig om de resultaten hiervan direct te<br />
correleren aan het exacte mechanisme dat aan de stabilisatie ten grondslag ligt.<br />
Hoewel het algemene beeld vrij complex was en het mechanisme voor de verschillende<br />
gebruikte ionen en buffers anders, is wel duidelijk dat de stabilisatie van de Cys 1,6<br />
disulfidebrug van het oxytocinemolecuul van essentieel belang is voor het stabiliseren van<br />
het peptide in waterige oplossing. Dit wordt het beste geïllustreerd in het NMR-onderzoek<br />
waaruit blijkt dat de carboxylaatgroep van aspartaat de positieve lading van N-terminus<br />
van Cys 1 neutraliseert, waardoor de interactie met zinkionen beter wordt. Deze interactie<br />
verklaart de bescherming van de disulfidebrug tegen intermoleculaire reacties die resulteren<br />
in dimerisatie en inactivatie.<br />
Zn 2+ veroorzaakt een verandering van de con<strong>for</strong>matie van oxytocine, hetgeen blijkt uit<br />
van de veranderingen in de chemical shifts van bijna alle aminozuren. De precieze aard van<br />
deze veranderingen is echter nog niet helemaal duidelijk maar kan mogelijk opgehelderd<br />
worden met een dynamische beschrijving van moleculaire dynamische simulaties gestuurd<br />
door afstand- en hoek-beperkingen die afgeleid kunnen worden uit NMR spectra.<br />
Het toepassen van moleculair dynamische simulaties kan een interessante benadering<br />
zijn om inzicht te krijgen in het stabilisatiemechanisme van de divalente metaalionen<br />
en de buffers op moleculair niveau. Bij dit type studies kan worden opgehelderd welke<br />
con<strong>for</strong>matieveranderingen worden veroorzaakt door de verschillende metaalionen en<br />
hoe deze con<strong>for</strong>matieveranderingen worden beïnvloed door de buffers in de oplossing.<br />
Wanneer er ook een relatie is tussen de con<strong>for</strong>matie van oxytocine en de stabiliteit, is het
mogelijk om een optimale combinatie van metaalionen en buffer te bepalen die leidt tot de<br />
meest stabiele con<strong>for</strong>matie van oxytocine. Een andere aanpak die uiteindelijk zou kunnen<br />
resulteren in een stabiele oxytocine<strong>for</strong>mulering is de kennis uit dit proefschrift te gebruiken<br />
in meer <strong>of</strong> minder door ‘trial and error’ geleide <strong>for</strong>muleringsstudie.<br />
Splitsing van de disulfidebrug bleek de belangrijkste degradatieroute voor oxytocine te<br />
zijn. Bepaalde carboxylaatbuffers in combinatie met bepaalde divalente metaalionen bleken<br />
te comlexeren met oxytocine hetgeen leidde tot stabilisatie van de disulfidebrug. Vooral de<br />
effecten van dit soort combinaties kunnen een interessant onderwerp voor verdere studies. Er<br />
bestaan veel verschillende carboxylaatbuffers en slecht drie werden onderzocht in het werk<br />
dat beschreven is in dit proefschrift. De complexiteit van de rol van de buffer wordt goed<br />
geïllustreerd door de dubbele rol die citraat kan spelen (stabilisator en adduc vormer), door<br />
de diverse effecten van de verschillende divalente metaalionen in aspartaatbuffer, en door de<br />
verschillen in effect van citraat- en acetaatbuffer in combinatie met divalente metaalionen.<br />
Het is onbekend <strong>of</strong> andere carboxylaatbuffers oxytocine beter kunnen stabiliseren en <strong>of</strong><br />
de stabiliteit verder kan worden verbeterd met andere metaalionen. In dit verband is het<br />
interessant om te vermelden dat uit een recente studie binnen onze onderzoeksgroep<br />
gebleken is dat malonaatbuffer in combinatie met zinkionen oxytocine beter stabiliseert dan<br />
aspartaat- en citraatbuffers in combinatie met zinkionen. Op dit moment is een real-time<br />
stabiliteitsstudie met <strong>for</strong>muleringen die gebaseerd zijn op malonaatbuffer en zinkionen<br />
gaande. De resultaten na een bewaartermijn van zes maanden zijn veelbelovend waardoor<br />
deze <strong>for</strong>mulering geschikt zou kunnen zijn voor gebruik in tropische ontwikkelingslanden.<br />
mondIaal perspectIef<br />
Zoals beschreven in de inleiding van dit proefschrift was de drijfveer voor ons onderzoek<br />
onvoldoende toegang in de tropische ontwikkelingslanden tot oxytocine, hetgeen leidt tot<br />
een jaarlijks sterfte van ongeveer 150.000 moeders. De beschikbaarheid van een geschikte<br />
vloeibare oxytocine<strong>for</strong>mulering zou mogelijk hun dood kunnen voorkomen. Bij de start<br />
van het onderzoek zijn we ervan uitgegaan dat de belangrijkste reden voor dit probleem<br />
het feit is dat geen van huidige oxytocine<strong>for</strong>muleringen kan worden opgeslagen buiten<br />
de “cold-chain” en dat er met een warmte-stabiele <strong>for</strong>mulering veel sterfgevallen zouden<br />
kunnen worden voorkomen. Dit uitgangspunt en de resultaten van ons onderzoek leidt tot<br />
twee belangrijke vragen:<br />
1. Is er in dit onderzoek een warmte-stabiele vloeibare oxytocine ormulering ontwikkeld?<br />
2. Wat is de beste manier om maternale sterfte als gevolg van een gebrek aan oxytocine in<br />
de tropen, bijvoorbeeld sub-Sahara Afrika, te verminderen?<br />
Helaas zijn we in onze studie niet in staat geweest om een vloeibare oxytocine<strong>for</strong>mulering<br />
te vinden, die voldoet aan de standaardeis van veel farmacopees, zoals stabiliteit van ten<br />
minste één jaar bij 40°C <strong>of</strong> twee jaar bij 30°C. De <strong>for</strong>muleringen gebaseerd op de combinatie<br />
van zinkionen en malonaatbuffer zijn echter wel veelbelovend. Maar zolang real-time<br />
stabiliteitsgegevens over een periode van ten minste één <strong>of</strong> twee jaar niet beschikbaar zijn,<br />
kunnen er nog geen definitieve conclusies worden getrokken.<br />
Vanuit technologisch oogpunt is er een aantal strategieën waarmee de onmiddellijke<br />
behoefte aan een warmtebesteding oxytocine<strong>for</strong>mulering op adequate wijze opgelost zou<br />
kunnen worden:<br />
SaMenvattInG, concluSIeS, en MondIaal perSpectIef<br />
&<br />
149
&<br />
150<br />
1. Op plaatsen waar geen koelkast beschikbaar is, is het opslaan van de huidige<br />
conventionele oxytocine<strong>for</strong>muleringen bij omgevingsomstandigheden mogelijk, op<br />
voorwaarde dat ze elke zes maanden worden vervangen.<br />
2. De introductie van koelkasten op zonne-energie op die plaatsen waar koelkasten op dit<br />
moment niet voorhanden zijn.<br />
3. Een warmtestabiele gevriesdroogde oxytocine<strong>for</strong>mulering (te reconstitueren vlak voor<br />
gebruik) is door ons ontwikkeld in samenwerking met MSD en kan direct worden<br />
toegepast.<br />
4. De ontwikkeling van een stabiele gesproeidroogde oxytocine<strong>for</strong>mulering voor<br />
pulmonale toediening.<br />
5. De introductie van de eerder genoemde <strong>for</strong>mulering gebaseerd op een combinatie<br />
van zinkionen en malonaatbuffer is waarschijnlijk de beste oplossing. Een definitieve<br />
conclusie over de geschiktheid van deze <strong>for</strong>mulering kan pas worden getrokken<br />
na afronding van een real-time stabiliteitsstudie van twee jaar. De resultaten van dit<br />
onderzoek zullen in november 2012 beschikbaar komen.<br />
Als we deze vijf mogelijke oplossingen voor het probleem evalueren, kunnen we<br />
concluderen dat voor de eerste twee alleen voldoende financiële middelen beschikbaar<br />
moeten komen om de oxytocinevoorraden elke half jaar te vervangen <strong>of</strong> om koelkasten op<br />
zonne-energie aan te schaffen.<br />
Een snelle introductie van een te reconstitueren gevriesdroogd poeder op de markt,<br />
de derde oplossing, is niet alleen afhankelijk van de bereidheid van de industrie om<br />
een stijging van proceskosten en materiaalkosten te accepteren. Veel belangrijker is de<br />
bereidheid van de verantwoordelijke autoriteiten om een dergelijk product op hun markt<br />
te toe te laten, ter vervanging van het huidige conventionele product, zonder dat daar<br />
een volledig registratiedossier voor aangelegd hoeft te worden. Als namelijk een dergelijk<br />
dossier moet worden voorbereid zullen de kosten enorm stijgen en zeker te hoog worden<br />
voor de betrokken landen. In dit verband moeten we ons realiseren dat de voorgestelde<br />
vervanging vanuit wetenschappelijk oogpunt niet tot relevante veiligheidsrisico’s leidt,<br />
aangezien alleen de huidige parenterale <strong>for</strong>mulering wordt vervangen door een andere<br />
parenterale <strong>for</strong>mulering.<br />
De vierde oplossing bestaat uit een alternatieve doseringsvorm. Vanuit wetenschappelijk<br />
oogpunt kan een droog poeder voor pulmonale toediening een goed alternatief zijn voor de<br />
parenterale toediening. Voor dit soort alternatieven is echter een volledig registratiedossier<br />
nodig waaruit de gelijkwaardigheid in werkzaamheid en veiligheid ten opzichte van het<br />
huidige injecteerbare product blijkt. De ontwikkelingsstudies die nodig zijn voor een<br />
dergelijk dossier kunnen jaren duren en de kosten zullen onaanvaardbaar hoog zijn.<br />
Op basis van deze overwegingen zal voor elk alternatief voor een injectie (<strong>of</strong> voor een<br />
reconstitueerbaar product) te veel ontwikkelstijd nodig zijn. Dit zal helaas ten koste gaan<br />
van te veel mensenlevens.<br />
De vijfde oplossing is mogelijk de meest ideale, maar ook hiervoor zullen autoriteiten<br />
bereid moeten zijn om hun eisen ten aanzien van het registratiedossier te beperken.<br />
Ten slotte, om het grote probleem op te lossen is niet meer nodig dan goede wil en financiering.<br />
De strategiën zoals we in het proefschrift hebben beschreven bieden de meest voor de hand<br />
liggende oplossing. Hier zou geld aan moeten worden besteed, in plaats van aan het circus van<br />
politici en ambtenaren die over de hele wereld reizen om “het probleem” te bespreken.
acknowledgements
&<br />
154<br />
Completing this thesis would not have been possible without the aid and support <strong>of</strong><br />
countless people over the past four years. I must first address the sincerest gratitude to<br />
Pr<strong>of</strong>essor Henderik W. Frijlink, my promotor, who allowed me to experience the life <strong>of</strong><br />
a PhD student. His involvement and crucial contribution has exceptionally inspired and<br />
enriched my growth as a scientist. I am grateful in every possible way and hope to keep up<br />
our collaboration in the future. To Dr. Wouter L. J. Hinrichs, my co-promotor, I am deeply<br />
indebted and thankful <strong>for</strong> being patience and relentless ef<strong>for</strong>t throughout my research<br />
and during the writing <strong>of</strong> this thesis, as well as providing unflinching encouragement<br />
and support in various ways. I would also like to thank Dr. Herman J. Woerdenbag and<br />
Dr. Gerad J. Bolhuis, who initiated the first contact that led to my PhD studentship in<br />
Groningen. As a Dutch Top Institute <strong>Pharma</strong> fellow, I also thank Louise Lammers and John<br />
den Engelsman, <strong>for</strong> their valuable suggestions.<br />
I was extraordinary <strong>for</strong>tunate in having Pr<strong>of</strong>essor Geny M.M. Groothuis, Pr<strong>of</strong>essor<br />
Arnold J.M. Driessen, and Pr<strong>of</strong>essor Wim Jiskoot as members <strong>of</strong> the reading committee, <strong>for</strong><br />
their constructive comments that significantly improved the content <strong>of</strong> this thesis.<br />
The completion <strong>of</strong> this thesis was made attainable with support and contribution from<br />
colleagues and friends, whom provided valuable discussion, asisstance with instruments,<br />
molecular modelling, and production process. On these, I shall address sincere thanks to<br />
Jean-Piere Amorij, Andrea Hawe, Robert Poole, Bazak Kukrer, Hjalmar Permentier, Alexej<br />
Kedrov, Angela Casini, Frans Mulder, Ruud Scheek, Alia Oktaviani, Peter van der Moelen,<br />
Jamshed Anwar, Hans de Waard, Vinay Saluja, Jan Fisher, Anko Eissen, Marinella Visser,<br />
Andy-Mark Thunnissen, Ali Rohman, Eni Ratnaningsih, Faizah Fulyani, Ryanto Boediono,<br />
M. Khalid, Alexej Kedrov, Annie van Dam, Ghea Schuurman, Syarif Riyadi, Caroline Visser,<br />
Hans van Doorne, Wangsa Tirta Ismaya, Wouter Tonis, and Jan Ettema.<br />
Collective and individual acknowledgment are also owed to my colleagues at the<br />
Department <strong>of</strong> <strong>Pharma</strong>ceutical Technology and Biopharmacy, University <strong>of</strong> Groningen,<br />
whose present refreshed, helpful and memorable at any time. Many thanks address to Paul,<br />
Dotie, Gieta, Doreenda, Peter, Stieneke, Lida, Fesia, Parinda, TT, Taufan, Senthil, Wouter T,<br />
Anne, Niels, Floris, Marcell, Elham, and Maarten. For Anne de Boer, to be my company on<br />
the trip to the hospital and cheering me up when I had the bike accident. Also thanks to the<br />
talented photographer BT, who provided nice picture <strong>for</strong> the thesis cover, which I am proud<br />
<strong>of</strong>. And to Milica, my “roomy”, it is really at my pleasure to have her as the paranymph.<br />
I also benefited by outstanding works from graduate and undergraduate students, whose<br />
works contributed to this thesis: Dewi, Jenny, Imma, Stashek, Ilja, Marriel, Ilvy, Esther,<br />
Erwin, and Marieke. My special thanks to Phillips, <strong>for</strong> his “<strong>Innovative</strong>” idea.<br />
To Henk, Kathy, Rika, Joke, Heleen, Anneke, Annete, Tim, Lisanne, Yvonne, Rijn and<br />
Sonja, I am thankful <strong>for</strong> their kind help with the administration works.<br />
I would also acknowledge the Rector and the Dean <strong>of</strong> the Faculty <strong>of</strong> <strong>Pharma</strong>cy <strong>of</strong> the<br />
University <strong>of</strong> Surabaya <strong>for</strong> granting the academic leave permit during my PhD study in<br />
Groningen, The Netherlands.<br />
Furthermore, I would also like to thank members <strong>of</strong> the Indonesian Students Association<br />
in Groningen which has given me the opportunity to involve as well as contribute in developing<br />
the organization. It is my wish that our splendid relationship would remain growing.<br />
Sincere thanks <strong>for</strong> my cousin, FS Widoyono and his family: Indah, Pandu, Pandji who<br />
helped me a lot with my daily routine. My companion in arms: Mariana, Kenzie, Ono, Neng,<br />
Insanu, Uyung, Adit, Wisnu, Adhi, Yota, Puri, Aramel, Muiz, Robby, Astri, Rahma, Lia,
Mutia, Shinta, Kadek, Wiwin, Arvie, and my beloved best friends: Fanny, Poppy, Awalia,<br />
Yayok, Desti, Yuli, Aini, Klara, Tara, Sita, Ismail, Reza, Pandji, Ratna, and many others,<br />
whom have been becoming part <strong>of</strong> my life during my PhD studentship. The joys and the<br />
sorrows we share certainly help me to learn how to become a better me.<br />
My parents deserve my highest gratitude, <strong>for</strong> their inseparable support and prayers.<br />
Spiritual bond and affection between us will never end at whatever cost, even when this life<br />
ends, someday. Dadang, Heru, and Ichwan, many thanks <strong>for</strong> being such a supportive and<br />
caring siblings. Also <strong>for</strong> the in-laws with their thoughtful support.<br />
Words fail me to express my appreciation to my little family: my beloved husband,<br />
Iskandar Zulkarnain, <strong>for</strong> his support, caring and gently love. My beloved daughter, Viny,<br />
my ray <strong>of</strong> sunshine in the cloudiest day, thank you <strong>for</strong> being always on my side. We are one<br />
and hand by hand together reaching our goals. I feel blessed that, despite my limited time<br />
as a Doctoral student, I am still allowed to spend enough time to accompany you finishing<br />
your Middle Years Program at International School <strong>of</strong> Groningen.<br />
Lastly, I would like to thank everyone who was important to the flourishing realization<br />
<strong>of</strong> this thesis, as well as expressing my request <strong>for</strong> <strong>for</strong>giveness that I would not be able to<br />
mention personally one by one.<br />
acknowledGeMentS<br />
&<br />
155