stability assessment of biopharmaceutical formulations · resumo o vasto uso de anticorpos com fins...

Stability Assessment of Biopharmaceutical Formulations

Joana Baltazar Domingues

Dissertation submitted for obtaining the Master’s degree in

Biological Engineering

Jury

President: Prof. Maria Raquel Murias dos Santos Aires Barros

Supervisor: Dr. Ana Margarida Nunes da Mata Pires de Azevedo

Co-Supervisor: Prof. José António Leonardo dos Santos

Examiner: Prof. Marília Clemente Velez Mateus

November 2011

ACKNOWLEDGEMENTS

First, I would like to thank my supervisor Dr. Ana Azevedo for giving me the opportunity to develop

this work and for her help throughout this process. It wouldn’t have been possible to conclude this

work without her knowledge and insightful ideas. I am profoundly grateful for her kindness and

constant support, availability and encouragement.

I would also like to express my gratitude to my co-supervisor, Professor José Santos for the

availability, as well as for the knowledge shared.

Special thanks to my laboratory colleagues for their warm welcoming into the group. I would like to

thank Luis Borlido for sharing with me some of his brightness and to José Oliveira for all the help he

gave me along this process.

To my family, the biggest thank for the interest, patience and support shown. To my father, thank

you for always trying to help me and for sharing your knowledge and insights. A special word to my

mother, for being the rock in my life and for inspiring me to always exceed myself. You represent

everything I one day hope to become.

I would like to thank my boyfriend, João Ferreira, for all the love, patience and motivation he gave

me at all times. Thank you for believing in me more than I believe in myself.

At last, but not least, I would like to thank all my friends for their friendship, as well as the support

and motivation that they have given me throughout this journey. I would like to pay a special

acknowledgement to my friends João Burgal and Joana Carmelo for the constant interest

demonstrated in my work, as well as for the motivation they have always given me.

iii

ABSTRACT

Antibodies are widely used for therapeutic purposes. A common problem associated with

therapeutic antibodies is to ensure their long term stability which constitutes a formulation challenge.

Therefore, in this thesis the stability of antibodies has been studied in the absence and in the presence

of common additives present in therapeutic formulations in order to identify which additive leads to

higher antibody stabilization.

First, IgG stability was evaluated against heat denaturation, pH induced denaturation and

mechanical shear stress induced degradation. Then, IgG stability in solutions with sucrose, glycine,

maltose, L-histidine and D-trehalose was tested against thermal and pH induced denaturation.

Aggregation studies were also conducted for different values of temperature and pH using dynamic

light scattering. The biological activity was determined by affinity chromatography and conformational

studies conducted by circular dichroism spectroscopy.

IgG was found to be resistant to shear stress up to 49 Pa induced in a concentric-cylinder shear

device and against induced denaturation at pH 3. At pH 2, irreversible denaturation occurred at some

extent. Regarding heat denaturation, IgG suffered only mild structural changes at 60°C but at 70°C

significant loss of biological activity, extensive structural changes and IgG precipitation were verified.

IgG showed higher tendency for aggregation upon incubation at 70°C than upon incubation at pH

values distant from the isoelectric range (pI ≈9).

Maltose was identified as the best IgG stabilizer since it provided complete stabilization at low pH

and the highest stabilization against heat denaturation. Generally, maltose was also able to reduce

IgG’s tendency for aggregation.

Keywords: Antibodies, Stability, Aggregation, Maltose, Circular Dichroism Spectroscopy, Dynamic

Light Scattering

v

RESUMO

O vasto uso de anticorpos com fins terapêuticos impõe que a sua estabilidade a longo prazo seja

assegurada, o que constitui um desafio no desenvolvimento da formulação terapêutica. Assim,

estudou-se a estabilidade do anticorpo na ausência e presença de aditivos comummente presentes em

formulações terapêuticas com o objectivo de identificar o aditivo que conduz a maior estabilização do

anticorpo.

Primeiramente avaliou-se a desnaturação da IgG induzida pela temperatura, pH e tensões de corte

de origem mecânica. A estabilidade da IgG em soluções contendo sacarose, glicina, maltose, L-histidina

e D-trealose foi testada em condições desnaturantes de temperatura e pH. Os efeitos da temperatura e

do pH na agregação foram estudados por dispersão dinâmica de luz. A actividade biológica foi

determinada por cromatografia de afinidade sendo os estudos conformacionais efectuados por

dicroísmo circular.

Verificou-se que a IgG consegue resistir a tensões de corte de 49 Pa induzidas num módulo de

cilindros concêntricos rotativo bem como à desnaturação induzida a pH 3. A pH 2 ocorreu

desnaturação irreversível em alguma extensão. A 60°C a IgG sofreu apenas leves mudanças

estruturais , mas a 70°C verificou-se perda significativa de actividade biológica, extensas alterações

estruturais e formação de precipitados. Verificou-se maior tendência para agregação após incubação

a 70°C do que após incubação a pH afastados do ponto isoeléctrico da IgG.

A maltose apresentou-se como o melhor estabilizador da IgG dado oferecer completa estabilização

a pH baixo e a melhor estabilização face à desnaturação induzida pela temperatura, contribuindo

também para a redução da tendência para agregação.

Palavras-chave: Anticorpos, Estabilidade, Agregação, Maltose, Dicroísmo Circular, Dispersão

Dinâmica de Luz

vii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ............................................................................................................. I

ABSTRACT ............................................................................................................................... III

RESUMO .................................................................................................................................. V

TABLE OF CONTENTS .............................................................................................................. VII

LIST OF FIGURES ...................................................................................................................... IX

LIST OF TABLES ...................................................................................................................... XV

LIST OF ABREVIATIONS .......................................................................................................... XVI

INTRODUCTION ....................................................................................................................... 1

1.1 BACKGROUND .........................................................................................................................1

1.2 OBJECTIVES .............................................................................................................................1

1.3 OUTLINE .................................................................................................................................2

2 LITERATURE REVIEW ........................................................................................................... 3

2.1 MONOCLONAL ANTIBODIES .......................................................................................................3

2.2 MABS’ HISTORICAL AND ECONOMICAL BACKGROUND ....................................................................7

2.3 STABILITY REQUIREMENTS .........................................................................................................9

2.4 PUBLISHED STUDIES ............................................................................................................... 10

2.5 DENATURATION CAUSED BY SHEAR STRESS ................................................................................ 13

2.5.1 Vorticular Flow Membrane Modules ........................................................................ 13

2.6 STABILITY ASSESSMENT .......................................................................................................... 15

2.6.1 Affinity Chromatography .......................................................................................... 15

2.6.2 Circular Dichroism ..................................................................................................... 17

2.6.3 Dynamic Light Scattering .......................................................................................... 23

3 MATERIALS AND METHODS ............................................................................................... 26

3.1 MATERIALS .......................................................................................................................... 26

3.1.1 Biologicals ................................................................................................................. 26

3.1.2 Chemicals .................................................................................................................. 26

3.2 METHODS ............................................................................................................................ 27

3.2.1 Sample Preparation .................................................................................................. 27

viii

3.2.1.1 Stability Assessment of IgG without additives ........................................... 27

3.2.1.2 Stability Assessment of IgG with additives................................................. 28

3.2.2 Analytical Methods.................................................................................................... 29

4 RESULTS AND DISCUSSION ................................................................................................. 31

4.1 STABILITY ASSESSMENT OF IGG WITHOUT ADDITIVES ................................................................... 31

4.1.1 Thermal induced denaturation ................................................................................. 31

4.1.2 pH induced denaturation .......................................................................................... 37

4.1.3 Mechanical Shear Stress induced degradation ......................................................... 40

4.2 STABILITY ASSESSMENT OF IGG WITH ADDITIVES ......................................................................... 44

4.2.1 Thermal induced denaturation ................................................................................. 44

4.2.2 pH induced denaturation .......................................................................................... 48

4.3 AGGREGATION STUDIES FOR IGG ............................................................................................. 50

4.3.1 pH effect on IgG aggregation .................................................................................... 51

4.3.2 Temperature effect on IgG aggregation .................................................................... 53

5 CONCLUSION AND FUTURE PRESPECTIVES ......................................................................... 57

REFERENCES ........................................................................................................................... 59

APPENDIX A - RELEVANT LITERATURE DATA ......................................................................... 63

APPENDIX B - RETAINED BIOLOGICAL ACTIVITY OVER TIME .................................................. 64

APPENDIX C - CD SPECTRA................................................................................................... 68

APPENDIX D - DLS SIZE DISTRIBUTIONS ................................................................................ 75

ix

LIST OF FIGURES

Figure 2.1 – IgG representation showing the light and heavy chains as well as the constant and variable

regions. Adapted from (6). .......................................................................................................................... 4

Figure 2.2 – Representation of IgG’s structure highlighting its several domains, as well as the Fab and Fc

fragments. The secondary structure of the Fab fragment is reptresented in detail (7). The Fv fragment

corresponds to an unstable fragment able to bind to an antigen with two V regions, VL and VH (2). ...... 5

Figure 2.3 – Representation of IgG’s secondary structure, evidencing the high content of β-sheets (8). . 6

Figure 2.4 – Schematic drawing of a CCSD, showing Taylor vortices (26). ...............................................14

Figure 2.5 – (A) Plane polarized light resolved into two circularly polarized components – left, L, and

right, R. As long as the intensities and phases of the two circularly polarized components remain equal,

their resultant will lie in a plane with oscillating magnitude. (B) If the right circularly polarized

component is less intense (more absorbed) than the left one, the electric vector of the light follows an

elliptical path, thus corresponding to elliptically polarized light (36). ......................................................18

Figure 2.6 – Typical CD spectra for particular secondary structural motifs (α-helix, β-sheet, random coil)

used in CD protein structure fitting programs (38). ..................................................................................21

Figure 4.1 – Retained Biological Activity (%) after 50 hours of incubation at 60°C of IgG solutions in PBS

(pH 7.4) with the initial concentrations of 1, 5, 10 and 15 g/L. ................................................................32

Figure 4.2 – Far-UV CD spectra of IgG in PBS (pH 7.4) before incubation () and after 50 hours of

incubation () at 60°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L. ....33

Figure 4.3 – Retained Biological Activity (%) after 2, 8 and 36 hours of incubation at 70°C of IgG

solutions in PBS (pH 7.4) with the initial concentrations of 1, 5, 10 and 15 g/L. ......................................34

Figure 4.4 – Precipitation occured after 8 hours of incubation at 70°C for the initial concentrations of

IgG in solution of: (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L. ..............................................................34

Figure 4.5 - Far-UV CD spectra of IgG in PBS (pH 7.4) before incubation () and after 8 hours of

incubation () at 70°C for the initial concentration of (A) 1g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L. .....36

Figure 4.6 – Retained Biological Activity (%) after 50 hours of incubation of IgG solutions with PBS at pH

3.02 with the initial concentrations of 1, 5, 10 and 15 g/L. ......................................................................37

Figure 4.7 - Far-UV CD spectra of IgG in PBS before incubation () and after 50 hours of incubation

() at pH 3.02 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L. .................38

Figure 4.8 - Retained Biological Activity (%) after 9 hours of incubation of IgG solutions with PBS at pH

2.1 with the initial concentrations of 1, 5, 10 and 15 g/L. ........................................................................39

x

Figure 4.9 - Far-UV CD spectra of IgG in PBS before incubation () and after 9 hours of incubation ()

at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L. .......................... 40

Figure 4.10 – Far-UV CD spectra of IgG in PBS at pH 7.4 for several periods of time spent in the shear

field. The initial concentration of IgG was of 1 g/L and the angular speed of the inner rotating cylinder

of 1000 rpm. ............................................................................................................................................. 41



of 2000 rpm. ............................................................................................................................................. 42



of 3000 rpm. ............................................................................................................................................. 42



of 3000 rpm. ............................................................................................................................................. 43

Figure 4.14 – Retained Biological Activity (%) of IgG after a period of circulation in the shear field of 30

minutes. .................................................................................................................................................... 44

Figure 4.15 - Retained Biological Activity (%) after incubation at 70°C for a period of 8h of IgG solutions

with the initial concentrations of 1, 5, 10 and 15 g/L. The buffers used consisted in PBS with 10% (w/v)

of additive except for L-histidine in which the concentration of additive in the buffer was of 2% (w/v).

.................................................................................................................................................................. 45

Figure 4.16 - Retained Biological Activity (%) after incubation at 70°C for a period of 8h of IgG solutions

with the initial concentrations of 1, 5, 10 and 15 g/L. The buffers used consisted in PBS with 2% (w/v) of

additive. .................................................................................................................................................... 46

Figure 4.17 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of maltose (pH ≈ 7) before incubation ()

and after 8 hours of incubation () at 70°C for the initial concentration of (A) 1g/L, (B) 5 g/L, (C) 10 g/L

and (D) 15 g/L. .......................................................................................................................................... 47

Figure 4.18 - Retained Biological Activity (%) after incubation of IgG solutions with the initial

concentrations of 1, 5, 10 and 15g/L at pH ≈ 2 for a period of 9h. The buffers used consisted in PBS with

10% (w/v) of additive except for L-histidine in which the concentration of additive in the buffer was of

2% (w/v). ................................................................................................................................................... 48

Figure 4.19 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of maltose before incubation () and

after 9 hours of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L

and (D) 15 g/L. .......................................................................................................................................... 49

xi

Figure 4.20 – Size (nm) of IgG in PBS at pH 2.1 before incubation and after 9 hours of incubation in PBS

at pH 2.1 and PBS with 10% (w/v) maltose at pH 2.1. ..............................................................................52

Figure 4.21 - Size (nm) of IgG in PBS at pH 11.3 before incubation and after 9 hours of incubation in PBS

at pH 11.3 and PBS with 10% (w/v) maltose at pH 11.5. ..........................................................................53

Figure 4.22 – Size (nm) of native IgG and IgG after 8 hours of incubation at 70 °C in PBS and PBS with

10% (w/v) maltose – pH ≈ 7. .....................................................................................................................54

Figure 4.23 – Variations in size of IgG in solution over time of incubation at 70°C in hours for the initial

concentration of IgG in solution of (A) 1g/L and (B) 15g/L, both in absence and presence of maltose.

The Fold Increase in Size corresponds to the ratio between the size at a given time and the initial

particle size................................................................................................................................................55

Figure B.1 - Retained Biological Activity (%) over time (h) of incubation at 60°C of IgG in PBS (pH 7.4)

with the initial concentrations of 1, 5, 10 and 15 g/L. ..............................................................................64

Figure B.2 - Retained Biological Activity (%) over time (h) of incubation at 70°C of IgG in PBS (pH 7.4)


Figure B.3 - Retained Biological Activity (%) over time (h) of incubation of IgG solutions with PBS at pH

3.02 with the initial concentrations of 1, 5, 10 and 15 g/L. ......................................................................65

Figure B.4 - Retained Biological Activity (%) over time (h) of incubation of IgG solutions with PBS at pH 2


Figure B.5 - Retained Biological activity (%) over time (min) of induced mechanical shear stress with the

inner cylinder rotating at 1000, 2000 and 3000 rpm and 3000 rpm for the initial concentrations of IgG

of 1 g/L and 15 g/L, respectively. ..............................................................................................................66

Figure B.6 – Comparison of the Retained Biological activity (%) over time (h) of induced thermal

denaturation (70°C) for the initial concentrations of IgG in PBS with (continuous line) and without

(dashed line) 10% (w/v) maltose of 1, 5, 10 and 15 g/L, respectively. .....................................................66

Figure B.7 – Comparison of the Retained Biological activity (%) over time (h) of pH induced

denaturation (pH ≈ 2) for the initial concentrations of IgG in PBS with and without 10% (w/v) maltose

of 1, 5, 10 and 15 g/L, respectively. ..........................................................................................................67

Figure C.1 – Far-UV CD spectra of IgG in PBS with 10% (w/v) of glycine (pH ≈ 7) before incubation ()

and after 8 hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L

and (D) 15 g/L.……………………………………………………………………..……………….................................................68

Figure C.2 – Far-UV CD spectra of IgG in PBS with 10% (w/v) of glycine before incubation () and after

9 hours of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D)

15 g/L. ........................................................................................................................................................69

xii

Figure C.3 – Far-UV CD spectra of IgG in PBS with 10% (w/v) of D-trehalose (pH ≈ 7) before incubation

() and after 8 hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C)

10 g/L and (D) 15 g/L. ……………….…………………………………………………………………………………………………………70

Figure C.4 – Far-UV CD spectra of IgG in PBS with 10% (w/v) of trehalose (pH ≈ 7) before incubation ()

and after 9 hours of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10

g/L and (D) 15 g/L. .................................................................................................................................... 71

Figure C.5 – Far-UV CD spectra of IgG in PBS with 10% (w/v) of sucrose (pH ≈ 7) before incubation ()

and after 8 hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L

and (D) 15 g/L. .......................................................................................................................................... 72

Figure C.6 – Far-UV CD spectra of IgG in PBS with 10% (w/v) of sucrose before incubation () and after

9 hours of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D)

15 g/L. ....................................................................................................................................................... 73

Figure C.7 – Far-UV CD spectra of IgG in PBS with 2% (w/v) of maltose before incubation () and after

8 hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D)

15 g/L. ....................................................................................................................................................... 74

Figure D.1 – Histogram showing the size distribution of an IgG solution in PBS (pH 2.1) before 9 hours

of incubation at pH ≈ 2, for the concentrations of row (A) 1 g/L and row (B) 15 g/L. The y-axis is the

intensity of the average scattered light (%) – Size distribution by Intensity - or total volume of particles

in the different size bins (%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters

(nm). ......................................................................................................................................................... 75

Figure D.2 – Histogram showing the size distribution of an IgG solution in PBS (pH 11.3) before 9 hours

of incubation at pH ≈ 11, for the concentrations of row (A) 1 g/L and row (B) 15 g/L. The y-axis is the

intensity of the average scattered light (%) – Size distribution by Intensity - or total volume of particles

in the different size bins (%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters

(nm). ......................................................................................................................................................... 76

Figure D.3 – Histogram showing the size distribution of an IgG solution in PBS with 10% (w/v) of

maltose after 9 hours of incubation at pH 2.06 for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis

is the intensity of the average scattered light (%) – Size distribution by Intensity - or total volume of

particles in the different size bins (%) – Size Distribution by Volume. The x-axis is the absolute size of

the scatters (nm). ..................................................................................................................................... 77


maltose after 9 hours of incubation at pH 11.45 for the concentrations of row (A) 1 g/L and row (B) 15

g/L. The y-axis is the intensity of the average scattered light (%) – Size distribution by Intensity - or total

xiii

volume of particles in the different size bins (%) – Size Distribution by Volume. The x-axis is the

absolute size of the scatters (nm). ........................................................................................................... 78

Figure D.5 – Histogram showing the size distribution of an IgG solution in PBS after 9 hours of

incubation at pH 2.1 for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the

average scattered light (%) – Size distribution by Intensity - or total volume of particles in the different

size bins (%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters (nm). ...........79


incubation at pH 11.3 for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the



Figure D.7 – Histogram showing the size distribution of an IgG solution (1 g/L) in row (A) PBS at room

temperature; row (B) PBS after 8 hours of incubation at 70°C; row (C) PBS with 10% (w/v) of maltose

after 8 hours of incubation at 70°C. The y-axis is the intensity of the average scattered light (%) – Size

distribution by Intensity - or total volume of particles in the different size bins (%) – Size Distribution by

Volume. The x-axis is the absolute size of the scatters (nm). ...................................................................81

Figure D.8 – Histogram showing the size distribution of an IgG solution (15 g/L) in row (A) PBS at room

temperature; row (B) PBS after 8 hours of incubation at 70°C; row (C) PBS with 10% (w/v) of maltose

after 8 hours of incubation at 70°C. The y-axis is the intensity of the average scattered light (%) – Size

distribution by Intensity - or total volume of particles in the different size bins (%) – Size Distribution by

Volume. The x-axis is the absolute size of the scatters (nm). ...................................................................82

Figure D.9 – Histogram showing the size distribution of an IgG solution in PBS after 0.5 hours of

incubation at 70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the




maltose after 0.5 hours of incubation at 70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is

the intensity of the average scattered light (%) – Size distribution by Intensity - or total volume of


the scatters (nm). ......................................................................................................................................84

Figure D.11 – Histogram showing the size distribution of an IgG solution in PBS after 1 hour of




xiv


maltose after 1 hour of incubation at 70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is



the scatters (nm). ..................................................................................................................................... 86




size bins (%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters (nm). ........... 87


maltose after 5 hours of incubation at 70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is



the scatters (nm). ..................................................................................................................................... 88









the scatters (nm). ..................................................................................................................................... 90









the scatters (nm). ..................................................................................................................................... 92

xv

LIST OF TABLES

Table 2.1 – Monoclonal Antibodies currently in the market – generic name, molecule type, marketing

company and approval status (11). ............................................................................................................. 8

Table 3.1 – Specifications of the chemical compounds used....................................................................26

Table 4.1 – Taylor number, shear rate (s-1) and shear stress (Pa) imposed on the IgG molecules during

the mechanical shear stress assay for the different angular speeds of the inner cylinder tested. ..........41

Table A.1 – Physicochemical properties (Relative Molecular Weigth of the molecule and the heavy

chain and Isoelectric point) of the several subclasses of Human IgG (2). .................................................63

Table A.2 – Binding and elution conditions commonly used with Protein A for Human IgG purification

(values for each subclass) (2). ...................................................................................................................63

Table A.3 – Data used in the calculations of the average and maximum shear rate and wall shear stress

at the membrane surface. Dimensions of the rotary separation chamber: inner and outter cylinder

radius (cm) an anular gap (mm) (22) and water cinematic viscosity at 20°C (45). ...................................63

xvi

LIST OF ABREVIATIONS

IgG – Immunoglobulin G

mAb – Monoclonal Antibody

CD – Circular Dichroism

DLS – Dynamic Light Scattering

PBS – Phosphate Buffer Saline

pI – Isoelectric point

CCSD – Concentric-cylinder Shear Device

1

INTRODUCTION

1.1 BACKGROUND

Therapies with biotechnology-derived products have been increasing steadily over the past years.

Antibodies, in particular, are widely used as convenient and valuable tools not only for therapy

purposes but as well as for immunochemical and biochemical analyses. Progress in monoclonal

antibody technology has led to the production of substantial amounts of highly specific monoclonal

antibodies (mAbs) (1). In fact, mAbs are the fastest growing class of biopharmaceuticals (2) with over

25 mAbs already approved for several diseases.

A common problem associated with these therapeutic proteins is that many of those proteins are

required in high doses, administered as part of high-dosing regimens (3), which makes necessary to

ensure stabilization of the active-product, not only during production and purification, but also during

storage. This constraint imposes the need to find proper formulations for those biopharmaceuticals,

which can be a challenge on its own. The successful formulation of proteins depends on a thorough

understanding of their physico-chemical and biological characteristics since the therapeutic activity of

proteins is highly dependent on their conformational structure. The protein conformational structure is

flexible and sensitive to external conditions (3). It is, therefore, essential to study, not only which

factors can cause disruption of the protein stability, but also to test different formulations in order to

identify which type of compounds may be relevant to the active-product stabilization and that may,

consequently, be added to the formulations to achieve that purpose.

1.2 OBJECTIVES

The main objective of the work developed in this thesis was to assess the stability of different

formulations of human therapeutic antibodies against temperature induced denaturation, pH induced

denaturation and mechanical shear stress. The goal was to identify which of the additives commonly

present in commercialized formulations of human therapeutic antibodies lead to higher antibody

stabilization.

2

1.3 OUTLINE

In order to assess the stability provided by preservatives added to the antibody formulations in the

market, several experiments were conducted with the purpose of evaluating the resistance to thermal

and pH denaturation as well as to mechanical shear stress.

The work conducted in this thesis is divided in three major parts: first the stability of the antibody

was tested against the three factors under analysis without the presence of any additive; then, and

based on the results obtained on the first part of the process, the stabilization potential of each

additive was evaluated and finally, aggregation studies were conducted for different conditions of

temperature and pH.

3

2 LITERATURE REVIEW

2.1 MONOCLONAL ANTIBODIES

Antibodies are key elements in the immune response in which they exert two main functions: At a

first level they are responsible to recognise and bind any foreign material (antigen) which is achieved

by the binding to molecular structures at the surface of the antigen – antigen determinants; A second

plane of action involves the binding of the so called effector molecules to the antibody-coated antigen,

thus triggering a serious of complex elimination mechanisms. The first task requires a great diversity of

antibodies in order to bind to each antigen while the second task imposes the need of those antibodies

to share some common features. These conflicting requirements are met by the antibody structure (4).

There is a basic configuration common to all immunoglobulin monomers consisting in two light

chains (L) and two heavy chains (H) (5). The two identical heavy chains span both the Fab and Fc

fragments while light chains are associated only with the Fab fragment (4). There are two types of light

chains, namely lambda (λ) and kappa (κ), being that, in humans, kappa chains are more prevalent than

lambda chains. Regarding the heavy chains, there are five main variants, namely: μ, γ, α, ε and δ, each

corresponding to one of the Immunoglobulin classes – IgM, IgG, IgA, IgE and IgD, respectively (5). The

heavy chains can be, furthermore, grouped into different forms or subclasses – these depending on the

species in evaluation. Since the focus of the study conducted in this thesis is on IgG, this particular class

will be addressed with further detail. Regarding the subclasses of the IgG molecule in humans, there

are four types with the heavy chains γ1, γ2, γ3 and γ4 which give rise to IgG1, IgG2, IgG3 and IgG4

subclasses, respectively.

The IgG molecule is organized into 12 homologous regions or domains, each one with a length of

approximately 110 amino acids. Each light chain consists in two domains: a variable (VL) and a constant

domain (CL). On the other hand, the heavy chain is constituted by four domains, three of which are

constant (CH1, CH2 and CH3) and only one variable (VH) (Figure 2.1). Disulfide bonds link the light chains to

the heavy chains as well as the two heavy chains to each other. Note that each domain possesses an

additional disulfide bond.

4

Figure 2.1 – IgG representation showing the light and heavy chains as well as the constant and variable regions.

Adapted from (6).

The three dimensional structure of the IgG molecule adopts a Y-shaped conformation, as illustrated

by Figure 2.1. It can be divided into three units, two of which are involved in antigen binding – Fab

(fragment antigen binding) and one involved with the binding to the effector molecules – Fc (fraction

crystallisable, since it crystallizes readily) (Figure 2.2). The shortest arms, each formed by four domains

(VL, CL, VH and CH1), contain the Fab fragments forming the antigen binding site (ABS). Note that, since

both the light chains have identical amino acid sequences, as well as the two heavy chains, each

immunoglobulin monomer has two abs. The Fc fragment corresponds to the stem of the ‘Y’ and is

formed by the constant domains of the heavy chains, CH2 and CH3. The Fab arms are linked to the Fc by

an extended polypeptide chain – hinge (which tends to be exposed and is sensitive to the attack of

proteases that cleave the molecule into its two functional groups) – being both the Fab and Fc

fragments roughly in the same plane.

In addition, all domains (except CH2) are in close lateral or in ‘sideways’ association with another

domain – the so called phenomenon of domain pairing. The CH2 domains have two sugar chains

interposed between them and exhibit as well a weak cis-interactions with the neighbour domains on

the same polypeptide chain.

Comparative studies of the sequence of several monoclonal IgG proteins indicate that the C-

terminal half of the light chain and approximately ¾ of the heavy chain (also C-terminal) show very

little sequence variation between IgG molecules, quite the opposite of the N-terminal of roughly 100

amino acid residues that showed considerable sequence variability in both the light and the heavy

chains. Comprised within these variable regions there are short sequences of extreme sequence

5

variability called hypervariable regions. These hypervariable regions are most likely associated with

the recognition of the diverse antigens – for that reason they are often called CDRs (complementary

determinant regions) (Figure 2.2). Three of those regions are located in the light chain while the fourth

is located in the heavy chain.

Figure 2.2 – Representation of IgG’s structure highlighting its several domains, as well as the Fab and Fc

fragments. The secondary structure of the Fab fragment is reptresented in detail (7). The Fv fragment

corresponds to an unstable fragment able to bind to an antigen with two V regions, VL and VH (2).

IgG molecule is very flexible and so, it can adopt many different conformations. This property may

help to achieve a better performance of its functions: Fab-Fab flexibility may provide a ‘wider range’ of

action, this meaning that the antibody may interact with determinants with different spacings; Fc-Fab

flexibility may help the antibody in different environments with the interaction with common effector

molecules.

Regarding the domains structure, each domain possesses a common pattern of polypeptide chain

folding, which consists in two β-sheets surrounding an internal volume of highly tight hydrophobic

residues (Figure 2.3). This arrangement is stabilized by an internal disulfide bond connecting the two β-

sheets in a central position. Both sheets are constituted by anti-parallel β-strands (one with three, the

other with four). These strands (also termed framework regions) are joined by bends or ‘loops’ that, in

general, show little secondary structure. The residues involved in the β-sheets tend to be quite

conserved, on the contraire of what happens with the residues in the ‘loops’ where there is a greater

diversity. It is important to note that the β-sheets of a variable domain are more distorted that those

from a constant domain. Furthermore, the variable domain possesses an extra ‘loop’.

Accounting the Fab fragment structure, the four individual domains are grouped in two different

ways. The VL and VH domains are paired trough the two respective three-strand- β-sheet layers while

6

the CH1 and CL domains are grouped trough the two four-strand- β-sheet layers. The interacting faces of

the domains are essentially hydrophobic and, consequently, the driving force for domain pairing is the

removal of those residues from aqueous environment. Further stabilization of the arrangement occurs

by a disulfide bond between the CH1 and the CL domains. The cis-interactions between CH1 and VH

domains and VL and CL domains are very limited thus allowing ‘elbow bending’.

For the Fc fragment the two CH3 domains are paired while the CH2 domains show no close

interaction. Still, they have interposed between them two branched N-linked carbohydrate chains that

have limited contact with each other. The CH2 domains contain the binding site for several important

effector molecules (4).

Figure 2.3 – Representation of IgG’s secondary structure, evidencing the high content of β-sheets (8).

Regarding Human IgG’s physicochemical properties, IgG presents a relative molecular weight (Mr)

that varies within the interval 150 000 - 160 000 Da. The Isoelectric point (pI) is comprised between 4

and 9, being in most cases higher than 6.0 (often more basic than other serum proteins). The

hydrophobicity of the IgG molecule is higher than most other proteins and so, precipitation occurs

more readily in ammonium sulphate for IgG than for other proteins. IgG is very soluble in aqueous

buffers but low near pI or in very low salt concentrations (depending on the antibody) (2). Specific

values for the several subclasses of human IgG can be consulted with further detail in Table A.1

(Appendix A).

7

2.2 MABS’ HISTORICAL AND ECONOMICAL BACKGROUND

Monoclonal antibodies (MAbs) are highly specific antibodies produced from single B-cell. The

high specificity of a monoclonal antibody is a significant advantage, particularly in therapeutic

applications (2).

The production of murine monoclonal antibodies (mAbs) was first reported in 1975. Initially,

they were derived from a murine hybridoma cell line (9). These hybridoma cells were created by

isolation of plasma cell precursors (B-cells) which were then fused with immortal cells (myeloma cells).

After combination and expansion, these hybridoma cells secrete only one antibody type, a murine mAb

(2). By 1980, mAbs were already being tested in humans (10). However, murine mAbs presented

several problems such as short serum half-life, weak immune response recruitment, development of

human anti-mouse antibody immunogenic response (HAMA response), the possibility of a fatal allergic

reaction, the loss of efficacy and function after small epitopic changes seen mostly in highly mutagenic

viruses often leading to drug resistance, and the need for repeated and higher doses of the antibody,

particularly in chronic disorders increasing the risk of side effects (11). Although Chimeric mAbs (fusion

of murine variable domains with human constant domains) – first reported in 1984 (10) - solved the

problem of interaction with the human immune system, the human anti-chimeric antibody response

still constituted a problem (11). Humanized mAbs, firstly reported in 1986, were developed to address

the remaining problems associated with chimeric mAbs (10). Approximately 85–90% human, these

humanized murine antibodies are less immunogenic than chimeric antibodies and have either limited

or eliminated the serious immune reactions that often occur in some patients (11). Currently, there are

several mAbs in the market, some of which are entirely human – in fact, the first fully human antibody

was only approved by the FDA in 2002 (9).

Table 2.1 shows some of those commercialized mAbs, specifying their type, marketing company and

approval status by the FDA and EMEA.

8

Table 2.1 – Monoclonal Antibodies currently in the market – generic name, molecule type, marketing company

and approval status (11).

Currently mAbs represent one of the successes of the biotechnology industry having a significant

impact on the market both due to their sales volume and their vast applications. In 2010, the sales

volume of mAbs exceeded 40 billion US Dollars and is expected to reach at least 70 billion US Dollars by

2015. The majority of the approved antibodies are targeting cancer and autoimmune diseases with the

top 5 grossing antibodies populating these two areas. In addition, over 100 monoclonal antibodies are

in Phase II and III of clinical development and numerous others are in various pre-clinical and safety

studies (9).

9

2.3 STABILITY REQUIREMENTS

According to the FDA (Food and Drug Administration), the product stability should meet the

requirements imposed by a previously developed clinical protocol. FDA states that product evaluation

should follow a stability program that includes tests for the physic-chemical integrity (fragmentation

and aggregation), potency, sterility, as well as, moisture, pH and preservative stability, at regular

intervals during the dating period. The FDA also requires the need of stability tests that assure the

biological activity, which should include a manufacturer’s in-house reference standard. Whenever

possible, a single lot of test antigen should be used throughout the study. A quantitative potency assay

should also be conducted in order to allow a meaningful comparison of activities. Finally, FDA requires

accelerated stability tests, i.e., stability testing after storage at temperatures higher than the normal

temperatures of storage. That can help to identify and to establish which tests are indicators of

stability. Specific parameters that are stability indicators should be monitored in all the lots of the

stability test program. Data from the accelerated stability tests can be used as support but should not

replace real-time data for product approval (12).

EMEA (European Medicines Agency) imposes the requirement to perform tests that help

determining and finding the optimal formulations to avoid the development of visible particulates

during storage. These visible particulates are the result of the natural tendency of immunoglobulins to

form aggregates, especially when in high concentration, which is common for several drugs since such

high concentrations of immunoglobulin are often needed to reach the intended therapeutic effect (13).

Antibody formulations may be aggregate-free after the last polishing step but that doesn’t imply that

aggregates won’t form during storage. In fact, protein aggregation is encountered routinely during

purification, refolding, sterilization, shipping and storage processes due to the existence of chemical

and physical stresses. Both native and denatured proteins may aggregate: aggregation of denatured

proteins is extremely common while aggregation of native proteins is most common upon long periods

of storage (it may also occur during freeze drying). Aggregation is undesirable because it causes a

decrease in the drug efficacy (14); in fact, irreversible aggregation has a direct impact on drug potency,

immunogenicity and on the unfolded protein response. Reversible aggregation may also lead to

problems during drug administration if the dissociation is slow on the physiological time scale (3). For

those reasons, aggregation levels even as low as 1% over a 2-year shelf life can cause a product to be

clinically unacceptable (14).

10

2.4 PUBLISHED STUDIES

Several studies have been conducted over the last years in order to assess the stability of IgG alone

as well as the influence of numerous factors that might contribute either to increase the stability of IgG

or to induce its degradation. The subsequent paragraphs present a brief description of some studies

conducted on this matter as well as of the information derived from them.

In 1995, Alexander and Hughes monitored the thermal stability and degradation of an internalizing

chimeric (human/mouse) monoclonal antibody (BR96) using capillary electrophoresis and MALDI-MS.

After 166 h of thermal stress at 60°C, the percentage of ionization by MALTI-MS carried by the intact

antibody molecular ions had clearly decreased, while that due to additional ion species had

significantly increased. These additional ion species correspond to fragments without one light chain,

without a Fab arm and to the formation of separated heavy-chain and light-chain moieties. Several of

these fragments resulted from simple disulfide linkage disruption. In addition, species less in mass than

common antibody subunits were also observed, demonstrating peptide as well as disulfide bond

cleavage. The observation that a small number of well-defined species were formed during the study

suggests that the cleavage induced by thermal stress is very site-specific within the IgG (15).

In 1998, Vermeer et al. studied the structural changes of IgG (a monoclonal mouse immunoglobulin

G from isotype 1) induced by heat treatment and by adsorption onto a hydrophobic Teflon surface.

Using differential scanning calorimetry and circular dichroism measurements it was shown that heat-

induced denaturation doesn’t lead to complete unfolding into an extended polypeptide chain, but

leaves a significant part of the IgG molecule in a globular or corpuscular form. It was verified that

heating dissolved IgG causes a decrease of the fractions of β-sheet and β-turn conformations, whereas

those of random coil and, to a lesser extent, α-helix increased (1).

Later, in 2000, the denaturation of IgG (a monoclonal mouse anti-rat antibody of isotype 2b) was

studied by the same group using different calorimetric methods (differential scanning calorimetry and

isothermal calorimetry) and by circular dichroism spectroscopy. The thermogram of the

immunoglobulin showed two main transitions that are a superimposition of distinct denaturation

steps. The transitions were independent and the unfolding was followed by an irreversible aggregation

step. Each transition has different sensibility to changes in temperature and pH: the Fab fragment is

most sensitive to heat treatment while the Fc fragment is most sensitive to decreasing pH. The

denaturation mode also affects the structure of the aggregates formed. It was also shown that below

the unfolding temperature, the denaturation rate is controlled by the unfolding step whereas at higher

temperatures (where a relatively high concentration of partially unfolded IgG molecules is present),

the rate of aggregation is so fast that the molecules become trapped in the aggregates before being

11

fully denatured. The CD spectra revealed a strong correlation between the changes in the secondary

structure and the denaturation transitions observed by calorimetry. After heat and pH-induced

denaturation, it was observed that a significant fraction of the secondary structure remained (16).

In the same year, this group also studied the effect of low molecular weight surfactants on the

thermal stability of immunoglobulin G (monoclonal mouse anti-rat antibody of isotype 2b) by

differential scanning calorimetry, being the corresponding change in the secondary structure

investigated using circular dichroism spectroscopy. The rate of aggregate formation, both in the

presence and absence of surfactant, was also monitored by dynamic light scattering. At low surfactant

concentrations (SDS:Tween 20 mixture) the thermal stability of the protein was not affected. With

increasing surfactant concentration the protein structure was perturbed, most probably due to

hydrophobic interaction with the surfactant, leading to a lower thermal stability. At even higher

concentrations, the surfactant molecules encapsulated the protein molecules, so that the unfolded

state was strongly suppressed due to restricted conformational freedom in a confined volume.

Interaction with the surfactant mixture at intermediate concentration showed a strong influence in the

secondary structure of IgG: α-helix and random coil conformations were promoted and the percentage

of β-sheets and β-turns was reduced (17).

Conformation, acid-induced conformational changes and stability of the murine monoclonal

antibody CB4-1 and its Fab and Fc fragments, were analysed by circular dichroism, fluorescence, and

differential scanning calorimetry measurements by Wefle et al. (1999). It was shown that lowering the

pH to 3.5, reduces the stability but does not change the conformation of the CB4-1 antibody. At pH

values between 3.5 and 2.0, conformational changes and the formation of new structures were

observed. Deconvolution of the bimodal DSC curves of the CB4-1 antibody revealed five `two-state'

transitions at pH 7.5. At pH 5 and below, only four transitions were found. Half transition

temperatures, Tm, and molar enthalpy changes, ΔHm, gradually decreased at pH 4 and 3.4. At pH 2.1,

two low-temperature and two high-temperature transitions were identified. The Fab and Fc fragments

behaved similarly. Deconvolution of their monophasic DSC curves yielded two `two-state' transitions

for each fragment. Tm and ΔHm values gradually decrease at pH 4.0 and 3.4; and at pH 2.1 and 2.8 for

Fab and Fc, respectively, one of the transitions was found at high temperature (18).

In 2004, Bermudez and Forciniti monitored the denaturation and aggregation of human polyclonal

IgG by aqueous two-phase partitioning, capillary electrophoresis and dynamic light scattering. The

denaturant agent used was urea, added in various concentrations. In the absence of urea, the size

obtained for native IgG was 10.9 nm, being also present a small fraction (≈1%) of dimmers and large

aggregates in the commercial IgG sample. Upon a concentration of 8M of urea, it was verified an

increase in the monomer size to 13.9 nm in an unfolded state. It was also observed that the larger the

12

extension of denaturation, the higher the number of dimmers and monomers present (because

although there is an increased aggregation, breakage of the larger aggregates had occured) (14).

The effect of solvent environment on the conformation and stability of human polyclonal IgG in

solution was evaluated in 2005 by Szenci et al. using different techniques. It was observed that the

conformational stability decreases with decreasing pH while the resistance to aggregation increases.

The optimum pH range determined for storage was 5.0-6.0, being a compromise between

conformational stability and the tendency for oligomerization. Regarding the thermal stability of IgG, it

was shown that additives in physiologically acceptable concentrations exert no influence. In contrast,

glucose or sorbitol, in concentrations as low as 1%, had a significant effect on the tertiary structures.

On its turn, 0.3% leucine, even though it didn’t increase the conformational stability, caused a decrease

in the tendency for aggregation more effectively than 1% glucose or sorbitol. Regarding the

temperature effect, it was observed an increase in the dimmer content upon storage at 5°C which was

partially reverted upon incubation at 37°C. It was also suggested that storage at temperatures higher

than 5°C may help to maintain the optimal proportion of dimmers (19).

Liu et al. (2006) characterized the stability of a fully human monoclonal IgG after 6 months

incubation at 40°C employing mass spectrometry and chromatography analyses. It was found that

deamidation, fragmentation and N-terminal glutamate cyclization to form pyroglutamate were the

major degradation pathways. Three major deamidation sites were identified and one site in a small

tryptic peptide accounted for more than 80% of the total. Peptide cleavage was observed at several

positions between different pairs of aminoacids. Most of the cleavage sites were located in the hinge

or other flexible regions of the IgG molecule (20).

Recently, in 2011, Kanmert et al. studied the thermal induction of an alternative folded state in

human IgG-Fc. It was reported the formation of a non-native, folded state of the Fc fragment of human

IgG4 (IgG4-Fc) induced by a high temperature at neutral pH and at a physiological salt concentration.

The obtained structure was similar to the molten globule state in that it displays a high degree of

secondary structure content and surface-exposed hydrophobic residues. However, it was highly

resistant to chemical denaturation. The thermally induced state of human IgG4-Fc was thus associated

with typical properties of the so-called alternatively folded state previously described for murine IgG,

Fab fragments, and individual antibody domains (VL, VH, CH1, and CH3) under acidic conditions in the

presence of anions. It was suggested that, like some of these molecules, human IgG4-Fc in its

alternative fold exists as a mixture of different oligomeric structures, dominated by an equilibrium

between monomeric and heptameric species. Further heating induced the formation of fibrous

structures in the micrometer range (21).

13

2.5 DENATURATION CAUSED BY SHEAR STRESS

The stability of a protein can be affected by both chemical and physical stress, which can ultimately

cause a protein to unfold. Among the physical stresses, the effects of temperature, shear stress and

pressure are relatively well characterized, and plenty of examples can be found in literature. Some

chemical stresses that are also known to induce changes in the structure of proteins include pH, metal

ions, chaeotropic salts, detergents, preservatives, and organic solvents.

Loss of stability induced by shear stress can be observed during routine protein processing or

handling, including pumping, mixing, stirring, membrane filtration, filling, and transportation. In this

thesis, a vorticular flow membrane module was used to induce shear forces on the antibody molecule.

2.5.1 Vorticular Flow Membrane Modules

Vorticular flow membrane modules are highly efficient separation systems which can provide fast,

reproducible separations of fermentation broths, cell culture supernatants, cell homogenates, column

fractions, tissue extracts and a wide variety of simple or complex solutions of proteins and other

biopolymers (22).

Rotating filters are currently used for several separations. They offer the unique advantage of a

greatly reduced plugging of the pores of the filter with particles when compared with standard

filtration techniques, which are severely limited by membrane fouling (23). Other advantages of these

systems are the possibility to obtain high fluxes at high values of transmembrane pressure and a

favourable effect on membrane selectivity. However, the complexity and limitations in membrane area

for some systems contribute to raise the equipment cost (24).

To perform the experiments conducted in this thesis the vorticular flow filtration system was

modified by removing the filter cartridge to leave the inner cylinder as a solid rod. The resulting system

– CCSD (concentric-cylinder shear device) – consisted of an electronic control unit and a closed rotary

separation unit.

The rotary separation unit consists of an inner cylinder rotating within a stationary outer concentric

cylindrical shell. The flow field is best described as supercritical Couette flow in an annulus with a

superimposed axial flow (23). Couette type systems, with an inner cylinder rotating inside a concentric

one, generate toroidal instabilities called Taylor vortices (24). In the presence of an axial velocity

component the Taylor vortices start moving helically along the rotor (25).

The rotation of the inner cylinder creates a centrifugal field that is maximum near the rotating inner

cylinder and decreases to zero at the stationary outer cylinder. This centrifugal field creates a net radial

14

force on non-neutrally buoyant particles. At the same time the Taylor vortices carry particles from near

the inner cylinder to near the outer cylinder and vice versa. The axial flow, on its turn, carries particles

along the length of the rotating cylinder annulus. Figure 2.4 illustrates the CCSD components, showing

the flow regimen and Taylor vortices in detail.

Figure 2.4 – Schematic drawing of a CCSD, showing Taylor vortices (26).

A concerning issue comprised by this system is related to the high shear forces generated by the

Taylor vortices, that may be adverse regarding some components of the solution circulating in the

system. For instance, they may be responsible for animal cells membrane damage, degradation or

structure alteration of proteins resulting in possible changes or loss of function, etc. (23).

The flow regimen in the annulus between the rotating inner cylinder and the stationary outer

cylinder is governed by the Taylor Number, Ta (dimensionless):

where ω denotes the angular speed, Ri the inner cylinder radius, e the annular gap and ν the

kinematic viscosity (26). Couette type systems generate Taylor vortices that appear when the Taylor

number exceeds 42 (24). For 42<Ta<800 laminar flow with vortices occurs in the system, for 800 < Ta <

2000, transitional vortex flow occurs and 2000 < Ta < 10000-15000 turbulent flow with vortexes

dominate flow (27). By measuring the torque exerted on the inner cylinder, Taylor proposed an

√

(2.1)

15

equation for the maximum shear rate (occurring at the wall of the inner rotating cylinder), γm (s-1), in

the inner cylinder that includes both laminar and turbulent regimens (24).

Knowing that shear rate, γm, is defined as

it should be possible to estimate the shear stress at the wall of the inner cylinder, τm, at which the

proteins are subjected if the viscosity, μ, is known (28).

2.6 STABILITY ASSESSMENT

There are different techniques that can be used to monitor and evaluate the stability of proteins. Some

of these techniques, including affinity chromatography, circular dichroism and dynamic light scattering

will be reviewed in this section.

2.6.1 Affinity Chromatography

Affinity Chromatography is a convenient method of isolation and purification of biological

molecules. The molecule to purify is passed through a column containing a chromatographic matrix -

an insoluble polymer or gel, chemical and physically inert- to which a specific ligand is covalently

bound. The separation thus occur based on the degree of affinity of the molecule towards that specific

ligand: molecules with no affinity pass through without being retained whereas molecules with affinity

for the ligand will bind and be retained (29). This technique is unique since it is the only technique that

enables the separation of biomolecules based on its biological function or its structural characteristics.

Affinity Chromatography has many wide ranging applications: it can be used to separate active

biomolecules from denatured or functionally different molecules, to isolate pure substances with a

very low concentration in a large crude sample or to remove specific contaminants thus contributing to

enhance the degree of purity of the biomolecule of interest. It offers high selectivity, hence high

resolution and usually high capacity for proteins of interest.

As previously stated, successful affinity purification requires a biospecific ligand capable of being

coupled in an active with a solid chromatographic matrix. In addition, the ligand/target complex must

⁄

(2.2)

(2.3)

16

be readily dissociated under certain specified conditions in order to release the target molecule in an

active form. In other words, the ligand should bind to the target molecule as strongly as to allow

efficient binding but no so strongly that it inhibits the target molecule of being eluted, this meaning

that the binding should be reversible.

The elution scheme is variable according to the system and may be selective or non-selective

applied in combination with group-specific ligands and highly specific ligands, respectively. Elution of

the target molecule can be achieved either by a change in the buffer composition, extreme conditions

of pH or concentrations of chaotropic agents or even with specific elution that consists in adding a

substance that competes for binding (30).

Regarding the purification of monoclonal IgG by affinity chromatography, the basis for such

separation is its high affinity for Protein A and/or Protein G. These proteins, both from bacterial origin,

create extremely useful and easy to wear media for IgG purification. Protein A is derived from a strain

of Staphylococcus aureus and is covalently bound to the peptidoglycan of the cell wall (31). It consists

of a single polypeptide chain in the form of a cylinder and contains five highly homologous antibody-

binding domains. The binding site for Protein A is located on the Fc fragment of IgG and it occurs

trough an induced hydrophobic fit. At the centre of the Fc binding site and on Protein A resides

histidine residues that are uncharged and hydrophobic at alkaline pH thus strengthening the

interaction between Protein A and the antibody. As the pH shifts to acidic values the histidine residues

become charged and repel each other, allowing the elution of the antibody (32). One molecule of

protein A can bind at least two molecules of IgG (30).

Protein A is suitable for the use in purification of antibodies because, besides being well

characterized and widely available, it has a dissociation constant of 10-7, which reflects its affinity to

the Fc region of IgG. Protein A is also stable in a wide range of pH (pH 2-11) (33). Typically, the elution

pH when using Protein A is approximately 3. Further details on the usual binding and elution conditions

commonly used for IgG purification with Protein A may be consulted in Table A.2 (Appendix A).

Protein G is a cell surface protein from Group G streptococci. The binding mechanism is a non-

immune mechanism. Like protein A, protein G binds specifically to the Fc region of IgG (it is a type III

Fc-receptor). However, it binds more strongly to several polyclonal IgGs and to human IgG. In fact,

protein G is a better choice for the general purpose of capturing antibodies since it binds a broader

range of IgG from eukaryotic species and more classes of IgG. Under standard buffer conditions,

protein G binds to all human subclasses and all mouse IgG subclasses, including mouse IgG1. Protein G

also binds rat IgG2a and IgG2b, which either bind weakly or do not bind at all to Protein A.

Furthermore, it exhibits minimal binding to albumin which results in cleaner preparations and greater

yields (2).

17

As both Protein G and Protein A bind to the Fc region of immunoglobulins, it can be used to probe

whether the Fc region maintains its structural conformation. This type of affinity chromatography is

not a measure of the antibody biological activity, as this is defined by the Fab fragment, but can be an

indication of the folding of the Fc regions, and thus of the all molecule.

2.6.2 Circular Dichroism

Circular Dichroism (CD) is a particularly suitable technique for studies on biological systems. It

explores the differential absorption of right and left circularly polarized light and so, it is uniquely

sensitive to the asymmetry of the system, being ideal to study chiral molecules in solution such as

proteins, polypeptides and nucleic acids (34).

Light, being described as an electromagnetic wave, possesses both an electric and a magnetic field.

These fields oscillate perpendicularly one to the other as well as to the propagation direction. The

polarization of a wave is defined by the orientation in which the electric field (or alternatively the

magnetic field) oscillates.

Linear or plane polarized wave refers to a wave in which the direction the electric vector is

constant, though it can also be the result of two circularly polarized components with the same

amplitude and opposite directions (left and right-handed). On the other hand, in circularly polarized

light, the electric vector rotates along the propagation axis with constant amplitude thus forming a

helix, which describes a circle, at any plane of intersection that is perpendicular to the direction of

propagation. Circularly polarized light will be either left or right-handed, depending whether the

orientation in which the field rotates is to left or to the right, as illustrated in Figure 2.5.

Circularly polarized light can be described as the superimposition of two plane polarized waves with

equal amplitudes oscillating perpendicularly to each other and with a phase difference of π/2 radians. If

the amplitude or the phase shift is different, elliptical polarization will occur instead (35).

Whenever light passes through an optically active molecule, both the left and right circularly

polarized rays will exhibit different velocities (CL≠CR) thus resulting in different wavelengths (λL≠λR) and

extinction coefficients (εL≠εR). Consequently, within the range in which the molecule absorbs,

absorptions will also be different (AL≠AR). The differential absorption (ΔA = AL - AR) or, alternatively, the

difference between the extinction coefficients (Δε = εL - εR) of left and right-handed circularly polarized

light, at a given wavelength, is determined as Circular Dichroism (36).

18

Figure 2.5 – (A) Plane polarized light resolved into two circularly polarized components – left, L, and right, R. As

long as the intensities and phases of the two circularly polarized components remain equal, their resultant will

lie in a plane with oscillating magnitude. (B) If the right circularly polarized component is less intense (more

absorbed) than the left one, the electric vector of the light follows an elliptical path, thus corresponding to

elliptically polarized light (37).

For the quantification and standardized description of CD there is the need to resort to the Beer-

Lambert-Bourguer law (equation (2.4)). If I0 is the intensity of the incident light on the cell and I the

intensity when the light is leaving the cell, then

In an optically active medium two absorptions of the previous kind can be recorded, one referring

to the left circularly polarized light, AL, and the other to the right circularly polarized light, AR. Note that

the intensities, measured when the light encounters the material, are equal for both the left and right

circularly polarized light.

Alternatively, CD can also be expressed in function of the ellipticity parameter (θ). As previously

stated, elliptical polarization derives from the preferential absorption of one of the circular

components. Ellipticity can be determined by a mathematical expression describing the shape of the

formed ellipse, therefore given by the angle corresponding to the arctangent of the ratio between the

minor and major axis of the ellipse, as shown in equation (2.5), where |E| refers to the magnitude of

the electric field vector.

(2.4)

(| | | |

| | | |) (2.5)

19

The angle involved is usually very small so tan (θ) can be approximated to θ in radians. In addition, if

one considers that the intensity of the light is proportional to the square of the electric field vector and

that, as mentioned above, there’s a dependency of the Absorption with the Intensity according to the

Beer-Lambert-Bourguer law, one can obtain the following equation:

If one expands the terms of the previous equation in a Taylor series, neglecting the terms in order

of ΔA by comparison to the unity, equation (2.7) can be obtained once the angle is converted to

degrees.

In order to normalize the experimental data obtained - this meaning to remove any dependency

with light path or solute concentration – the molar ellipticity is defined as:

where C stands for the solute concentration (M) and l the length path of the light (cm).

Combining the Lambert-Beer-Bourguer law with equation (2.9) it is shown that:

At last, it is important to refer that each of the previously described variables are wavelength

dependent and so, they can be plotted in function of the wavelength in order to obtain the CD

spectrum of the molecule under study.

CD data is usually represented by the mean residue ellipticity that accounts the number of peptide

bonds as well as the molecular weight of the protein – this due to the fact that the peptide bond is the

main chromophore in the Far-UV region of the CD spectrum. The mean residue ellipticity can be

determined according to the following equation (35), (38):

(2.6)

( ) (

) (2.7)

[ ] ( )

(2.8)

[ ] ( ) (2.9)

20

where c is the solute concentration (g/L), mrw the mean residue molecular weight and the remaining

variables have the same meaning as above. The mrw can be determined dividing the protein molecular

weight (Mw) by the number of peptide bonds (number of peptide residues (NA) minus 1).

CD of biomolecules has two major strands: it is used to probe either changes in the conformation of

the molecule or its interactions with small achiral molecules (which by themselves show no intensity in

a CD spectrum).

Proteins have well-defined sequences of amino acids (primary structure) folded in a much defined

way. The overall shape of a protein is crucial to its biological activity. The regular and well defined

secondary structures formed are the result of the peptide link between amino acids (O=C-N-) being

planar and rigid but still allowing a somewhat high degree of free rotation regarding the bonds to the

rest of the protein chain. This constringes the possible relative orientation of the neighbour residues

and allows several intramolecular H-bonding between the C-O of one peptide and the N-H of the other.

Each set of common chiral secondary structural units have a somewhat well defined CD spectrum.

Conversely, CD is very useful to study the structure of proteins.

Regarding protein UV spectroscopy, the UV spectra can be divided into near-UV and far-UV spectra.

Near-UV refers to wavelengths between 250 and 300 nm and it is also described as aromatic region, on

account that these are the transitions that play the major role on the intensity of the CD signal.

However, transitions of disulfide bonds (cystines) also contribute for the absorbance intensity in this

region. The far-UV region refers to wavelengths under 250 nm and is primarily influenced by

transitions on the backbone of the protein chain although transitions in the lateral chains may also

have some contribution (in particular, if the α-helix content is low it may occur erroneous

determinations of the protein structure due to that fact). Empirical analysis of the far-UV CD spectrum

provides information on the secondary structure whereas the near-UV spectrum offers inlet on the

tertiary structure.

The lowest energy transition of the peptide chromophore is n→π*transition (similar to those of

ketones) followed by π→π* transition. The first transition occurs at 210-230 nm and its electric

character is polarized roughly along the lines of the carbonyl bond. The latter, dominated by the

[ ] ( )

( )

( )⇔ [ ]

(2.10)

21

carbonyl π-bond, is also influenced by the involvement of the amide nitrogen in the π-orbitals, being its

electric dipole transition moment polarized along the line between nitrogen and oxygen and centred at

190 nm. For α-helix the previous transition occurs around 208 nm.

All the lateral aromatic chains – phenylalanine, tyrosine and tryptophan - have transitions in the

near-UV region. The tryptophan’s indole has two or more transitions at 240-290 nm (εmax at 279 nm)

while both tyrosine and phenylalanine has only one transition (εmax at 279 nm and 258 nm,

respectively). There is also absorption from a cystine disulfide bond. In crescent order of intensity the

transitions are phenylalanine, cystine, tyrosine and tryptofan. Even though the more intense

transitions in this region correspond to tryptophan, they aren’t necessarily dominant because most

proteins have few tryptophans comparatively to the other aromatic groups.

CD can be used to trace changes in the conformation of proteins and for that it is required the

ability to determine – even if with some degree of inaccuracy – the protein structure. All the pure

conformations have their CD spectrum described. In principle, the CD spectrum of a native protein can

be expressed as a linear combination of the appropriate percentages of each component spectrum.

Nevertheless, relative arrangements of structural units and motifs, such as disulfide bonds, contribute

to the observed spectrum. The individual spectra of each motif, described below, are shown in Figure

2.6.

.

Figure 2.6 – Typical CD spectra for particular secondary structural motifs (α-helix, β-sheet, random coil) used in

CD protein structure fitting programs (39).

22

The α-helix motif is a well defined motif in which the nth peptide unit forms H-bonds between its C-

O and the N-H of the (n+4)th peptide as well as its N-H and the C-O of the (n+4)th peptide. The CD

spectra of α-helices exhibit a negative band with separate maxima with similar magnitude at 222 nm

(n→π* transition) and at 208 nm (π→π* transition). It’s the only motif where the π→π* transition has

such a long wavelength component, being therefore easily identifiable. Its magnitude is the highest of

all the motifs and is variable according to the helix and helix length variations, as well as with

interactions with neighbour structural units.

The β-sheet motif results of an alternative efficient formation of H-bond between parallel or anti-

parallel runs of amino acids. These runs of amino acids face in alternate directions so that alternated

amino acids form H-bonds with neighbour runs of amino acids. The β-sheets spectra are characterized

by a negative band at approximately 216 nm and a positive band of similar magnitude at 195 nm (34).

Furthermore, the intensity of absorption is approximately zero at 206.5 nm, while the other

components have a significant contribution at this wavelength. At this wavelength the ellipticity due to

the fraction of β-turns is positive whereas α-helices and random coil structures give a negative

contribution to the ellipticity (1). Typically the strands of an anti-parallel β-sheet are connected by β-

turns where the nth peptide forms an H-bond with the (n+3)rd peptide unit. The term β-turn also refers

to any turn that might happen in the chain. The characteristics of such spectra include a weak red-

shifted negative band (n→π*) near 225 nm, a strong positive π→π* transition between 200 and 205

nm and a strong negative band between 180 and 190 nm.

The completely denatured state of a protein is referenced as ‘random coil’. The term is also used to

designate sections of the protein that don’t fit the standards of neither of the secondary structural

motifs. In general, the CD spectrum shows only negative contribution, approaching to zero around 230

nm.

In order to determine the secondary structure of complex proteins several algorithms have been

developed. All of them rely on common ground, this is, they are based on five hypotheses which, in

general, are valid. The assumptions on which the construction of those algorithms is based are the

following: (i) the contributions of individual secondary structures are additive; (ii) the ensemble-

averaged solution structure and the time-averaged solid structure are equivalent; (iii) the CD

contributions of non peptide chromophores have no influence on the analysis; (iv) the effect of the

tertiary structure on CD is negligible; (v) effects of the geometric variability of the secondary structure

don’t need to be explicitly considered (36).

23

2.6.3 Dynamic Light Scattering

When light collides on matter, the electric field of light induces an oscillating polarization of the

electrons of the molecules. The molecules act then as secondary sources of light and, subsequently,

radiate (scatter) light. The frequency shifts, angular distribution, polarization and intensity of the

scattered light are determined by the size, shape and molecular interactions of the scattering medium.

As a result, it should be possible to obtain information about the structure and molecular dynamic of

the scattering medium based on its light-scattering characteristics (40).

There are two types of light scattering experiments – classical and dynamic - and both can be used

to characterize proteins. Classical light scattering allows the direct measurement of the molecular

mass, so it is useful to determine whether the native form of a protein is a monomer or an oligomer. It

can also be used to measure the mass of aggregates or of other non-native forms present in solution.

On the other hand, dynamic light scattering (DLS) uses the diffracted light to determine the ratio of

diffusion of protein particles (41).

In a DLS experiment, light from a laser passes through a polarizer that defines the polarization of

the monochromatic beam incident on the scattering medium. The scattered light then passes through

an analyser that selects a given polarization and enters at last in the detector. The detector’s position

defines the scattering angle, θ. The intersection of the incident beam with the beam intercepted by the

detector defines a scattering region of volume V – the region under analysis (40).

The data resulting from DLS experiments is usually processed in order to derive a size distribution to

the sample in study, being the size given by the hydrodynamic radius (alternatively called Stokes

radius) of the protein particles. This hydrodynamic size is dependent on the molecular mass and shape

(conformation) of the protein. DLS is particularly suitable for determine the presence of very small

amounts of aggregates (less than 0.01% by weight) and to study samples with a wide variety of masses.

It can also have great utility when comparing the stability of different formulations – monitoring

inclusively real time changes at high temperatures.

With DLS one measures the time dependency of refracted light of a small region of the solution

over a time range of tenths of microseconds to milliseconds. The fluctuations in the intensity of the

scattered light are related with the diffusion ratio of molecules in and out of the region under analysis

(Brownian motion) (41). In other words, Brownian motion is the random movement of particles due to

the bombardment by the solvent molecules that surround them. The larger the particle, the slower the

Brownian motion will be (42).

The data can be analysed to give directly the diffusion coefficients of the light scattering particles.

When multiple species are present, a distribution of the diffusion coefficients is seen.

24

In general, instead of the data being present in function of the diffusion coefficients, it is processed

in order to give the “size” of the particles (radius or diameter). The correlation between the diffusion

and the particle size is based on the theoretical relationships for the Brownian motion of spherical

particles – the Stokes-Einstein equation.

where d(H) is the hydrodynamic diameter; k, the Boltzmann’s constant; D, the translational diffusion

coefficient; T, the absolute temperature; and η, the viscosity. The translational diffusion coefficient will

depend not only on the size of the particle “core”, but also on any surface structure, as well as the

concentration and type of ions in the medium (42). The particle size can be obtained by different

methods: distribution of size by intensity (amount of scattered light by the particles in the different size

bins), by volume (total volume of the particles is the different size bins) or by number (number of

particles in the different size bins) (43).

The hydrodynamic diameter or the Stokes radius, Rh, derived from this method, corresponds to the

size of a spherical particle that would have a diffusion coefficient equal to the one of the native

protein. The data is usually presented in terms of the fraction of particles in function of their size.

Most proteins are not spherical and, consequently, their hydrodynamic size will depend on both

their molecular mass and conformation. Furthermore, the diffusion is affected by water molecules

bound or entrapped by the protein. Therefore, the hydrodynamic size can significantly differ from the

real physical size of the protein. In addition, it isn’t, as well, a reliable measure of the molecular mass.

Despite that, the hydrodynamic size can be direct and accurately determined by DLS.

DLS is, in theory, capable of distinguish whether a protein is a monomer or dimmer - the radius

would be measurably different. However, it cannot resolve monomers from small oligomers neither

can quantify fractions of small oligomers. Generally, two species must differ on their hydrodynamic

radius in a factor of 2 or higher to be resolved into two separate peaks. That factor of 2 regarding the

radius difference is translated into a factor of 8 with regard to the difference in the molecular mass. As

a result, monomers, dimmers, trimmers up to octamers are resolved in one average peak, giving DLS

less use when it comes to analysing small oligomers.

DLS offers the possibility to study samples directly in their formulation buffers as well as samples

with high protein concentration (at least up to 50 mg/mL, sometimes higher). However, the data

interpretation for samples with higher concentrations – in particular, the hydrodynamic radius – can be

somewhat difficult to interpret due to the non ideality effects of the solution. Nevertheless, as in the

( )

(2.11)

25

data for lower protein concentration, the fact that the sample is not homogeneous and contains large

aggregates is not subject to errors.

Another advantage of DLS is its utility to assess the formation of aggregates over the time and to

directly compare the degradation ratio of different formulations. It is possible to conduct somewhat

fast stability studies by monitoring a single sample in situ or by periodic sampling from samples kept at

high temperature.

The major inconvenient of DLS is that it is frequently difficult to measure accurately the amount of

aggregates that might be present in solution, being, in spite of that, a very good technique for relative

comparison studies (41).

26

3 MATERIALS AND METHODS

3.1 MATERIALS

3.1.1 Biologicals

To replicate the formulations as well as to assess the stability of the antibody without additives, the

therapeutic antibody used was Gammanorm from Octapharma with a nominal concentration of 165

mg/mL of human normal immunoglobulin G. The purity degree is 95%, thus corresponding to an

effective concentration of IgG of 156.8 mg/mL. Several subclasses of IgG are present in the following

proportions: IgG1 59%, IgG2 36%, IgG3 4.9% and IgG4 0.5%. The formulation contains a maximum of

82.5 μg/mL of IgA. The list of excipients includes glycine, sodium chloride, sodium acetate (resulting in

approximately in 2.5 mg of sodium per mL of solution) and water for injection (WFI). The product has a

shelf life of 3 years and should be stored at 2-8°C.

3.1.2 Chemicals

The chemical compounds used throughout this thesis are listed in Table 3.1

Table 3.1 – Specifications of the chemical compounds used.

Substance Chemical Formula Purity Molecular

Weight (Da) Company

Glycine NH2CH2COOH ≥99.0% 77.05 Sigma

D-trehalose dihydrate C12H22O11 · 2H2O ≥99.0% 378.33 Sigma

Maltose monohydrate

crystalline C12H22O11 · H2O - 360.32 Merck

Sucrose C12H22O11 ≥99.5% 342.31 Sigma

L-histidine C6H9N3O2 ≥99.0% 155.15 Sigma - Aldrich

27

3.2 METHODS

3.2.1 Sample Preparation

In order to achieve the initial concentrations of 1, 5, 10 and 15 g/L, samples were prepared by

dilution of the stock solution of IgG. For the experiments without additives the preparation was done

with PBS (10 mM phosphate, 150 mM NaCl, pH ≈ 7.4). For the second set of experiments samples were

prepared with PBS with each additive added individually in concentrations of 10% (w/v) for glycine,

maltose, sucrose and D-trehalose, and of 2% (w/v) for L-histidine due to its lower. For comparison

purposes maltose was also studies at a 2% (w/v) concentration.

3.2.1.1 Stability Assessment of IgG without additives

Temperature Assay I – Incubation at 60°C

The solutions were incubated in a bath at constant temperature of 60°C and 200 μL samples

(duplicates from each tube) were collected every 2 hours for a period of 50 hours. The samples were

diluted 1:5 in PBS and then stored at 4°C until further analysis to assess the irreversible temperature

induced denaturation.

Temperature Assay II – Incubation at 70°C

The incubation took place in a bath at constant temperature of 70°C and 200 μL samples were

collected (from each of the two tubes incubated) every 2 hours up to a total time of 36 hours. The

samples were diluted 5 times by the addition of 800 μL of PBS and then stored at 4°C until further

analysis to assess the irreversible temperature induced denaturation.

pH Assay I – Incubation at pH 3.02

Every 2 hours, 200 μL samples (duplicates from each tube) were collected up to a total time of 50

hours. The samples were diluted 5 times by the addition of 800 μL of PBS (pH 7.4). Samples were kept

at all times at 4°C to eliminate any temperature induced denaturation effect. Preparation of PBS at pH

3.02 was achieved by the addition of HCl until the pH was approximately 3.

pH Assay II – Incubation at pH 2.1

Samples of 200 µL were recovered at time intervals of 1 hour up to a total time of 9 hours. The

samples were diluted 5 times with PBS (pH 7.4) and 5 µL of NaOH 1M were added to restore the pH

back to approximately 7 in order to stop the denaturation process and to allow a representative

28

analysis. Samples were kept at all times at 4°C to eliminate any temperature induced denaturation

effect. PBS at pH 2.1 was prepared by gradual addition of HCl until the solution reached a pH of

approximately 2.

Mechanical Shear Stress Assay

The mechanical shear stress assay was performed in a vorticular flow filtering system, Benchmark

Gx from Membrex (Garfield, N.J., USA) without a filtration membrane. 200 µL samples were collected

at several times of circulation - 0, 2, 5, 10, 20 and 30 minutes - in order to evaluate the resistance of

IgG against mechanical shear stress induced degradation. The samples were diluted 5 times with PBS

stored at all times at 4°C to eliminate any temperature induced denaturation effect.

The recirculation rate of the IgG solutions (100 mL) through the annulus was 116.4 mL/min, with a

corresponding superficial velocity of 0.003 m/s. The recirculation of the solutions was accomplished

with a peristaltic pump (from Watson-Marlow, Massachussets, USA).The operating velocities of the

inner cylinder of the CCDS were 1000, 2000 and 3000 rpm. Two initial concentrations of IgG were

tested: 1 g/L – tested at the three velocities previously mentioned - and 15 g/L – tested only at 3000

rpm.

3.2.1.2 Stability Assessment of IgG with additives

Temperature Assay – Incubation at 70°C

The incubation took place in a bath at constant temperature of 70°C and 200 μL samples were

collected from each tube every 2 hours up to a total time of 8 hours. The samples were diluted up to 5

times with PBS and then stored at 4°C until further analysis to assess the irreversible temperature

induced denaturation.

pH Assay – Incubation at pH 2.1

200 µL samples were recovered at time intervals of 1 hour up to a total time of 9 hours. The

samples were diluted 5 times with PBS and 5 µL of NaOH 1M were added to restore the pH back to

approximately 7 in order to stop the denaturation process and to allow further analysis. Samples were

kept at all times at 4°C to eliminate any temperature induced denaturation effect. The several buffers

used were prepared by gradual addition of HCl until the solution reached a pH of approximately 2.

29

3.2.2 Analytical Methods

3.2.2.1 Protein A and Protein G Affinity Chromatography

The protein quantification was conducted in an Äkta 10 System (GE Healthcare, UK) with an

analytical POROS Protein A Affinity Column and/or an analytical POROS Protein G Affinity Column both

from Applied Biosystems (Foster City, CA, USA). Samples from the initial concentration of IgG of 5 g/L,

10 g/L and 15 g/L were further diluted with PBS to reach approximately the same initial antibody

concentration of 200 mg/L before analysis. IgG adsorption to the column was performed in PBS.

Elution was accomplished with the desorption buffer (12 mM HCl, 150 mM NaCl, pH 2-3). After analysis

the column was re-equilibrated with the storage buffer (10 mM NaH2PO4, 0.02% NaN3, pH ≈ 7.4).

Absorbance was measured at 280 nm and IgG concentration estimated based on a calibration curve

prepared by dilution of a 1 g/L stock solution of IgG.

3.2.2.2 Circular Dichroism

The CD spectra were measured with an Applied Photophysics spectropolarimeter, model PiStar –

180 (Leatherhead, UK). For the far-UV measurements (200-250 nm), a Quartz cuvette (Hellma, USA)

with a light path of 1 mm was used. The antibody concentration in the measured samples was of 100

mg/L. In the glycine and maltose assays at pH 2.1, the IgG concentration had to be reduced to 80 mg/L

due to buffer interference in the spectra. Furthermore, measurements for the glycine and maltose at

pH 2.1 assay were done from 205 to 250 nm. For the L-histidine assay, no CD measurements were

taken due to high buffer interference which prevented obtaining reproducible results.

The CD spectra obtained are the result of the average of 20 scans – except for pH assay I and

temperature assays I and II, in which only 10 scans were retrieved. The monochromator bandwidth

was set to 2 nm and each measurement done with a 1 nm step. A blank was performed with the

corresponding control buffer. A tool from the Pro-Data Analysis software was used to smooth the

curves by a parameter of 10.

The mean residue ellipticity was determined by equation (2.10) using a mean residue weight of

113.16 Da (44).

3.2.2.3 Dynamic Light Scattering

The dynamic light scattering experiments were carried out using a Malvern Zetasizer nano- ZS

(Malvern Instruments Ltd, Worcestershire, UK) with a 5 mW He-Ne laser (633 nm) and a fixed

scattering angle of 173 degrees. The direct measurements of the particle size were performed at a

constant temperature of 25°C. The particle size was determined for IgG in PBS and also for IgG after 9h

of incubation at pH ≈ 2 and pH ≈ 11 (in PBS and in PBS with 10% (w/v) of maltose for both pH values).

30

Similar measurements were also performed for IgG solutions in PBS without and with 10% (w/v) of

maltose after incubation at 70°C for a period of 8 hours. For the latter, samples were collected at

regular intervals in order to monitor the formation of aggregates over time. The IgG concentrations

studied in this assay were of 1 and 15 g/L.

For the measurements 1.5 mL disposable polystyrene cuvettes were used. The number of scans was

determined by the equipment according to the characteristics of each sample. The results obtained

were processed and analyzed with the Zetasizer nano-ZS software.

31

4 RESULTS AND DISCUSSION

4.1 STABILITY ASSESSMENT OF IGG WITHOUT ADDITIVES

Ideally the determination of the biological activity of IgG would be performed based on the affinity

of IgG for a specific antigen in order to achieve a higher specificity of the results. However, since the

tests are performed with a mixed population of antibodies, the previous would be impossible from a

practical approach. For that reason, the determination of the IgG concentration in the samples was

done based on the affinity of the Fc fragment of IgG for Protein A and/or Protein G. This mixed

population of antibodies may also be accountable for some variability in the results, given that the

concentration of the different IgG subclasses may vary for each sample and each subclass of IgG may

present a distinct resistance facing different denaturation conditions.

In order to standardize the results from all experiments, the data will be presented not in the form

of concentration but in the form of the percentage of Retained Biological Activity which can be defined

as the ratio between the IgG concentrations of the sample under analysis and the initial sample. In

other words, the Retained Biological Activity represents the fraction of the initial IgG molecules that

remain biologically active at a given time.

4.1.1 Thermal induced denaturation

The thermal induced denaturation was firstly assessed by the incubation of IgG in PBS at 60°C for a

period of time of 50 hours. The evolution of the percentage of Retained Biological Activity over time

can be consulted in Figure B.1 (Appendix B).

Figure 4.1, which shows the retained biological activity for the studied concentrations of IgG after

50 hours of incubation, suggests that there are no major losses of biological activity. In fact, the lowest

value of biological activity, obtained for 10 g/L, was of 74% having the remaining samples at least 89%

of active molecules.

32

Figure 4.1 – Retained Biological Activity (%) after 50 hours of incubation at 60°C of IgG solutions in PBS (pH 7.4)

with the initial concentrations of 1, 5, 10 and 15 g/L.

Despite this minor denaturation, it was important to evaluate the secondary structure of the IgG

molecules in order to determine whether any loss of structure had occurred. This was performed by CD

spectroscopy.

Analysing the graphics shown in Figure 4.2A-D, it is seen that the CD spectra of intact IgG – before

incubation at 60°C - presented are that of a typical immunoglobulin, with a minimum at approximately

217nm, a maximum close to 200 nm and zero intensity around 207 nm, thus representing a high

content of β-sheet. For the initial concentration of IgG of 10 g/L, the minimum at 217 nm broadens and

shifts to a lower wavelength and the wavelength corresponding to zero intensity also shifts to a lower

value. The last modification occurs as well for the sample of initial concentration of 15 g/L which also

shows an overall decrease in the ellipticity. Nevertheless, the general shape of the curve remains

similar to the typical β-sheet curve for all the samples. The CD data is thus in conformity with the

results obtained by HPLC, suggesting that only minor denaturation of the IgG molecules occurs. These

minor denaturation is however more important for higher concentrations of IgG.

90

74

89

0

10

20

30

40

50

60

70

80

90

100

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG Concentration

1 g/L 5 g/L 10 g/L 15 g/L

33

Figure 4.2 – Far-UV CD spectra of IgG in PBS (pH 7.4) before incubation () and after 50 hours of incubation

() at 60°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

The thermal induced denaturation of IgG was then studied at more extreme conditions - incubation

at 70°C for 36 hours. Under these conditions no more than 9% of the IgG molecules remained

biologically active indicating a much higher extent of denaturation. The evolution of the percentage of

Retained Biological Activity over time can be consulted in Figure B.1 (Appendix B).

Observing Figure 4.3 one can conclude that the denaturation of IgG at 70°C is fast and doesn’t

change significantly between 8 and 36 hours of incubation. In fact, the major loss of biological activity

is almost immediate and it occurs between 0 and 2 hours of incubation, as can be seen in Figure 4.3

and Figure B.2, that plots the percentage of active molecules over time of incubation. Therefore, the

total time of incubation at 70°C adopted for further assays and data comparison was of 8 hours.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

(A)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

(B)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

(C)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

(D)

34

Figure 4.3 – Retained Biological Activity (%) after 2, 8 and 36 hours of incubation at 70°C of IgG solutions in PBS

(pH 7.4) with the initial concentrations of 1, 5, 10 and 15 g/L.

Upon incubation at 70°C for both 8 and 36 hours, it was clearly observed the precipitation of IgG, as

it can be seen in Figure 4.4.

Figure 4.4 – Precipitation occured after 8 hours of incubation at 70°C for the initial concentrations of IgG in

solution of: (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

IgG unfolds as a consequence of the heat treatment, causing the exposure of the hydrophobic

residues that were located in the interior of the protein (17). The exposure of these hydrophobic

patches in solution triggers the formation of aggregates through intermolecular hydrophobic binding

(16). The aggregates result from the partial unfolding of the Fab-fragments, which then become

associated leaving the Fc-fragments exposed to the aqueous environment (1), (16). The formation of

these aggregates result in the observed precipitation of the IgG molecules.

It has been demonstrated that the heating process of IgG doesn’t lead to a complete unfolding into

an extended polypeptide chain, but leaves a significant part of the IgG molecule in a globular or

8

26 19

24

4 8 6 8 9 7 5 5

0

10

20

30

40

50

60

70

80

90

100

1 g/L 5 g/L 10 g/L 15 g/L

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG Concentration

2h 8h 36h

35

corpuscular form. This occurs because the rate of aggregation is so fast that the molecules become

locked in the aggregates formed before being completely denatured. The aggregation step is thus the

step that induces the irreversibility of the heating denaturation process (reversible unfolding step,

followed by an irreversible process that locks the unfolded protein in a state from which it can no

longer refold) (16).

It is clear from Figure 4.4 that more precipitation occurs for higher values of initial concentration of

IgG in solution that for lower values. Now, if the irreversibility of the denaturation process is due to the

aggregation step which ultimately leads to the precipitation of the antibodies in solution, and that the

amount of precipitate is dependent on the initial concentration, it is patent that increasing the initial

concentration of IgG in solution will ultimately lead to a higher irreversible denaturation.

In face of such extent loss of biological activity, it was important to evaluate the changes that surely

have occurred in IgG’s secondary structure. The CD spectra obtained for the samples are presented in

Figure 4.5A-D.

In all the images it’s patent that a major change occurs in the secondary structure given the

difference between the curves of the samples before and after 8 hours of incubation at 70°C.

Observing Figure 2.6, one can easily identify the similarities between the curve of the random coil

motif and the curve correspondent to the samples after incubation at 70°C. In general, the CD

spectrum of random coils shows only negative contribution, approaching to zero around 230 nm. The

samples after incubation present a curve with only negative contributions but generally close to zero

while the initial ones present a curve in general similar to the one of a regular immunoglobulin.

Furthermore, the ellipticity near 206.5 nm is negative which is consistent with an increased

contribution of random coil motifs. The previous observations suggest that the samples after

incubation possess a higher content of random coils and virtually none of β-sheet. This depletion in β-

sheet content shows that denaturation had occurred in a large extent.

36

Figure 4.5 - Far-UV CD spectra of IgG in PBS (pH 7.4) before incubation () and after 8 hours of incubation ()

at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

Accounting the two temperatures tested to induce heat denaturation of IgG, at 70°C it appeared to

occur irreversible denaturation of the Fc-fragment of IgG while at 60°C the Fc-fragment seems to suffer

only minor alterations. Based on the fact that the percentage of Retained Biological Activity is

determined on the affinity of Protein A and Protein G to the Fc fragment of IgG, it is clear that if the

percentage of retained biological activity is diminished, the Fc fragment has suffered alterations since it

no longer binds to Protein A or G. Nonetheless, at 60°C there are mild changes in the secondary

structure of IgG molecule, suggesting that other fragments of the IgG molecule may be affected upon

incubation at this temperature. This situation may arise from IgG being a multi-domain protein which

results on the fact of (at least) two existing domains denature at different temperatures. In fact, these

evidences are consistent with results obtained in previous studies which suggested that the Fab and Fc

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

37

fragments appear to denature independently from each other, occurring at the temperatures of 61°C

and 71°C, respectively (16).

4.1.2 pH induced denaturation

The pH firstly chosen to assess pH induced denaturation was a pH of approximately 3, since

this is the common elution pH of protein A.

Figure 4.6 – Retained Biological Activity (%) after 50 hours of incubation of IgG solutions with PBS at pH 3.02


After 50 hours of incubation at pH 3.02 at 4°C, it was only observed a residual loss of biological

activity for the initial concentrations of IgG of 5 and 10 g/L, respectively with 98% and 97% of the initial

molecules remaining biologically active (Figure 4.6). The CD spectra (Figure 4.7A-D) are in agreement

with the data previously shown, since in all of them the curves for the samples before and after

incubation are for all purposes coincident. This indicates that the secondary structure of IgG remains

intact after 50 hours of incubation in PBS at pH 3.02.

98 97

0

10

20

30

40

50

60

70

80

90

100

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG Concentration

1 g/L 5 g/L 10 g/L 15 g/L

38

Figure 4.7 - Far-UV CD spectra of IgG in PBS before incubation () and after 50 hours of incubation () at pH

3.02 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

Since there was no denaturation at pH 3.02 and based on a previous study that revealed an increase

on random coil configuration upon exposure of IgG molecules to a pH of 2 at the temperature of 20°C

(44), it was chosen to incubate the IgG solutions in PBS at pH 2.1 at 4°C – to, as previously stated,

eliminate any contribution of temperature to the induced denaturation.

After a period of 9 hours, the percentage of retained biological activity was significantly lower when

compared with pH ≈ 3, except for the initial concentration of 15 g/L where all IgG remained biologically

active. The loss of activity was especially more pronounced for the concentrations of 5 and 10 g/L of

IgG.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

39

Figure 4.8 - Retained Biological Activity (%) after 9 hours of incubation of IgG solutions with PBS at pH 2.1 with

the initial concentrations of 1, 5, 10 and 15 g/L.

The CD spectra of these samples (Figure 4.9A-D) are in concordance with the results shown above.

For the solution with the initial concentration of IgG of 15 g/L the curves of the samples before and

after incubation are very similar and, for most values of wavelength, superimposed, indicating that the

secondary structure of the molecule didn’t suffer alterations. The spectrum corresponding to the IgG

solution with the initial concentration of 1 g/L shows a curve after incubation with the same shape but

lower values of ellipticity than the curve of the sample before incubation at pH 2.1. This suggests that a

change in the secondary structure has occurred but not in major extension, which can be accurately

related to the HPLC data that revealed a value of retained biological activity relatively high (82%). The

samples that revealed a significantly higher loss of function – 5 and 10 g/L – present CD spectra with

more distinct curves of samples before and after incubation: the minimum observed at 217nm

broadens and shifts to a lower wavelength and the wavelength corresponding to zero intensity also

shifts to a lower value. The pronounced differences between the curves indicate a structural change in

a higher extension which had already been suggested by the previous results. It has been reported the

formation of a new, well-defined IgG structures upon exposure at low pH values (pH < 3). The

protonation of amino acid side chains is supposed to cause the reorganization of the native state into a

so called A-state. This new acid-induced structure is characterized by a high degree of secondary

structure, increased accessibility of hydrophobic clusters, increased fluorescence intensity and a

native-like hydrophobic surrounding of the aromatic rings. The A-state is described to be relatively

compact [ (17) , (18), (44)].

82

64

43

0

10

20

30

40

50

60

70

80

90

100

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG Concentration

1 g/L 5 g/L 10 g/L 15 g/L

40

Figure 4.9 - Far-UV CD spectra of IgG in PBS before incubation () and after 9 hours of incubation () at

pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

4.1.3 Mechanical Shear Stress induced degradation

Shear forces encountered by Taylor vortices can partition proteins to the air-water interface which

encourages partial unfolding on exposure to the more hydrophobic air phase (3). So, before assessing

the effect of mechanical shear stress on IgG, it was pertinent to determine the extent of the forces

imposed on the IgG molecules during the assay. Resorting to equations (2.1), (2.2) and (2.3) it was

possible to estimate the Taylor number, the shear rate and the shear stress imposed on the IgG

molecules.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

41

Table 4.1 – Taylor number, maximum shear rate (s-1

) and shear stress at the wall of the inner rotating cylinder

(Pa) imposed on the IgG molecules during the mechanical shear stress assay for the different angular speeds of

the inner cylinder tested.

ω (rpm) 1000 2000 3000

Ta 1664 3327 4991

γm (s-1) 9.40x103 2.66x104 4.88x104

τm (Pa) 9 27 49

Initially, tests were performed recirculating an IgG solution with a concentration of 1 g/L in a

concentric-cylinder shear device with an angular speed of the inner cylinder of 1000 rpm.

Figure 4.10 – Far-UV CD spectra of IgG in PBS at pH 7.4 for several periods of time spent in the shear field. The

initial concentration of IgG was of 1 g/L and the angular speed of the inner rotating cylinder of 1000 rpm.

The CD spectra (Figure 4.11) show the behavior of a regular immunoglobulin, being the curve

representative of a high content of β-sheet. The curves are similar in shape and ellipticity values

suggesting that the secondary structure of IgG is maintained throughout the time spent in the shear

field. It is then implied that IgG is capable to resist the induced degradation of shear forces of

approximately 9 Pa.

The angular speed of the inner cylinder was then raised to 2000 rpm in order to evaluate whether

an enhancement in this parameter and, consequently, in the shear stress imposed on IgG would cause

degradation. Figure 4.11 shows again the typical behaviour of an immunoglobulin with a high content

of β-sheet that is maintained throughout the period of permanence in the shear field, showing no

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

0 min 2 min 5 min 20 min 30 min 10 min

42

alterations in the secondary structure. The results indicate that IgG in solution at a concentration of 1

g/L is also capable to resist shear stress up to 27 Pa.



Given that no degradation was observed for 2000 rpm, the inner cylinder angular speed was then

raised to 3000 rpm.



The CD spectra show, once again, the behavior of a regular immunoglobulin with elevated β-sheet

content as well as no alterations in the secondary structure of the IgG. Thus, it is possible to infirm that

IgG in solution with a concentration of 1 g/L is capable to resist an imposed shear stress of 49 Pa.

-5

-3

-1

1

3

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)


-5

-3

-1

1

3

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm) 0 min 2 min 5 min 20 min 30 min 10 min

43

Since for the initial concentration of 1 g/L no degradation has occurred, it was tested the

mechanical shear stress induced degradation at the highest concentration used in these studies – 15

g/L – and at the highest angular speed of the inner rotating cylinder – 3000 rpm – in order to create

more extreme conditions.



Figure 4.13 shows that, for the initial concentration of 15 g/L, no alterations in the secondary

structure over the time spent in the shear field, being the general aspect of the curve correspondent to

high β-sheet content.

The percentage of molecules that remain biologically active was plotted for all the samples tested.

The retained biological activity was of or close to 100% for all the situations, supporting the data

obtained from CD spectroscopy. These combined data are strong indicators that the IgG molecule is

resistant against mechanical sheer stress induced degradation, being able to sustain shear stresses up

to 49 Pa.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)


44

Figure 4.14 – Retained Biological Activity (%) of IgG after a period of circulation in the shear field of 30 minutes.

Previous studies on different proteins showed that the effect of shear and shear rate (>105 s-1) on

protein activity (rhGH and rhDNase) wasn’t significant (28). The results obtained contribute to include

IgG in the lot of proteins capable to resist the partial unfolding induced by shear forces.

When compared to other biomolecules, namely with Plasmid DNA (pDNA), IgG shows better

response facing the shear stress induced degradation. Previous studies report that of pDNA suffers

severe injuries upon shear forces of 10.9 Pa (26) whereas IgG is able to resist degradation induced by

shear forces up to 49 Pa.

4.2 STABILITY ASSESSMENT OF IGG WITH ADDITIVES

In order to evaluate which of the selected additives commonly used in human therapeutic

antibodies formulations is the best stabilizer, IgG solutions with each individual additive were

incubated in the denaturation inducing conditions previously studied (70°C and pH ≈ 2).

4.2.1 Thermal induced denaturation

The retained biological activity of IgG after incubation at 70ºC after a period of 8 hours is shown in

Figure 4.15 for different initial IgG concentrations – 1, 5, 10 and 15 g/L.

0

10

20

30

40

50

60

70

80

90

100

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG Concentration

1000 rpm (1g/L) 2000 rpm (1g/L) 3000 rpm (1g/L) 3000 rpm (15g/L)

45

Figure 4.15 - Retained Biological Activity (%) after incubation at 70°C for a period of 8h of IgG solutions with the

initial concentrations of 1, 5, 10 and 15 g/L. The buffers used consisted in PBS with 10% (w/v) of additive

except for L-histidine in which the concentration of additive in the buffer was of 2% (w/v).

After 8 hours of incubation it was still observed a quite extent denaturation induced by heat, except

for the initial IgG concentration of 1 g/L incubated with maltose in which all the IgG molecules

remained biologically active. This means that maltose is able to fully stabilize IgG at this concentration

against heat induced denaturation. Generally, all the values of retained biological activity obtained

were higher in the presence of these stabilizers when compared with IgG in PBS. Only L-histidine

showed lower percentage of biologically active molecules than IgG alone – at 5 and 15 g/L – and equal

values at 10 g/L.

Figure 4.15 reveals that maltose is the more suitable stabilizer against thermal induced

denaturation of IgG since it achieved the highest values of retained biological activity for most of the

initial concentrations. In fact, maltose allowed a complete stabilization of IgG at the initial

concentration of 1 g/L without any loss of biological function. For the initial concentration of 15 g/L,

glycine had a better performance in stabilizing IgG but the value obtained for the retained biological

activity was very close of the one obtained with maltose.

Once again, as observed before, the higher the initial concentration of IgG, the higher was the

extent of the denaturation.

It was also observed, for all the solutions under study, the formation of IgG precipitates after the 8

hours of incubation at 70°C in amounts correlated with the initial concentration of IgG in solution, as

had already happened upon heat treatment of IgG in PBS.

4 8 6 8

17 17

8 9

47

24

16 18

100

36 33

15

46

7 7 9 7 6 6 5

0

10

20

30

40

50

60

70

80

90

100

1 g/L 5 g/L 10 g/L 15 g/L

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG Concentration

Control Trehalose Glycine Maltose Sucrose L-histidine

46

In view of the fact that the buffer concentration of L-histidine was quite lower than the

concentration of the other additives - 2% (w/v) compared to 10% (w/v) - it was pertinent to evaluate

whether maltose at the same buffer concentration as L-histidine would still provide higher stabilization

against thermal induced denaturation than L-histidine, as shown in Figure 4.16.

Figure 4.16 - Retained Biological Activity (%) after incubation at 70°C for a period of 8h of IgG solutions with the

initial concentrations of 1, 5, 10 and 15 g/L. The buffers used consisted in PBS with 2% (w/v) of additive.

After 8 hours of incubation at 70°C, maltose still revealed a better behavior as a stabilizer of IgG

against heat denaturation, with fairly higher values of retained biological activity than L-histidine. It is

noticeable that the values of the retained biological activity for maltose at a concentration of 2% (w/v)

are significantly lower than for a concentration of 10% (w/v), which denotes the importance of the

stabilizer concentration in the resistance provided against denaturation.

The effect of the presence of a stabilizer on the changes occurring in the secondary structure of IgG

upon heat denaturation was evaluated by CD spectroscopy. The CD spectra for incubation with

maltose are shown in Figure 4.17A-D. For the remaining additives studied the CD spectra are presented

in Appendix C. It is pertinent to observe that it was verified significant structural changes upon heat

treatment in the presence of all the remaining additives.

7 6 6 5

21

10 10 9

0

10

20

30

40

50

60

70

80

90

100

1 g/L 5 g/L 10 g/L 15 g/L

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG concentration

L-histidine 2% (w/v) Maltose 2% (w/v)

47

Figure 4.17 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of maltose (pH ≈ 7) before incubation () and

after 8 hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15

g/L.

The curve observed for the sample before incubation at 70°C at the initial concentration of 1 g/L is

consistent with high β-sheet content. Despite the ellipticity values between 200 and 220 nm of the

sample after 8 hours of incubation at 70°C being much lower and the minimum at 217 nm being a little

broader when compared with the sample before incubation, the curve still presents a similar shape to

the initial one, indicating that some content of β-sheets is maintained. However, at approximately

206.5 nm, where a typical β-sheet shows an intensity of zero, the curve presents a negative intensity at

that wavelength, thus revealing the existence of random coil and maybe α-helix motifs (which exhibit

negative contributions at that wavelength).

For the remaining studied concentrations, the CD spectra show a slight improvement on the

conservation of the secondary structure upon comparison with the CD spectra for IgG without any

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm)

48

additive. While in the one without any additive the curve resembled the typical curve of a random coil

motif, in the presence of maltose the curve shows some resemblance in shape with the curve of a β-

sheet motif, with positive contributions between 200 and approximately 210 nm and a minimum, even

though broader than the one of a typical immunoglobulin, around 215 nm. Even so, the distinction

between the curves of the initial and final samples is fairly pronounced suggesting that changes in the

secondary structure have definitely occurred.

4.2.2 pH induced denaturation

The retained biological activity of IgG after incubation at pH 2.1 after a period of 9 hours is shown in

Figure 4.19 for different initial IgG concentrations – 1, 5, 10 and 15 g/L.

Figure 4.18 - Retained Biological Activity (%) after incubation of IgG solutions with the initial concentrations of

1, 5, 10 and 15g/L at pH ≈ 2 for a period of 9h. The buffers used consisted in PBS with 10% (w/v) of additive

except for L-histidine in which the concentration of additive in the buffer was of 2% (w/v).

Upon incubation at pH ≈ 2 in the presence of the several additives, it was observed some variability

in the stabilization provided by the additives. Depending on the initial concentration of IgG in solution,

the retained biological activity was largely different for the same additive. For example, D-trehalose

behaved fairly well except for the solution of 10 g/L in which the percentage of retained biological

active was drastically lower than for the other initial concentrations and than for IgG alone itself.

Observing Figure 4.18, one can infer that maltose would also be the best stabilizer against pH induced

denaturation since it allows achieving a percentage of molecules biologically active after incubation

82

64 65

97 91

41

82

64 65

96 97

66 71

38

68 71 70 69

0

10

20

30

40

50

60

70

80

90

100

1 g/L 5 g/L 10 g/L 15 g/L

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

IgG concentration

Control Trehalose Glycine Maltose Sucrose L-histidine

49

equal or very close to 100% and it is, furthermore, the additive that showed less variability with the

initial concentration of IgG in solution.

The secondary structure of IgG after incubation was studied to determine whether the remaining

biologically active molecules maintained or not their original secondary structure. Figure 4.19A-D

shows the CD spectra for IgG with maltose as the additive – since it showed the higher percentages of

retained biological activity – while the CD spectra of IgG with each of the remaining additives are

shown in Appendix C. It should be mentioned that, in general, the behavior of the remaining additives

was consistent with the retention of the biological activity achieved in their presence.

Figure 4.19 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of maltose before incubation () and after

9 hours of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

The CD spectra exhibit curves identical to those of a typical immunoglobulin with high β-sheet

content characterized by a minimum near 217 nm and zero intensity at approximately 206.5 nm.

Furthermore, in Figure 4.19A-D one can observe that the curves of the samples before and after

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03 d

eg

cm2

dm

ol-

1)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245[θ

] m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03 d

eg

cm2

dm

ol-

1)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03 d

eg

cm2

dm

ol-

1)

λ (nm)

50

incubation are virtually equal, being almost superimposed. Hence, it is possible to assert that no

alterations on the secondary structure have occurred.

Considering the results obtained both for the thermal and pH induced denaturation it is fair to

conclude that maltose is the most suitable additive for IgG stabilization from amongst the additives

studied. This hardly consists in a great surprise given that sugars have been used for many years as

stabilizing agents for the maintenance of the biological activity of several macromolecules.

It has been recognized that sugars increase the transition temperatures of some proteins in

aqueous solution, being this effect ascribed to the induction of a decrease in H-bond rupturing

potency. Another source of stabilization is attributed to the preferential interaction of the protein with

solvent components in the presence of high concentration of additives - such as the concentration of

maltose used in the studies performed. The surface free energy perturbation by sugars plays a

predominant role in their preferential interaction with proteins. Other contributing factors are the

exclusion volume of the sugars and the chemical nature of the protein surface.

The protein preferential hydration is a common feature of aqueous sugar solutions within the sugar

concentration range ordinarily employed, regardless of the type of protein and the solvent conditions

used. In a three-component system, preferential hydration of a protein is a good indication that the

third component is a stabilizer of the structure of the macromolecule. The stabilizing effect of the third

solution component on proteins can be explained in terms of the positive change in the chemical

potential of the proteins, induced by the addition of these substances. Since the addition of sugars

results in a positive surface free-energy change, and this change is assumed to increase with an

increase in the surface area of the macromolecule, the macromolecule in the denatured state will

experience a greater increase in chemical potential when sugar is added than that in the native form,

i.e., contact of sugar with the denatured state is thermodynamically more unfavorable than with the

native state. This must be reflected in the equilibrium constant, which, in sugar solutions, will shift to a

value favoring the native conditions. It is also strongly suggested that the cohesive force of sugars

responsible for the increase in the surface tension of water is a very important factor governing the

preferential interaction of proteins with solvent components in aqueous sugar solutions and hence the

stabilization of proteins (45).

4.3 AGGREGATION STUDIES FOR IGG

Aggregation properties of IgG solutions are dependent on the conformational stability and

structural properties of individual IgG molecules (19). In a large multi-domain protein like IgG,

51

relatively gentle conditions can be sufficient to initiate aggregation. Partially unfolded states are much

more susceptible to aggregation than the native or completely unfolded state, due to the exposure of

contiguous hydrophobic regions that are hidden in the native state or absent in the denatured state

(3). It is thus extremely important to study the effect on aggregation of environmental factors capable

to lead IgG to these partially unfolded states, such as pH – approximately 2 and 11 - and heat exposure

- 70°C – by the measurement of the size of particles in solution.

Ideally, the size distribution intensity should be the one used to determine the particle size, as those

data are closest to what is really measured (43). However, if the plot shows a substantial tail, or more

than one peak, then Mie theory (named after Gustav Mie) can make use of the input parameter of

sample refractive index to convert the intensity distribution to a volume distribution, which will then

give a more realistic view of the importance of the tail or second peak present (42). Based on that, the

data presented for the pH assay correspond to the mode of the curves of the size distribution by

volume.

On the other hand, the size presented in the temperature assay corresponds to the average size

measured because, in most cases, both the plots of intensity and volume distributions present several

peaks. In addition, the presence of the larger precipitates may compromise the results by masking the

size of smaller aggregates in solution and so, the most adequate approach would be to consider the

contribution of all the peaks that correspond to different identities in solution. Nonetheless, the size

distributions are useful to assess the number of different identities present in solution that provide a

contribution to the average size determined.

This makes it, however, impossible to compare the data of the pH assay with the data from the

temperature assay because only data analyzed with the same distribution can be compared (43).

4.3.1 pH effect on IgG aggregation

When the pH is shifted away from the isoelectric point (in which all the molecules are electrically

neutral), the molecules acquire charges – positive or negative, depending whether the pH is lower or

higher than the isoelectric point. At high pH (pH ≈ 11), the IgG molecules are negatively charged since

the pH is higher than the isoelectric point of IgG (pI ≈ 9). At low pH (pH ≈ 2), a similar situation occurs

but, this turn, with positively charged molecules given that the pH is lower than IgG’s isoelectric point.

It is common knowledge that molecules with the same charge tend to repel each other, so it would be

expected that with the increase of the molecule charge (either positive or negative) the tendency for

aggregation would decrease. It is also expectable that the higher the initial concentration of IgG in

solution, the lower the tendency for aggregation, since there are a higher number of molecules present

52

in solution that may become charged causing more repulsion as result. It has been reported indeed

that oligomerization/aggregation of polyclonal IgG decreases at lower pH, when distancing from the

isoelectric range, while the tendency for denaturation increases (19).

The plots of the size distribution by intensity and by volume for IgG before and after incubation at

pH 2 or 11 in the absence and presence of maltose for two initial concentrations of IgG – 1 g/l and 15

g/l are given in the Appendix D.

Figure 4.20 – Size (nm) of IgG in PBS at pH 2.1 before incubation and after 9 hours of incubation in PBS at pH

2.1 and PBS with 10% (w/v) maltose at pH 2.1.

The tendency expected was not observed upon incubation in PBS at pH 2.1 in which, for both initial

concentrations, the measured size after 9 hours of incubation was higher than for the sample prior to

incubation (Figure 4.20). A size reduction would also be expected due to the formation of the

previously mentioned A-state at low pH values which consists in a fairly compact structure. This may be

due to the fact that these measurements were performed at a salt concentration of 150 mM NaCl, and

thus charge shielding may have occurred. Nevertheless, in the presence of maltose, for both initial

concentrations of IgG in solution, the particle size was somehow reduced, verifying both the low

tendency for aggregation at low pH as well as the capacity of maltose for inhibiting the aggregation

process more than the effect caused by incubation at low pH values alone (repulsion between

molecules caused by the increase of the number of molecules with the same charge).

12,86 11,69

13,74 13,43 11,92

9,63

0

2

4

6

8

10

12

14

1g/L 15g/L

Size

(n

m)

IgG Concentration IgG in PBS (pH 2.1)

IgG in PBS (after 9 h of incubation at pH 2.1)

IgG in PBS with 10% (w/v) maltose (after 9 h of incubation at pH 2.1)

53

Figure 4.21 - Size (nm) of IgG in PBS at pH 11.3 before incubation and after 9 hours of incubation in PBS at pH

11.3 and PBS with 10% (w/v) maltose at pH 11.5.

Regarding the samples incubated at pH ≈ 11 (Figure 4.21), a reduction in size was observed both in

the presence and absence of maltose when compared to IgG at pH ≈ 11 at the beginning of the

incubation. This goes along with the theoretical previsions. It is further observable that the size

determined in the presence of maltose in solution was smaller than in its absence, especially at the

highest initial concentration of IgG in solution tested, 15 g/L. This, once again, indicates that maltose

has the ability to inhibit the aggregation process at high pH values in further extension than the pH

effect alone.

4.3.2 Temperature effect on IgG aggregation

The plots of the size distribution by intensity and volume for native IgG and IgG after incubation at

70°C in PBS both in the absence and presence of maltose for each initial concentration – 1 g/L and 15

g/L - are shown in Figure D.7 and Figure D.8, respectively.

The average particle size is plotted in Figure 4.22 for native IgG and IgG after 8 hours of incubation

at 70°C in PBS with and without maltose, for the initial IgG concentrations of 1 and 15 g/L.

The reported size of native IgG is ≈ 10 nm (14) which is lower than the sizes determined

experimentally (Figure 4.22). This difference may be ascribed as the result of long term storage at low

temperatures of the commercial sample used, which is commonly accepted to happen (19).

10,92

12,57

10,73

12,17

10,57

7,96

0

2

4

6

8

10

12

14

1g/L 15g/L

Size

(n

m)

IgG Concentration IgG in PBS (pH 11.3)

IgG in PBS (after 9 h of incubation at pH 11.3)

IgG in PBS with 10% (w/v) maltose (after 9h of incubation at pH 11.5)

54

Figure 4.22 – Size (nm) of native IgG and IgG after 8 hours of incubation at 70 °C in PBS and PBS with 10% (w/v)

maltose – pH ≈ 7.

For the initial concentration of 1 g/L, it is only observable one peak in the size distribution plots

correspondent to one individual identity present in solution for both IgG before and after 8 hours of

incubation at 70°C. The average size determined after incubation is much higher than the size of native

IgG. The latter would be expected since, as previously described, the mechanism of heat denaturation

comprises an aggregation step which leads to IgG precipitation. In the presence of maltose, the

average size determined – in which there is also the contribution of only one identity - was significantly

lower than in its absence which is in agreement with the results previously obtained: if aggregation is

prevented to a certain extent, the heat denaturation is expected to be gentler.

For the initial concentration of IgG in solution of 15 g/L, only one peak was found in the size

distribution plots for IgG in PBS, before and after heat treatment, corresponding to only one individual

identity in solution. However, in the presence of maltose three peaks are observed in the size

distribution by intensity plot, revealing the presence of three different identities with distinct sizes that

contribute to the average size determined. The average particle size determined after incubation in

PBS alone was enormously higher than the size of native IgG. The same situation occurred for the

sample incubated in PBS with 10% (w/v) of maltose but, in this case, the particle size was quite smaller

than in its absence.

Comparing the solutions of IgG in PBS with 10% (w/v) maltose, the size determined for 1 g/L was

higher than for 15 g/L which contrasts with the results previously obtained: for 1 g/L the retained

biological activity of the IgG molecules was of 100% while, in contrast, for 15 g/L it was of 15%. In

addition, the changes in the secondary structure observed for the sample with the initial concentration

14,25 13,85

1762

2556

413,9

119,2

0

500

1000

1500

2000

2500

3000

1g/L 15g/L

Size

(n

m)

IgG Concentration Native IgG

IgG in PBS (after 8 h of incubation at 70°C)

IgG in PBS with 10% (w/v) maltose (after 8 h of incubation at 70°C)

55

of 1 g/L were less extent than for 15 g/L. Since aggregation plays an important part in heat

denaturation it should be expectable to observe less aggregation for samples more stable against heat

denaturation.

The variations of the particle size with time were plotted in terms of the fold increase in size (ratio

between the particle size at a given time and the initial particle size) to express the aggregation rate

(Figure 4.23).

Figure 4.23 – Variations in size of IgG in solution over time of incubation at 70°C in hours for the initial

concentration of IgG in solution of (A) 1g/L and (B) 15g/L, both in absence and presence of maltose. The Fold

Increase in Size corresponds to the ratio between the size at a given time and the initial particle size.

It has been suggested in previous studies that at higher heating rate and/or incubation around or

above the denaturation temperature, the unfolding occurs at a higher speed, leading to a high

0

500

1000

1500

2000

2500

0 1 2 3 4 5 6 7 8 9

Fold

Incr

eas

e in

Siz

e

Time (h)

IgG in PBS (1g/L) IgG in PBS with 10% (w/v) maltose (1g/L)

0

500

1000

1500

2000

2500

0 1 2 3 4 5 6 7 8 9

Fold

Incr

eas

e in

Siz

e

Time (h)

IgG in PBS (15g/L) IgG in PBS with 10% (w/v) maltose (15g/L)

56

concentration of (partially) unfolded IgG molecules. The rate of aggregation is correspondingly faster,

and it could well be that at such high aggregation rates IgG molecules are incorporated in the

aggregate before they have had sufficient time for complete unfolding (16). In this experiment the

incubation took place at a temperature near the denaturation temperature (T≈71°C for the Fc-

fragment) and so the rate of aggregation would be expected to be high.

The shape of the aggregation curves is significantly different, fact that might be attributed to the

poor quality of the results given the precipitates effect on the light scattering measurements. Even

though the results comprise a certain error, it is possible to verify the expected high rate of

aggregation upon incubation in PBS alone for the two initial concentrations tested. In the presence of

maltose, the rate of aggregation was significantly lower than in its absence, which correlates perfectly

with the already done observation of maltose being able to stabilize IgG facing heat induced

denaturation.

It is also important to refer that the decreasing aggregation tendency observed for the final period

of incubation may be related to native domains initially present in solution being locked in the formed

aggregates.

Maltose proved to give, once again, a positive contribution to IgG stabilization, being generally able

to reduce the tendency for aggregation. The addition of sugars to the protein solution is a popular

approach to stabilize the protein and thereby reduce access to partially unfolded states that favor

aggregation by hydrophobic contacts. The mechanism is the same described before: the sugars are

preferentially excluded from the surface of the protein hence favoring a compact state (3). Protein

structures that exist in such a confined volume tend toward their state of minimum hydrodynamic

volume, being the protein structure stabilized by a confined environment (17).

At last, it is important to include an extra remark. Even though the results for the pH and the

temperature effect on aggregation cannot be quantitatively compared (once the sizes measured do

not result from the same distribution), a qualitative comparison indicates a higher tendency for

aggregation upon heat treatment. This should have been expected given the observed formation of IgG

precipitates at those conditions which didn’t occur upon incubation at pH ≈ 2.

57

5 CONCLUSION AND FUTURE PRESPECTIVES

Long-term stability of IgG depends on the conformational stability and structural properties of

individual IgG molecules (19). The protein chain can assume many possible conformations, the relative

stabilities of which are very sensitive to pH, ionic strength and temperature, amongst other factors (3).

It was observed that IgG is stable upon pH induced denaturation at a pH of approximately 3 but, for

lower pH values (pH ≈ 2) irreversible denaturation occurs in a certain extent. Regarding the thermal

induced denaturation, IgG was found to be relatively stable at 60°C since that most molecules in

solution maintain their biological activity. However, some gentle alterations in the secondary structure

occur upon incubation at this temperature. When the temperature is increased to 70°C extensive

irreversible denaturation occurs given that the major portion of molecules in solution loses their

biological activity and significant changes occur at the secondary structure of IgG. Large IgG

precipitates are also formed after heat treatment at 70°C, being this aggregation step the one that

confers irreversibility to the heat denaturation process. Upon imposed mechanical shear forces, IgG

was capable of maintaining its structure and remaining biologically active, being able to sustain shear

stress up to 49 Pa. Finally, it was observed a tendency for aggregation upon incubation at 70°C at an

elevated aggregation rate which is consistent with the observation made during the heat treatment of

IgG. For the pH effect on aggregation, it was observed a lower tendency for aggregation at elevated pH

(pH ≈ 11), while for a low pH (pH ≈ 2) the particle size increased contrasting with the theoretical

prevision of a lower tendency for aggregation at this pH range.

Maltose at a concentration of 10% (w/v) was found to be the best IgG stabilizer from amongst the

additives studied. In fact it proved to be a quite appropriate stabilizer for IgG since it accomplishes a

full stabilization at low pH (approximately 2), with no loss of biological activity and no alterations in the

secondary structure. It also allows a complete stabilization of IgG in a concentration of 1 g/L against

heat denaturation at 70°C in terms of the ratio of IgG molecules that remain biologically active.

However, the secondary structure showed some alterations after incubation which may only guarantee

that the Fc fragment is stabilized in this situation (since the determination of the retained biological

activity is done based on the affinity of Protein A and G for the Fc-fragment). For the remaining initial

concentrations of IgG in solution the stabilization achieved wasn’t as good but it was still higher than in

the absence of maltose. In general, maltose was also able to reduce the tendency for aggregation both

facing thermal and pH denaturation inducing conditions.

58

It is important to note that the conditions studied are extreme conditions that normally would not

occur in a regular storage environment. This means that if maltose is added to a therapeutic

formulation in a concentration similar to the one studied it should be expected to provide full

stabilization of the IgG molecules, since the environmental conditions are far gentler than the ones

studied.

It should also be noticed that the results from previous studies can be useful as a starting point to

understand the mechanisms adjacent to the induced denaturation processes studied, but should not

be used to strictly compare the resulting data because the type of IgG used in the tests may have a

significant impact on the results obtained.

Future Perspectives

The fact that IgG can resist the shear forces imposed by the CCSD suggests that vorticular flow

membrane modules may be used as an alternative/complementary purification process for IgG.

However, in order to use this system in the downstream process further study is required, namely,

studying the mechanical shear stress induced degradation at higher protein concentrations that many

therapeutic formulations require.

It is known that aggregation is generally quite site specific involving well-defined oligomerization

interfaces but the molecular details of the aggregation process lack some insight. (3) For this reason, at

the present the prevention of aggregation remains largely empirical. Further studies on the

aggregation molecular mechanisms may contribute to find a more specific and more appropriate way

to prevent the formation of aggregates.

Sugars are commonly applied for IgG stabilization, frequently in somewhat high concentrations (5-

10%). However, some sugars, such as maltose and sucrose, may cause adverse effects upon a regular

use which can include renal failure (19). Further work should thus comprise the study of alternative

formulations combining two or more additives that would allow the decrease of the sugar

concentration to a more physiological acceptable one. Alternative compounds that might stabilize the

antibody without endangering patient health by adverse side effects should also be investigated.

Hydrophobic amino acids have been reported as promising alternatives to sugars (19).

59

REFERENCES

[1] 1. Vermeer, A. W.P., Bremer, M. G.E.G. and Norde, W. Structural changes of IgG

induced by heat treatment and by adsorption onto a hydrophobic Te£on surface

studied by circular dichroism spectroscopy. Biochimica et Biophysica Acta. 1425:1-12,

1998.

[2] 2. Antibody Purification Handbook. GE Healthcare Life Sciences. [Online] [Citação: 30

de Março de 2011.]

http://www.gelifesciences.com/aptrix/upp00919.nsf/content/LD_153975280-R350.

[3] 3. Frokjaer, S. and Otzen, D. E. Protein Drug Stability: A Formulation Challenge.

Nature Reviews. 4:298-305, 2005.

[4] 4. Delves, P.J., Martin, S.J., Burton, R.D. and Roitt, I.M. Roitt's essential

immunology. Massachussets, USA : Blackwell Publishing, 2006.

[5] 5. Kayser, F.H., Bienz, K.A., Eckert, J. and Zinkernagel, R.M. Medical

Microbiology. Stuttgart, Germany : Thieme, 2004.

[6] 6. Monoclonal Antibodies: Structures and Characteristics. The University of New

Mexico. [Online] [Citação: 5 de Junho de 2011.]

http://www.unm.edu/~mpachman/Antibodies%20Project/heavy.htm.

[7] 7. Brekke, O.H. and Sandlie, I. Therapeutic antibodies for human diseases at the

dawn of the twenty-first century. Nature Reviews Drug Discovery . 2:52-56, 2003.

[8] 8. Little, M., Kipriyanov, S.M., Le Gall, F. and Moldenhauer, G. Of mice and men:

hybridoma and recombinant antibodies. Immunology Today. 21:364-370, 2000.

[9] 9. Chon, J.H. and Zarbis-Papastoitsis, G. Advances in the production and

downstream processing of antibodies. New Biotechnology. 28:458-463, 2011.

[10] 10. Pavlou, A.K.and Belsey, M.J. The therapeutic antibodies market to 2008.

European Journal of Pharmaceutics and Biopharmaceutics. 59:389-396, 2005.

[11] 11. ElBakri, A., Nelson, P.N. and Abu Odeh, R.O. The state of antibody therapy.

American Society for Histocompatibility and Immunogenetics. 71:1243-1250, 2010.

[12] 12. U.S. Food and Drug Administration. [Online] [Citação: 10 de Maio de 2011.]

http://www.fda.gov/.

[13] 13. European Medicines Agengy. [Online] [Citação: 10 de Maio de 2011.]

http://www.ema.europa.eu/ema/index.jsp?curl=/pages/home/Home_Page.jsp&jsen

abled=true.

60

[14] 14. Bermudez, O. and Forciniti, D. Aggregation and denaturation of antibodies: a

capillary electrophoresis, dynamic light scattering, and aqueous two-phase

partitioning study. Journal of Chromatography B. 807:17-24, 2004.

[15] 15. Alexander, A.J. and Hughes, D.E. Monitoring of IgG Antibody Thermal Stability

by Micellar Electrokinetic Capillary Chromatography and Mat rix-Assisted Laser

Desorpt ion/lonizat ion Mass Spectrometry. Analytical Chemistry. 1995, Vol. 67.

[16] 16. Vermeer, A.W.P. and Norde, W. The Thermal Stability of Immunoglobulin:

Unfolding and Aggregation of a Multi-Domain Protein. Biophysical Journal. 78:394-

404, 2000.

[17] 17. —. The influence of the binding of low molecular weight surfactants on the

thermal stability and secondary structure of IgG. Colloids and Surfaces. 161:139-150,

2000.

[18] 18. Welfle, K., Misselwitz, R., Hausdorf, G., Höhne, W. and Welfle, H.

Conformation, pH-induced conformational changes, and thermal unfolding of anti-p24

(HIV-1) monoclonal antibody CB4-1 and its Fab and Fc fragments. Biochimica et

Biophysica Acta. 1431:120-131, 1999.

[19] 19. Szenczi, A., Kardos, J., Medgyesi, G.A. and Závodszky, P. The effect of solvent

environment on the conformation and stability of human polyclonal IgG in solution.

Biologicals. 34:5-14, 2006.

[20] 20. Liu, H., Gaza-Bulseco, G. and Sun, J. Characterization of the stability of a fully

human monoclonal IgG after prolonged incubation at elevated temperature. Journal

of Chromatography B. 837:35-43, 2006.

[21] 21. Kanmert, D., Brorsson, A., Jonsson, B. and Enander, K. Thermal Induction of an

Alternatively Folded State in Human IgG-Fc. Biochemistry. 50:981-988, 2011.

[22] 22. Downstream processing with hydrophilic membranes in rotary filtration systems -

Benchmark manual by Membrex (Garfield, NJ, USA).

[23] 23. Wereley, S. T. and Lueptow, R. M. Inertial particle motion in a Taylor Couette

rotating filter. Physics of Fluids. 11:325-333, 1999.

[24] 24. Jaffrin, M.Y. Dynamic shear-enhanced membrane filtration: A review of rotating

disks, rotating membranes and vibrating systems. Journal of Membrane Science.

324:7-25, 2008.

[25] 25. Kroner, K. H., Nissinen, V. and Ziegler, H. Improved Dynamic Filtration of

Microbial Suspensions. Biotechnology. 5:921-926, 1987.

[26] 26. Wu, M.L., Freitas, S.S., Monteiro, G.A., Prazeres, D.M.F. and Santos, J.A.L.

Stabilization of naked and condensed plasmid DNA against degradation induced by

61

ultrasounds and high-shear vortices. Biotechnology Applied Biochemistry. 53:237-246,

2009.

[27] 27. Belfort, G., Pimbley, J.M., Greiner, A. and Chung, K.Y. Diagnosis of membrane

fouling using a rotating annular filter. Journal of Membrane Science. 77:1-22, 1993.

[28] 28. Maa, Y. and Hsu, C.C. Effect of High Shear on Proteins. Biotechnology and

Bioengineering. 51:458-465, 1996.

[29] 29. Cuatrecasas, P. Protein Purification by Affinity Chromatography. Journal of

Biological Chemistry. 245:3059-3065, 1970.

[30] 30. Affinity Chromatography, Principles and Methods. GE Healthcare Life Sciences.

[Online] [Citação: 20 de Março de 2011.]

http://www.gelifesciences.com/aptrix/upp00919.nsf/content/LD_149605979-F640.

[31] 31. Hjelm, H., Hjelm, K. and Sjöquist,J. Protein A from Staphylococcus Aureus. Its

isolation by affinity chromatography and its use as an immunosorbent for isolation of

immunoglobulins. FEBS Letters. 28:73-76, 1972.

[32] 32. Bailon, P., Ehrlich, G.K., Fung, W. and Berthold, W. Affinity Chromatography,

Methods and Protocols. New Jersey, USA : Humana Press, 2000.

[33] 33. Huse, K., Böhme, H. and Scholz,G.H. Purification of antibodies by affinity

chromatography. Journal of biochemical and biophysical methods. 51:217-231, 2002.

[34] 34. Gore, M.G. Spectrophotometry and Spectrofluorimetry. New York, USA : Oxford

University Press, 2000.

[35] 35. Borlido, L. Extraction of human IgG in thermo-sensitive aqueous two-phase

systems: Assessment of structural stability by circular dichroism. 2009.

[36] 36. Berova, N., Nakanishi, K.and Woody, R.W. Circular dichroism: principles and

applications. USA : Wiley-VCH, 2000.

[37] 37. Circular Dichroism Spectroscopy. [Online] [Citação: 7 de Abril de 2011.]

http://www.ruppweb.org/cd/cdtutorial.htm.

[38] 38. Kelly, S. M., Thomas, J.J. and Price, N.C. How to study proteins by circular

dichroism. 1751:119-139, 2005.

[39] 39. Circular Dichorism (CD) Spectroscopy. [Online] [Citação: 2 de Abril de 2011.]

http://mach7.bluehill.com/proteinc/cd/cdspec.html#Theory.

[40] 40. Berne, B.J. and Pecora, R. Dynamic light scattering: with applications to

chemistry, biology, and physics. New York, USA : Courier Dover Publications, 2000.

[41] 41. Alliance Protein Laboratories. [Online] [Citação: 15 de Abril de 2011.]

http://www.ap-lab.com/light_scattering.htm.

62

[42] 42. Dynamic Light Scattering: An Introduction in 30 Minutes. Malvern Innovative

solutions in material characterization. [Online] [Citação: 5 de Maio de 2011.]

http://www.malvern.com/malvern/kbase.nsf/search?readform&t=product&q=Zetasiz

er%20Nano&start=11.

[43] 43. Intensity – Volume – Number Which Size is Correct? Malvern Innovative solutions

in material characterization. [Online]

http://www.malvern.com/malvern/kbase.nsf/search?readform&t=product&q=Zetasiz

er%20Nano&start=11.

[44] 44. Borlido, L., Azevedo, A.M. and Aires-Barros, M.R. Extraction of Human IgG in

Thermo-Responsive Aqueous Two-Phase Systems: Assessment of Structural Stability

by Circular Dichroism. Separation Science and Technology. 45:2171-2179, 2010.

[45] 45. Arakawa, T. and Timasheff, S.N. Stabilization of Protein Structure by Sugars.

Biochemistry. 21:6536-6544, 1982.

[46] 46. The Engineering ToolBox. [Online] [Citação: 20 de Outubro de 2011.]

http://www.engineeringtoolbox.com/.

[47] 47. Chang, L., Sheperd, D., Sun, J., Ouellette,D., Grant, K.L., Tang, X. and Pikal,

M.J. Mechanism of Protein Stabilization by Sugars During Freeze-Drying and Storage:

Native Structure Preservation, Specific Interaction, and/or Immobilization in a Glassy

Matrix? Journal of Pharmaceutical Sciences. 94:1427-1442, 2005.

63

APPENDIX A - RELEVANT LITERATURE DATA

Table A.1 – Physicochemical properties (Relative Molecular Weigth of the molecule and the heavy chain and

Isoelectric point) of the several subclasses of Human IgG (2).

Immunoglobulin Relative Molecular Weight (Mr) (kDa) Mr (heavy chain) (kDa) pI

IgG1 146 50 5.0 – 9.5

IgG2 146 50 5.0 – 8.5

IgG3 170 60 8.2 – 9.0

IgG4 146 50 5.0 – 6.0

Table A.2 – Binding and elution conditions commonly used with Protein A for Human IgG purification (values

for each subclass) (2).

Immunoglobulin Protein A binding pH Protein A elution pH

IgG1 6.0 – 7.0 3.5 – 4.5

IgG2 6.0 – 7.0 3.5 – 4.5

IgG3 8.0 – 9.0 ≤7.0

IgG4 7.0 – 8.0 3.0 – 6.0

Table A.3 – Data used in the calculations of the average and maximum shear rate and wall shear stress at the

membrane surface. Dimensions of the rotary separation chamber: inner and outter cylinder radius (cm) an

anular gap (mm) (22) and water dynamic and cinematic viscosity at 20°C1 (46).

Rinner (cm) 2.20

e (mm) 2.3

μwater (Pa.s) at 20°C 1.002x10-3

νwater (m2/s) at 20°C 1.004x10-6

1 The properties considered for the IgG solution were the water properties.

64

APPENDIX B - RETAINED BIOLOGICAL ACTIVITY OVER TIME

B.1. IgG INCUBATED IN PBS

B.1.1. Incubation at 60°C

Figure B.1 - Retained Biological Activity (%) over time (h) of incubation at 60°C of IgG in PBS (pH 7.4) with the

initial concentrations of 1, 5, 10 and 15 g/L.


Figure B.2 - Retained Biological Activity (%) over time (h) of incubation at 70°C of IgG in PBS (pH 7.4) with the

initial concentrations of 1, 5, 10 and 15 g/L.

0

20

40

60

80

100

0 10 20 30 40 50 60

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

Time (h)

1 g/L 5 g/L 10 g/L 15 g/L

0

10

20

30

40

50

60

70

80

90

100

0 5 10 15 20 25 30 35 40

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

Time (h)

1g/L 5g/L 10g/L 15g/L

65

B.1.3. Incubation at pH 3.02

Figure B.3 - Retained Biological Activity (%) over time (h) of incubation of IgG solutions with PBS at pH 3.02


B.1.4. Incubation at pH 2.1

Figure B.4 - Retained Biological Activity (%) over time (h) of incubation of IgG solutions with PBS at pH 2 with

the initial concentrations of 1, 5, 10 and 15 g/L.

0

20

40

60

80

100

0 10 20 30 40 50 60

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

Time (h)

1 g/L 5 g/L 10 g/L 15 g/L

0

20

40

60

80

100

0 2 4 6 8 10

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

Time (h)

1 g/L 5 g/L 10 g/L 15 g/L

66

B.1.5. Mechanical Shear Stress

Figure B.5 - Retained Biological activity (%) over time (min) of induced mechanical shear stress with the inner

cylinder rotating at 1000, 2000 and 3000 rpm and 3000 rpm for the initial concentrations of IgG of 1 g/L and

15 g/L, respectively.

B.2. INCUBATION IN THE PRESENCE OF MALTOSE


Figure B.6 – Comparison of the Retained Biological activity (%) over time (h) of induced thermal denaturation

(70°C) for the initial concentrations of IgG in PBS with (continuous line) and without (dashed line) 10% (w/v)

maltose of 1, 5, 10 and 15 g/L, respectively.

0

20

40

60

80

100

0 5 10 15 20 25 30 35Re

tain

de

Bio

logi

cal A

ctiv

ity

(%)

Time (min)

2000 rpm 1000 rpm 3000 rpm 3000 rpm (15g/L)

0

20

40

60

80

100

0 2 4 6 8 10

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

Time (h)

1 g/L 5 g/L 10 g/L 15 g/L

1g/L in PBS 5g/L in PBS 10g/L in PBS 15g/L in PBS

67

B.1.7. Incubation at pH ≈ 2

Figure B.7 – Comparison of the Retained Biological activity (%) over time (h) of pH induced denaturation (pH ≈

2) for the initial concentrations of IgG in PBS with and without 10% (w/v) maltose of 1, 5, 10 and 15 g/L,

respectively.

0

20

40

60

80

100

0 2 4 6 8 10

Re

tain

ed

Bio

logi

cal A

ctiv

ity

(%)

Time (h)

1 g/L 5 g/L 10 g/L 15 g/L

1g/L in PBS 5g/L in PBS 10g/L in PBS 15g/L in PBS

68

APPENDIX C - CD SPECTRA

Figure C.1- Far-UV CD spectra of IgG in PBS with 10% (w/v) of glycine (pH ≈ 7) before incubation () and after 8

hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03 d

eg

cm2

dm

ol-

1)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03 d

eg

cm2

dm

ol-

1)

λ (nm)

69

Figure C.2- Far-UV CD spectra of IgG in PBS with 10% (w/v) of glycine before incubation () and after 9 hours

of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

-7

-5

-3

-1

1

3

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-7

-5

-3

-1

1

3

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-7

-5

-3

-1

1

3

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-7

-5

-3

-1

1

3

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

70

Figure C.3 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of D-trehalose (pH ≈ 7) before incubation () and

after 8 hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

71

Figure C.4 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of trehalose (pH ≈ 7) before incubation () and

after 9 hours of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D)

15 g/L.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

72

Figure C.5 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of sucrose (pH ≈ 7) before incubation () and after

8 hours of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

73

Figure C.6 - Far-UV CD spectra of IgG in PBS with 10% (w/v) of sucrose before incubation () and after 9 hours

of incubation () at pH 2.1 for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03

de

g cm

2 d

mo

l-1

)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03 d

eg

cm2

dm

ol-

1)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

205 225 245

[θ]

mrw

(1

03 d

eg

cm2

dm

ol-

1)

λ (nm)

74

Figure C.7 - Far-UV CD spectra of IgG in PBS with 2% (w/v) of maltose before incubation () and after 8 hours

of incubation () at 70°C for the initial concentration of (A) 1 g/L, (B) 5 g/L, (C) 10 g/L and (D) 15 g/L.

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm)

-5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm) -5

-4

-3

-2

-1

0

1

2

3

4

5

200 210 220 230 240 250

θ m

rw (

10

3 d

eg

cm2

dm

ol-

1)

λ (nm)

75

APPENDIX D - DLS SIZE DISTRIBUTIONS

D.1. EFFECT OF PH ON IgG AGGREGATION

Figure D.1. - Histogram showing the size distribution of an IgG solution in PBS (pH 2.1) before 9 hours of

incubation at pH ≈ 2, for the concentrations of row (A) 1 g/L and row (B) 15 g/L. The y-axis is the intensity of

the average scattered light (%) – Size distribution by Intensity - or total volume of particles in the different size

bins (%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters (nm).

76

Figure D.2. - Histogram showing the size distribution of an IgG solution in PBS (pH 11.3) before 9 hours of

incubation at pH ≈ 11, for the concentrations of row (A) 1 g/L and row (B) 15 g/L. The y-axis is the intensity of

the average scattered light (%) – Size distribution by Intensity - or total volume of particles in the different size

bins (%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters (nm).

77

Figure D.3. - Histogram showing the size distribution of an IgG solution in PBS with 10% (w/v) of maltose after

9 hours of incubation at pH 2.06 for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the

average scattered light (%) – Size distribution by Intensity - or total volume of particles in the different size bins

(%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters (nm).

78


9 hours of incubation at pH 11.45 for the concentrations of row (A) 1 g/L and row (B) 15 g/L. The y-axis is the

intensity of the average scattered light (%) – Size distribution by Intensity - or total volume of particles in the

different size bins (%) – Size Distribution by Volume. The x-axis is the absolute size of the scatters (nm).

79

Figure D.5. - Histogram showing the size distribution of an IgG solution in PBS after 9 hours of incubation at pH

2.1 for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the average scattered light (%) –

Size distribution by Intensity - or total volume of particles in the different size bins (%) – Size Distribution by

Volume. The x-axis is the absolute size of the scatters (nm).

80

Figure D.6. - Histogram showing the size distribution of an IgG solution in PBS after 9 hours of incubation at pH

11.3 for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the average scattered light (%) –



81

D.2. EFFECT OF TEMPERATURE ON IgG AGGREGATION

D.2.1. Initial and final (8 hours of incubation at 70°C) samples

Figure D.7. - Histogram showing the size distribution of an IgG solution (1 g/L) in row (A) PBS at room

temperature; row (B) PBS after 8 hours of incubation at 70°C; row (C) PBS with 10% (w/v) of maltose after 8

hours of incubation at 70°C. The y-axis is the intensity of the average scattered light (%) – Size distribution by

82

Intensity - or total volume of particles in the different size bins (%) – Size Distribution by Volume. The x-axis is

the absolute size of the scatters (nm).

Figure D.8. - Histogram showing the size distribution of an IgG solution (15 g/L) in row (A) PBS at room

temperature; row (B) PBS after 8 hours of incubation at 70°C; row (C) PBS with 10% (w/v) of maltose after 8

83

hours of incubation at 70°C. The y-axis is the intensity of the average scattered light (%) – Size distribution by

Intensity - or total volume of particles in the different size bins (%) – Size Distribution by Volume. The x-axis is

the absolute size of the scatters (nm).

D.2.2. 0.5 hours

Figure D.9. - Histogram showing the size distribution of an IgG solution in PBS after 0.5 hours of incubation at

70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the average scattered light (%) –



84


0.5 hours of incubation at 70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the



85

D.2.3. 1 hour

Figure D.11. - Histogram showing the size distribution of an IgG solution in PBS after 1 hour of incubation at




86


1 hour of incubation at 70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the



87

D.2.4. 5 hours

Figure D.13. - Histogram showing the size distribution of an IgG solution in PBS after 5 hours of incubation at




88


5 hours of incubation at 70°C for the concentrations of (A) 1 g/L (B) 15 g/L. The y-axis is the intensity of the



89

D.2.5. 6 hours





90





91

D.2.6. 7 hours





92





stability assessment of biopharmaceutical formulations · resumo o vasto uso de anticorpos com fins...

Documents