-
Department of Physics, Chemistry and Biology
MASTER’S THESIS
New SPR based assays for plasma protein titer determination.
Johan Kärnhall
Performed at GE Healthcare Bio-Sciences AB
Linköping, February 2011
LITH-IFM-A-EX—11-2388--SE
The Department of Physics, Chemistry and Biology
Linköping University
SE-581 83 Linköping, Sweden
-
- ii -
-
- iii -
Department of Physics, Chemistry and Biology
New SPR based assays for plasma protein titer determination.
Johan Kärnhall
Performed at GE Healthcare Bio-Sciences AB
Linköping, February 2011
Supervisors:
Åsa Frostell-Karlsson
Dr. Camilla Estmer Nilsson
Examiner:
Prof. Bo Liedberg
GE Healthcare Bio-Sciences AB
SE-750 15 Uppsala, Sweden
-
- iv -
-
- v -
Abstract Reliable analytical tools are important for time
efficient and economical process development,
production and batch release of pharmaceuticals. Therapeutics
recovered from human plasma,
called plasma protein products, involve a large pharmaceutical
industry of plasma fractionation.
In plasma fractionation of human immunoglobulin G (hIgG) and
albumin (HSA) recommended
analysis techniques are regulated by the European Pharmacopoeia
and are including total protein
concentration assays and zone electrophoresis for protein
composition and purity. These
techniques are robust, but more efficient techniques with higher
resolution, specificity and less
hands-on time are available.
Surface plasmon resonance is an optical method to study
biomolecular interactions label-free
in real time. This technology was used in this master thesis to
set up assays using Biacore systems
for quantification of HSA and hIgG from all steps of
chromatographic plasma fractionation as a
tool for process development and in-process control. The
analyses have simplified mass balance
calculations to a high extent as they imply specific detection
of the proteins compared with using
total protein detection. The assays have a low hands-on time and
are very simple to perform and
the use of one master calibration curve during a full week
decreases analysis time to a minimum.
Quick, in-process control quantification of one sample is easily
obtained within
-
- vi -
-
- vii -
Acknowledgement
I would like to thank:
My supervisors Åsa Frostell-Karlsson and Camilla Estmer Nilsson
at GE Healthcare
Bio-Sciences for their great support and help throughout the
project and for giving me
the opportunity to perform my master thesis project at GE
Healthcare.
Members of the Protein Analysis R&D, Applications division
for support and for
answering any Biacore-related questions.
Members of the BioProcessing section for their very friendly and
supporting manner
during the three weeks of guidance and evaluation of the
purification process and the
associated analyses. And for providing me process samples
throughout the project.
Klara Pettersson, my opponent for carefully reading through this
report and giving me
valuable feedback.
Bo Liedberg, for taking the time to be my examiner for this
master‟s thesis project.
-
- viii -
-
- ix -
Table of Contents
1 Introduction
...............................................................................
1
1.1. Background
.......................................................................................................
1
1.2. Aim
.....................................................................................................................
2
1.3. General approach
...............................................................................................
2
2 Theory
.......................................................................................
3
2.1. Plasma
................................................................................................................
3
2.1.1. Plasma fractionation process
.........................................................................................
3
2.1.2. Immunoglobulin G
.........................................................................................................
6
2.1.3. Albumin
............................................................................................................................
7
2.2. Protein characterization and quantification
...................................................... 8
2.2.1. Protein composition
.......................................................................................................
8
2.2.2. Molecular size distribution
.............................................................................................
8
2.2.3. Protein quantification
.....................................................................................................
8
2.2.4. International reference material
....................................................................................
9
2.2.5. Coefficient of Variation (CV)
......................................................................................
10
2.3. Surface plasmon resonance biosensor technology
.......................................... 11
2.3.1. Biacore system
...............................................................................................................
12
2.3.2. Sensor chip
.....................................................................................................................
13
2.3.3. Immobilization
..............................................................................................................
14
2.3.4. Concentration measurements
......................................................................................
15
3 Materials and Methods
............................................................ 17
3.1. Materials
..........................................................................................................
17
3.1.1. Chemicals
.......................................................................................................................
17
3.1.2. Reagents
..........................................................................................................................
19
-
- x -
3.1.3.
Materials..........................................................................................................................
20
3.2. Methods
...........................................................................................................
21
3.2.1. pH scouting
....................................................................................................................
21
3.2.2. Immobilization
..............................................................................................................
21
3.2.3. Regeneration
..................................................................................................................
22
3.2.4. Biacore concentration assay development
.................................................................
23
3.2.5. Activity and cross-reactivity experiment with capture
antibodies .......................... 26
3.2.6. Value transfer from international reference material to
calibrator ........................ 27
3.2.7. Biuret, total protein concentration assay
...................................................................
31
3.2.8. SDS-PAGE
....................................................................................................................
32
3.2.9. ELISA
.............................................................................................................................
34
4 Results
......................................................................................
37
4.1. Total IgG concentration assay
........................................................................
37
4.1.1. Evaluations of reagents for total IgG concentration assay
..................................... 37
4.1.2. Assay development total IgG concentration
.............................................................
37
4.1.3. International reference material calibration for IgG
standard ................................ 41
4.1.4. Results total IgG assay on plasma-derived process samples
.................................. 42
4.2. IgG subclass distribution assay
.......................................................................
46
4.2.1. Evaluations of reagents for IgG subclass distribution
assay .................................. 46
4.2.2. Assay development IgG subclass distribution
.......................................................... 52
4.2.3. International reference material calibration
IgGSc-standard .................................. 55
4.2.4. Results IgG subclass distribution assay on plasma-derived
samples ..................... 57
4.3. Albumin concentration assay
..........................................................................
64
4.3.1. Evaluations of reagents for albumin concentration assay
....................................... 64
4.3.2. Assay development albumin concentration
..............................................................
67
4.3.3. International reference material calibration for albumin
standard ........................ 68
4.3.4. Results albumin assay on plasma-derived process samples
.................................... 70
4.4. Albumin specificity assay
................................................................................
74
4.4.1. Evaluations of reagents for albumin specificity assay
.............................................. 74
-
- xi -
5 Discussion
................................................................................
77
5.1. Total IgG concentration assay
........................................................................
77
5.2. IgG subclass distribution assay
.......................................................................
77
5.3. Albumin concentration assay
..........................................................................
78
5.4. Biacore assays, performance and comparison
................................................ 78
5.4.1. Specificity
.......................................................................................................................
78
5.4.2. Sensitivity
........................................................................................................................
79
5.4.3. Resolution
......................................................................................................................
79
5.4.4. Robustness
.....................................................................................................................
80
5.4.5. Hands-on and analysis time
.........................................................................................
80
5.4.6. Consumables cost
.........................................................................................................
82
6 Recommendations
...................................................................
83
7 References
................................................................................
85
Appendix A Regeneration scouting
α-hIgG2........................................ 89
Appendix B Hands-on and analysis time
.............................................. 91
Appendix C Protocol total IgG concentration assay
............................ 92
Appendix D Protocol IgG subclass distribution assay
......................... 94
Appendix E Protocol albumin concentration assay
............................. 97
-
- xii -
-
- xiii -
List of abbreviations CM5 Carboxymethylated Dextran 5 CV
Coefficient of Variance EA Ethanolamine EDC
1-ethyl-3-dimethylaminopropyl-carbodiimide EDTA Ethylene
diamintetra acetic acid HBS-EP+ 10 mM Hepes pH 7.4, 150 mM NaCl,
0.5 mM EDTA, 0.5 %
surfactant P20 IFC Integrated microfluidic cartridge IgG
Immunoglobulin G IgGSc Immunoglobulin G subclass IVIG Intravenous
immunoglobulin NHS N-hydroxysuccinimide P20 Surfactant P20 (Tween
20) RI Refractive Index RM Reference Material RU Resonance Unit
SDS-PAGE Sodium dodecyl sulphate polyacrylamide gel electrophoresis
SPR Surface Plasmon Resonance TM Target Material
-
- xiv -
-
- INTRODUCTION -
- 1 -
Chapter 1
1Introduction
1.1. Background
Today, plasma protein products recovered from human plasma is a
major class of
therapeutics. A large pharmaceutical industry for fractionation
of human plasma in the world
with over 70 factories exists [1]. During the development of
fractionation processes, during the
execution of the process and for quality control (QC) there are
high demands on good and
sensitive analytical tools. Analysis of plasma protein products
is highly regulated for safety
reasons and current approved methods are presented in the
European Pharmacopoeia by the
European Directorate for the Quality of Medicine and HealthCare
[2].
Albumin has been used as a therapeutic for over 50 years and its
main usage is for colloid
replacement and maintaining of blood volume at blood loss [3].
Intravenous Immunoglobulin G
has been used for over 25 years and mainly for replacement
therapy in primary
immunodeficiency syndromes and for myeloma or chronic lymphatic
leukaemia, but new areas of
use are emerging [4].
GE Healthcare Bio-Sciences AB in Uppsala, Sweden has a
chromatographic plasma
fractionation process for the protein products coagulation
factor VIII, factor IX, human serum
albumin and Immunoglobulin G from human blood plasma. The
sensitivity, specificity, analysis-
and hands-on time of the available analysis methods were not
satisfactory for the involved parties
who required new and better methods.
GE Healthcare‟s platform Biacore, which employs surface plasmon
resonance biosensor
technology and is a highly sensitive label-free analysis tool
for biomolecular interactions, was
chosen for the study.
-
- CHAPTER 1 -
- 2 -
1.2. Aim
The first aim of this study was to perform a feasibility study
to see which of the plasma
protein products that was possible to quantify satisfactory with
a Biacore-assay, with focus on
albumin, Immunoglobulin G (IgG) and the relative distribution of
Immunoglobulin G subclasses
(IgGSc) 1-4. The second aim was to develop the most viable assay
as far as time allowed, in
addition the results and methods were to be compared with
current alternative analyses.
1.3. General approach
Several antibody reagents will be tested and conditions
optimized for the Biacore-system. The
extreme salt and pH conditions that occur from the purification
steps could possibly interfere
with the interactions required for the analysis and these
parameters needed investigation. Process
samples will be analysed with the new Biacore assay as it is
developed as well as with current
methods as a comparison. The plasma fractionation process will
be examined for insight into the
actual experimental situation.
-
- THEORY -
- 3 -
Chapter 2
2Theory
2.1. Plasma
2.1.1. Plasma fractionation process
Methods used for plasma fractionation has been developed since
the 1946 with methods
varying from traditional cold ethanol fractionation with ethanol
precipitation and centrifugation
as the major techniques to modern chromatographic processes [3].
The use of a chromatographic
process enables a larger variety of products to be extracted
from the plasma other than traditional
albumin processes, it is also less damaging and generally gives
a higher yield.
There are two types of human plasma differentiated by the means
of collection. The major
type is plasma collected with plasmapheresis or apherisis where
blood is filtered or continuously
centrifuged and the blood cells returned to the donor. The
second type is plasma recovered
through double centrifugation of whole blood donations. Plasma
from plasmapheresis
corresponds to 65 % and recovered plasma to 35 % of the total
plasma fractionated in the world
today [1]. Both the plasmapheresis donations (category A plasma)
and whole blood donations
(category B plasma) are to be frozen within 6 hours, if frozen
within 24 hours of donation
(category C plasma) it can only be used in the production of
immunoglobulin G and albumin [5].
The current process of interest is a chromatographic method
using several steps of buffer
exchange chromatography, gel filtration chromatography, anion-
and cation exchange
chromatography together with ultra- and diafiltration and
numerous other steps. Ultrafiltration is
used to increase the concentration while diafiltration also
replaces the buffer. An overview of the
process is displayed in Figure 2-1. The process is structured
with factor VIII being the first
product to be separated, thereafter factor IX followed by
albumin and finally IgG. This leads to
-
- CHAPTER 2 -
- 4 -
four segments that can be called: factor VIII-trail, factor
IX-trail, albumin-trail and finally IgG-
trail. The crude plasma has to be treated with heparin, which is
a highly-sulphated
glycosaminoglycan acting as an anti-coagulant. All the products
have to undergo virus
inactivation and sterile filtration in order to be safe to use
as a pharmaceutical [1, 5]. Virus
inactivation is typically done by addition of solvent and
detergent chemicals, such as tween-80,
TNBP, or triton X-100, or by pasteurisation and finally sterile
filtration.
The chromatographic purification requires a variety of different
buffers with different pH and
salt levels to elute the wanted proteins. Sodium Chloride (NaCl)
levels vary between 0 and 500
mM and pH levels vary from pH 4.0 to pH 9.0. Together, this can
yield quite extreme conditions
complicating the quantification methods.
-
- THEORY -
- 5 -
Figure 2-1: Process overview plasma fractionation by GE
Healthcare The four sections are denoted Factor VIII-trail, Factor
IX-trail, Albumin-trail and IgG-trail. Blue boxes represent
chromatography steps and yellow boxes represent filtration steps.
In this study only the Albumin- and IgG-trail were studied, each
time starting from plasma following the black arrows. Samples were
taken and analysed from the entire process, at least before and
after every major chromatography and filtration step. For example
the DEAE Sepharose FF step in the Albumin-trail was denoted “Alb
DEAE” and the second ultrafiltration in the IgG-trail was denoted
“IgG UF2”.
Plasma
Pre-treatment
Sepharose 4 FF
Q Sepharose HP
Chemical addition
Virus inactivation
SP Sepharose HP
Superose 12 pg
Formulation
Ultrafiltration
Sterile filtration
Filling
Lyophilisation
Severe heat treatment
DEAE Sepharose FF
Chemical addition
Virus inactivation
Heparin Sepharose FF
Q Sepharose FF
Ultra-diafiltration
Sterile filtration
Filling
Lyophilisation
Severe heat treatment
Ultrafiltration
Sephadex G-25 C
Euglobulin precipitation
Centrifugation
DEAE Sepharose FF
CM Sepharose FF
Ultrafiltration
Heat treatment
Centrifugation
Sephacryl S-200 HR
Ultra-diafiltration
Formulation
Sterile filtration
Ultrafiltration
Q Sepharose FF
Ultrafiltration
Chemical addition
Virus inactivation
CM Sepharose FF
Ultra-diafiltration
Formulation
Sterile filtration
Filling
Pasteurization
Filling
Factor VIII Albumin
Factor IX
IgG
-
- CHAPTER 2 -
- 6 -
2.1.2. Immunoglobulin G
Immunoglobulins, also known as antibodies, are protein molecules
part of the immune system
used to specifically identify and bind antigens leading to an
immune response. Antibodies usually
bind the antigens tightly, sometimes not even leaving space for
water molecules, by interactions
primarily formed by hydrogen bonds and electrostatic
interactions. In the bloodstream the most
common class of immunoglobulins are immunoglobulin G class
(IgG), which will hereby be
described more thoroughly. In a normal pool of plasma, the total
IgG level is on average 8.5
mg/ml [6]. IgG is a globular, water-soluble protein with a
molecular weight of approximately
150 000 Dalton (150 kDa). IgG is composed of two light chains
consisting of two domains each
and two heavy chains consisting of four domains each, linked
together with disulphide bonds, see
Figure 2-2 for a structural overview [7]. All domains possess
the characteristic immunoglobulin
fold consisting of two sandwiched antiparallel β-sheets [8].
Immunoglobulins are glycoproteins containing of 82-96 % protein
and 4-18 % carbohydrate
attached to the heavy chains [8]. Each IgG has two antigen
binding sites located at the N-termini
of the light and heavy chains in the variable domains (Figure
2-2) [7]. The region on an antigen
recognized by the antibody is called the epitope; there can be
several epitopes on one antigen
recognized by different antibodies.
Figure 2-2: Structural overview of Immunoglobulin G An
illustration of Immunoglobulin G showing the heavy (red) and light
(blue) chain and also the Fc and Fab regions [7].
The pharmaceutical product intravenous immunoglobulin (IVIG or
IGIV in the US) has
many clinical uses but with potential risks and an inevitable
limited supply due to its human
origin. The United States Food and Drug Administration (FDA)
currently have six clinical
indications licensed for IVIG, they are: primary
immunodeficiency disease, idiopathic
thrombocytopenic purpura, Kawasaki disease, B-cell chronic
lymphocyticleukemia, HIV
infection, bone marrow transplantation [4]. In recent studies,
it has also been found to work for
autoimmune diseases [9] and Alzheimer‟s Disease [10].
Antigen binding site
Fab
Fab
Fc
hinge region
-
- THEORY -
- 7 -
2.1.2.1. IgG subclasses
There are four different isotypes, or subclasses, of IgG named
IgG1, IgG2, IgG3 and IgG4.
The distribution of these subtypes in the blood varies with
individuals, depending mainly on age
and sex. The average distribution is as followed: IgG1 (58,9 %)
> IgG2 (21,1 %) > IgG3 (4,3 %)
≈ IgG4 (4,7 %) [7]. These different IgG subclasses, illustrated
in Figure 2-3, show differences in
structure where IgG3 is larger (170 kDa) than the others (146
kDa) with the main difference in
the hinge region, with 62 amino acids in IgG3 rather than 12 in
the others. IgG3 is also more
susceptible to proteolytic enzymes and has a shorter biological
half-time, 7 days compared to 21
days [7].
Figure 2-3: Immunoglobulin G subclasses Illustrations of the
four IgG subclasses. The major visible differences are the
hinge-region which is uniquely elongated in IgG3 and shorter in
IgG4 [7].
2.1.3. Albumin
Albumin is the most abundant protein in the plasma and
corresponds to approximately 60 %
of the total protein by mass. On average, in a normal pool of
plasma, the albumin level is 34
mg/ml [6]. It is a very stable, highly water-soluble protein
with a molecular weight of 66 500
Dalton (66.5 kDa) [11]. Albumin maintains the colloid osmotic
pressure which ensures retaining
of water in the circulation. The protein is also a carrier for
several hormones, enzymes, fatty-
acids, metal ions and medical products [3]. In the blood,
albumin is generally composed with 0.5 -
1.5 moles fatty-acids per mole albumin [11]. The most frequent
fatty-acids are: Oleic < 33 %,
Palmitic 25 %, and Linoleic < 20 % [11]. During purification,
some of the fatty-acid composition
will be depleted and by special steps it can be completely
removed yielding a fatty-acid free
albumin preparation [11].
-
- CHAPTER 2 -
- 8 -
2.2. Protein characterization and quantification
2.2.1. Protein composition
The protein composition in a plasma sample is generally
determined by sodium dodecyl
sulphate polyacrylamide gel electrophoresis (SDS-PAGE). Proteins
are separated on a gel by
electrophoresis, solely based on their molecular weight. By
comparing the molecular mobility of
the samples with molecular markers, the protein composition and
the purity may be concluded.
Other techniques for determination of protein composition are 2D
gel electrophoresis (2DGE)
and capillary zone electrophoresis.
2.2.2. Molecular size distribution
Size exclusion chromatography, also called gel filtration
chromatography, is used to determine
the molecular size distribution of the purified protein
products. According to the European
Pharmacopoeia, for albumin at least 95 % of the total peak area
has to be composed of monomer
or dimer and polymers and aggregates may not represent more than
5 % of the total peak area
[2]. For IgG the peaks of polymers and aggregates should not be
more than 10 % of the total
peak area [2].
2.2.3. Protein quantification
Quantification of proteins is generally carried out with an
assay based on analysis of a
calibrator of known concentration in several dilutions. In
Biacore, there exists an alternative to
using a calibrator called Calibration Free Concentration
Analysis (CFCA), more on this in section
2.3.4. The measured signal is used to construct a standard curve
where standard points are fitted
with either a linear or non-linear mathematical fitting model.
Samples with unknown
concentration with different dilutions is analysed and
interpolated on the standard curve to give
the concentration. Preferably, a control sample with known
concentration is also analysed and
the concentration interpolated on the standard curve is compared
with the true concentration
[12].
Modern surface plasmon resonance based biosensor systems as well
as nephelometric or
turbidimetric optical systems and ELISA use an antibody to
recognize the targeted antigen in the
sample and these assays are called immunoassays or
immunochemical assays. Other techniques
than immunoassays such as biuret-assay, Kjeldahl nitrogen-assay
and absorbance spectroscopy
are less sensitive and not specific to a certain protein.
Quantification assays have a high demand on instrument and
antibody reagents as well on
calibrators and controls. Immunoassays for human plasma protein
measurements are highly
influenced by several factors that are not always met [13]. The
nature of the antibody and antigen
is vital, with demand on highly specific antibodies and a
homogenous invariable antigen. This is
-
- THEORY -
- 9 -
not always the case when analysing samples throughout a
purification process as the antigen may
change as it becomes purer, for example the removal of
fatty-acids bound to albumin (mentioned
in section 2.1.3) which may impact the interaction. Further,
changes in salt levels as well as pH
might interfere with the antibody recognition in the
immunoassay. Finally, the calibrator used has
to behave identically with the measured analyte in order to
yield a comparable signal.
The leading techniques for protein quantification in clinical
chemistry today are nephelometry
and turbidimetry [12]. The two methods are both immunochemical
fluid phase optical sensors,
where nephelometry measures an increase in side-scattered light
while turbidimetry measures a
decrease in light transmission. Calibrators or samples are
injected into a reaction tube. Antibodies
against for example human IgG1 are added and antibody-antigen
complexes are formed. After a
fixed time, the side-scattered light is recorded. Standard
curves are constructed and sample
measures are interpolated and concentrations calculated [7].
There are several assays available for quantitative
determination of IgG subclasses. The most
common ones are radial immunodiffusion (RID), nephelometry,
turbidimetry and ELISA [7].
RID is performed in ready-to-use agar plates integrated with
specific antibodies against the
IgGSc. Standards, controls and samples are added in holes in the
agar. As the IgGSc migrates
into the agar and forms complexes with the integrated antibodies
precipitation rings will emerge.
The diameter is proportional to the level of that specific IgG
subclass. The method requires 48-
60 hours incubation time with a moderate hands-on time and no
automation [7].
Nephelometry and turbidimetry are discussed above. The detection
limit is in μg/ml range
with a fairly short analysis time and an automated system
[7].
Enzyme-linked immunosorbent assay (ELISA) which was the method
chosen to compare
with in this study is a well-known and widely used
immunochemical method. The IgG subclasses
are captured by a coated anti-human IgG subclass-specific
antibody. A secondary enzyme-linked
antibody is added and quantified by a coloured enzyme reaction
upon addition of a substrate.
The ELISA method has a very low detection limit but demands a
high hands-on time and a long
analysis time [7].
2.2.4. International reference material
In order to ensure the use of good and correct standards for
quantification and to reduce the
observed variation of up to 50 – 100 % depending on the
calibrator used, international reference
material has been introduced [13]. A variety of international
reference materials has been used for
decades and has previously been produced by amongst others the
World Health Organisation
(WHO), Community Bureau of References of the Commission of the
European Communities
(BCR) and today by the Committee on Plasma Protein
Standardisation of the International
Federation of Clinical Chemistry (IFCC) [13].
The latest recognised international reference material for
plasma proteins is called ERM®-
DA470k/IFCC and is valid for twelve common plasma proteins:
α2-macroglobulin, αl-acid
-
- CHAPTER 2 -
- 10 -
glycoprotein (orosomucoid), αl-antitrypsin (αl-protease
inhibitor), albumin, complement 3c,
complement 4, haptoglobin, immunoglobulin A, immunoglobulin G,
immunoglobulin M,
transferrin and transthyretin (prealbumin) [14]. This type of
reference material is called a certified
reference material (CRM) and is provided with a certificate of
analysis with certified and traceable
values, accompanied with a value of uncertainty. CRMs are
generally short on stock and are not
to be used on a daily basis [15].
Development and use of analytical tools requires large
quantities of reference material and
with CRMs this would become quite costly. Instead it is
recommended and practical to use other
reference materials or standards that are purchased or produced
in-house to act as the calibrator
[15]. This calibrator is to be calibrated against the CRM using
determined procedure and protocol
to transfer the value from the reference material to the target
material [16-17].
2.2.5. Coefficient of Variation (CV)
The coefficient of variation (CV) is a normalized measure of
reliability expressed in
percentage. It has the advantages to be a dimensionless number
enabling the user to compare the
CV between different data sets without taking into consideration
the mean value. When the mean
value is closer to zero the CV is very sensitive to small
changes and are therefore not as useful.
CV is normally presented in percentage and with the number of
data in the set as n. CV is
calculated with Equation 2-1 below.
Equation 2-1
Where σ = standard deviation and µ = mean.
100%
CV
-
- THEORY -
- 11 -
2.3. Surface plasmon resonance biosensor technology
Surface plasmon resonance (SPR) biosensor technology is a
powerful tool in label-free
biomolecular interaction analysis used in drug discovery and
proteomic research. Today, several
biosensor systems employing SPR technique exists on a growing
market, the leader in SPR
biosensors is Biacore from GE Healthcare [18]. Also, other
technologies for label-free
biomolecular interaction analysis are available, such as
bio-layer interferometry (BLI) used in
ForteBio‟s instruments and quartz crystal microbalance used in
Attana‟s and Q-sense‟s
instruments [19].
The application of surface plasmon resonance biosensors on
biomolecules was first
demonstrated in 1983 [20]. When a beam of plane-polarized light
passes through a prism with a
thin metal film it is totally internally reflected if the angle
is above a certain critical angle of
incidence [21]. The reflected light is monitored and the
intensity measured.
As the angle of incidence is changed the reflected light will
decrease in intensity at a specific
angle showing a dip in reflected light. At this specific angle,
surface plasmons in the metal film
are excited by the light inducing surface plasmon resonance
(SPR) [21]. When the wave vector of
the incident light matches the wavelength of the surface
plasmons, the free electrons in the metal
film resonate, hence the term surface plasmon resonance. The
angle with the maximum loss of
intensity is called the SPR angle or resonance angle. This angle
is dependent on the optical
properties of the media adjacent to the metal film.
Figure 2-4: Principle of SPR and schematic sensorgram Left: The
principle of a SPR biosensor. Right: A schematic sensorgram showing
the response upon association of analyte during injection and the
dissociation post injection followed by regeneration.
1 0
Regeneration Analyte injection phase
Post-injection
phase
-
- CHAPTER 2 -
- 12 -
On one side is the glass prism with an unaltered high refractive
index (RI) and on the other
side the measured medium with a low RI [21]. Molecules such as
proteins are bound and thereby
increasing the mass, the refractive index is changed leading to
a shift in the SPR angle [21]. Figure
2-4 illustrates the principles of SPR described here and also
shows a schematic sensorgram. The
shift in angle is translated to resonance units (RU), 1 RU is
equivalent to 10-4° corresponding to
about 1 pg/mm2 bound protein and is linear all the way to the
upper limit of the dynamic range
[22].
The surface plasmon creates an electromagnetic field, called the
evanescent field, which
propagates into the media [22]. Any change in mass, and thereby
a change in RI, occurring
within this evanescent field is detected by the sensor [22]. The
molecule immobilized on the
surface is called ligand and the molecule injected sample is
called the analyte.
Since the response is dependant of the refractive index of the
solution in the flow channel,
when a solution with a different RI is injected a bulk response
will be visible. When the injection
is completed, the bulk response will disappear. This can for
example be visible when using
different buffers and variable concentrations of salts, such as
NaCl in the solutions. Extreme
levels of NaCl (very high or low) might also affect interactions
in other ways, as many
interactions are governed by electrostatic attractions.
2.3.1. Biacore system
The Biacore system from GE Healthcare can monitor a biomolecular
interaction in real-time
and label-free. The system consists of three main units, the SPR
optics, the liquid handling
system and the sensor chip [21]. The sensor chip will be
discussed in section 2.3.2 and the SPR
optics and the principle of SPR technology was brought up in
section 2.3. The instrument used in
this study was Biacore T100 system and in some cases T200, see
Figure 2-5. These instruments
are very similar but with a higher sensitivity in the T200. In
this system the liquid handling system
consists of an IFC with four flow-cells, sample injection loops,
highly accurate pumps and
pneumatic valves [23]. For different applications the flow-cells
can be used independently or
serially as in Figure 2-6.
Figure 2-5: Biacore T100 instrument A Biacore T100 instrument
that was used during this study.
-
- THEORY -
- 13 -
Figure 2-6: flow-cells. Illustration of the flow-cells in a
Biacore T100 system [24]. The flow-cells are formed when the sensor
chip is docked on the IFC (top). The four flow-cells can be used
either serially (left) or independently (right).
2.3.2. Sensor chip
The sensor chip (see Figure 2-7) consists of a plastic cassette
designed so the sensor chip is
easily moved and positioned automatically onto the integrated
microfluidic cartridge (IFC) in the
instrument. The chip itself is composed of a thin glass covered
with a 50nm gold film, coated
with a monolayer of hydroxyalkanethiol linkers [23]. There are
several different sensor chips
available with different surface chemistry attached to the
linkers; they are suitable for different
interactions, applications and immobilization techniques. The
most common sensor chip, and the
chip used in this study, is the CM5; which has a
carboxymethylated dextran matrix attached [22].
The dextran matrix is used as an anchor for the immobilization
of ligands (see section 2.3.3).
There are more advantages by using a dextran matrix; firstly it
enables the ligands to be
positioned in a three dimensional space increasing the number of
interactions sensed by the
evanescent field and thereby increasing the binding response,
secondly it enables the interaction
to proceed under conditions that mimics a fluidic and thirdly it
minimises non-specific binding to
the gold surface [24].
-
- CHAPTER 2 -
- 14 -
Figure 2-7: Series S sensor chip CM5 The carboxymethylated
dextran matrix spans 100 nm into the flow-cell and is attached to
the gold surface with a layer of linkers (grey).
2.3.3. Immobilization
There are several available coupling chemistries to immobilize
proteins to the sensor chip
surface. To the carboxymethylated dextran matrix on CM5 sensor
chip it is possible to do several
different coupling chemistries [25]. Amine coupling, the most
used technique and the one used
in this study, will be further described below [26]. Carboxyl
groups on the matrix forms covalent
bonds with primary amines on the ligand protein. This reaction
does not occur spontaneously
and an activation of carboxyl groups into esters is necessary.
This is done with a mixture of 1-
ethyl-3-dimethylaminopropyl-carbodiimide (EDC) and
N-hydroxysuccinimide (NHS) as
illustrated in Figure 2-8.
First, EDC reacts with the carboxyl group forming a reactive
intermediate. Second, the NHS
reacts and forms a NHS ester which is a good leaving group.
Finally, as the protein is injected the
activated ester will spontaneously react with primary amines on
the protein forming a covalent
bond. The final step is to inject an ethanolamine (EA) solution
that reacts with the remaining of
the activated esters.
Figure 2-8: EDC NHS chemistry The chemical reaction during
EDC/NHS immobilization. EDC reacts with the carboxyl group on the
dextran matrix. This forms a reactive intermediate which reacts
with NHS, leaving an ester. The NHS ester is a good leaving group,
reacting with a primary amine on the ligand, forming a covalent
bond.
Carboxyl group / Ligand
Dextran
Linker layer
Gold film
Glass
-
- THEORY -
- 15 -
During the immobilization, in order to attract the ligand to the
surface to ensure the reaction
to occur to a satisfactory extent it has to be attracted by
electrostatic forces in something called
pre-concentration. By dissolving the ligand in a buffer with a
pH below the pI of the protein, this
will result in the protein having a net positive charge and it
will be attracted to the slightly
negatively charged dextran matrix.
The optimum pH can be determined by a pH scouting experiment
described in section 3.2.1.
The properties that can be modified in order to vary the level
of final immobilized ligand to the
desired level are concentration, pH of buffer, flow-rate and
contact time.
2.3.4. Concentration measurements
The SPR technology can be used for concentration measurements in
a robust, accurate,
precise and specific manner [27]. In several recent studies, SPR
biosensors have been used for
quantification, for example quantification of bovine IgG in milk
(2010) [28], estriol metabolites in
liquid media (2009) [29] and yessotoxin from marine
dinoflagellates (2008) [30]. In general, three
methods for quantification exist. A traditional method using the
relative response after injection
[23], a method using the binding rate (RU/s) [31] and the most
recent calibration free
concentration analysis (CFCA) using two different flow-rates
[32].
-
- MATERIALS AND METHODS -
- 17 -
Chapter 3
3Materials and Methods
3.1. Materials
3.1.1. Chemicals
Chemical Cat. No. Supplier
Milli-Q filtered H2O Millipore
HBS-EP+ 10X BR-1006-69 GE Healthcare
MgCl2 M2670 Sigma-Aldrich
50 mM NaOH BR-1003-58 GE Healthcare
3M MgCl2 BR-1008-39 (capture kit) GE Healthcare
Glycine pH 2.0 BR-1003-55 GE Healthcare
Amine coupling
Ethanol amine BR-1000-50 GE Healthcare
EDC BR-1000-50 GE Healthcare
NHS BR-1000-50 GE Healthcare
Immobilization pH scouting
Acetate pH 4.0 BR-1003-49 GE Healthcare
Acetate pH 4.5 BR-1003-50 GE Healthcare
Acetate pH 5.0 BR-1003-51 GE Healthcare
Acetate pH 5.5 BR-1003-52 GE Healthcare
-
- CHAPTER 3 -
- 18 -
Regeneration Scouting
Glycine pH 1.5 Regeneration scouting kit GE Healthcare
Glycine pH 2.0 BR-1005-56 GE Healthcare
Glycine pH 2.5 BR-1005-56 GE Healthcare
Glycine pH 3.0 BR-1005-56 GE Healthcare
SDS 0.5% BR-1005-56 GE Healthcare
NaCl 5M BR-1005-56 GE Healthcare
MgCl2 4M BR-1005-56 GE Healthcare
NaOH 200mM BR-1005-56 GE Healthcare
Ethylene Glycol BR-1005-56 GE Healthcare
SDS-PAGE NuPAGE Novex 4-12% Bis-Tris Gel, 1.9 mm, 12 well
NP0322BOX Invitrogen
Precision Plus Protein Dual Color Standard 161-0374 Bio-Rad
Laboratories AB
NuPAGE LDS sample buffer 4X NP0007 Invitrogen
β-Mercapthoethanol M6250 Sigma-Aldrich
NuPAGE MOPS SDS Running buffer 20X NP0001 Invitrogen
GelCode Blue Stain Reagent #24592 Thermo Scientific
ELISA
Peliclass human IgG subclass kit M1551 Sanquin
Chromatography media
Ion exchanger media ( GE Healthcare)
Matrix: Highly cross-linked agarose, 6%
Particle size: average 90 μm (45 – 165 μm)
Q Sepharose™ FF Quaternary ammonium strong anion exchanger
Cat. No. 17-0510-05
DEAE Sepharose™ FF Diethylaminoethyl weak anion exchanger
Cat. No. 17-0709-05
CM Sepharose™ FF Carboxymethyl weak cation exchanger
Cat. No. 17-0719-05
Gel filtration media (GE Healthcare)
Sepharose 4 Fast Flow Highly cross-linked 4% agarose
Cat. No. 17-0149-01 Particle size: 45 – 165 µm
Fractionation range: 6 × 104 – 3 × 107 Da
Sephacryl S-200 HR Spherical allyl dextran and N,
N’-methylenebisacrylamide
Cat. No. 17-0584-10 Particle size: 50 µm
Fractionation range: 5 × 103 – 2.5 × 105 Da
Sephadex G-25 Cross-linked dextran
Cat. No. 17-0034-01 Particle size: 75 – 510 µm
Fractionation range: 1 × 103 – 5 × 103 Da
-
- MATERIALS AND METHODS -
- 19 -
3.1.2. Reagents
Name Denotation Supplier / Cat. No.
human IgG hIgG Sigma / I4506
human IgG1 κ (myeloma) hIgG1 Millipore / AG502
human IgG2 κ (myeloma) hIgG2 Millipore / AG504
human IgG3 κ (myeloma) hIgG3 Millipore / AG506
human IgG4 κ (myeloma) hIgG4 Millipore / AG508
Peliclass human IgG subclass standard IgGSc-standard Sanquin /
M1551
Peliclass human IgG subclass control IgGSc-control Sanquin /
M1551
HSA "Essentially fatty acid free" HSAa Sigma / A-3782
HSA "Fraction V" HSAb Sigma / A-1653
HSA “internally purified” HSAc GE Healthcare / internal
BSA BSA Sigma / P9418
HSA and gamma-globulins Sigma / P8119
International RM - ERM-DA470k/IFCC ERM-DA470k Sigma /
ERMDA470KIFCC-1VL Antibodies
Species +
Specificity Denotation Clone Isotype Supplier / Cat. No.
human IgG (Fc) α-hIgG GE Healthcare / BR-1008-39
human IgG1 α-hIgG1poly sheep (poly) The binding site / AU006
human IgG1 (Fc) α-hIgG1a HP6091 mouse IgG2a The binding site /
MC003
human IgG1 (Fc) α-hIgG1b HP6069 mouse IgG1 Invitrogen /
MH1013
human IgG1 (Fc) α-hIgG1 HP6070 mouse IgG1 Invitrogen /
MH1015
human IgG2 (Fab) α-hIgG2 HP6014 mouse IgG1 The binding site /
MC005
human IgG3 (Fab2) α-hIgG3 HP6050 mouse IgG1 The binding site /
MC006
human IgG4 (pFc) α-hIgG4 HP6025 mouse IgG1 The binding site /
MC009
mouse IgG (Fc) α-mIgG Rabbit (poly) GE Healthcare /
BR-100838
HSA α-HSApoly Rabbit (poly) GE Healthcare / internal
HSA α-HSAmab mouse IgG1 Abcam / Ab399
BSA α-BSAa 2A3E6 mouse IgG1 Santa Cruz Biotech / sc-32816
BSA α-BSAb 0.N.32 mouse IgG1 Santa Cruz Biotech / sc-70445
BSA α-BSAc BGN/D1 mouse IgG1 Santa Cruz Biotech / sc-80704
-
- CHAPTER 3 -
- 20 -
3.1.3. Materials
Material Cat. No. Supplier
Microplate 96 well
Microplate cover-foil 96 well
Microplate flat bottom 96 well
Pipette and pipette tips, 10-100ul Eppendorf
Pipette and pipette tips, 20-200ul Eppendorf
Pipette and pipette tips, 100-1000ul Eppendorf
Finnpipette, 5mL Labsystems
Finntips, 5mL 940 20 50 Thermo Scientific
Pipette Multi channel, 30-300ul Eppendorf
Pipette Multi channel automatic, 10-200ul Eppendorf
Pipette Multi channel automatic, 100-1000ul Eppendorf
Series S Sensor chip CM5 BR-1006-68 GE Healthcare
Plastic vials, ø 7mm BR-1002-12 GE Healthcare
Glass vials, ø 16mm BR-1002-09 GE Healthcare
Rubber cap, type 3 (for ø 7mm) BR-1005-02 GE Healthcare
Rubber cap, type 2 (for ø 16mm) BR-1004-11 GE Healthcare
Instrument Software Supplier
Biacore T100 Control software v2.0.3, Evaluation software v2.0.3
GE Healthcare
Biacore T200 Control software v1.0, Evaluation software v1.0 GE
Healthcare
Milli-Q Advantage A10 Millipore
Electrophoresis power supply – EPS 301 GE Healthcare
miniVE – Vertical electrophoresis system GE Healthcare
ImageScanner III Labscan 6.0 GE Healthcare
ImmageQuant TL 6.0 GE Healthcare
SPECTRA Max PLUS 384 SoftMax Pro v5.4 Molecular Devices
Microplate-shaker
ÄKTA pilot Unicorn v5.11 GE Healthcare
-
- MATERIALS AND METHODS -
- 21 -
3.2. Methods
If nothing else is stated, all Biacore-experiments were
performed at 25°C with HBS-EP+ as
sample and running buffer. For longer (>12 hours) experiments
the sample compartment
temperature was decreased to 10°C from 25°C, but the analysis
temperature remained unaltered.
3.2.1. pH scouting
In order to determine the optimal pH for pre-concentrating the
ligand to the matrix during
immobilization, as described in section 2.3.3, a pH scouting was
performed. The ligand was
diluted to 20 μg/ml in buffers with different pH and injected
during 2.5 minutes over an
unmodified sensor chip. After each injection the surface was
regenerated with 50 mM NaOH to
ensure no ligand remains non-specifically bound to the surface.
The most neutral pH was
injected first followed by more acidic injections. The aim was
to obtain a sufficiently high
increase of response but with the most neutral pH possible in
order to maintain the native state
of the ligand. The buffers used were 10 mM sodium acetate with
pH ranging from 4.0 to 5.5, 10
mM maleate pH 6.0 to 6.5 and 10 mM phosphate pH 7.0. An example
of a pH scouting can be
seen in Figure 4-8.
3.2.2. Immobilization
Immobilization of ligands to the sensor chip surface was
performed with amine coupling
chemistry, as described in section 2.3.3. Chemicals from amine
coupling kit (GE Healthcare)
were utilised. The surface was activated with a 7 minute
injection of 1:1 mixture of EDC and
NHS. The ligand injection was optimized for each antibody and
specified under each result
section; typically a 7 minute injection of 20 μg/ml antibody
diluted in pre-concentration buffer
was used. The surface was deactivated with a 7 minute injection
of ethanol amine (EA). An
example of an immobilization sensorgram is displayed in Figure
3-1.
Figure 3-1: Typical immobilization sensorgram
0
10000
20000
30000
0 400 800 1200 1600
Re
spo
nse
(RU
) .
Time (s)
EA Ligand ~10000 RU EDC/NHS
-
- CHAPTER 3 -
- 22 -
3.2.3. Regeneration
For experiments when the affinity of the interaction is high,
and the analyte does not
dissociate by itself it is required to regenerate the surface
between cycles. This is generally the
case for concentration analysis with high affinity antibodies
and high responses. The principle of
regeneration is that the interactions between the analyte and
the ligand are broken at the same
time as the analyte may be partly denatured whilst the ligand
maintains its activity. Therefore, for
an easier regeneration the less stable protein should be the
analyte.
Different results that might occur during regeneration are
illustrated in
Figure 3-2. A and B show optimal and acceptable regeneration
when the analyte response and
the baseline remains the same. C and D illustrate incomplete
regeneration. The last two show
irreversible changes of the ligand due to the regeneration, E
has a loss of ligand activity and in F
the ligand is lost from the surface.
Figure 3-2: Illustration of regeneration results Common
regeneration results are illustrated. A and B show good
regeneration. C and D illustrate incomplete regeneration. E and F
indicate an irreversible change on the ligand due to
regeneration.
3.2.3.1. Regeneration scouting
The protocol from the regeneration scouting kit was followed. A
freshly immobilized and
previously unused surface was used for each regeneration
solution tested. An analyte with a high
concentration was injected and the binding response and baseline
was compared to the initial
cycle. The mildest condition for each solution was used first
with a successively tougher
Analyte response
Baseline
A Optimal regeneration.
C Incomplete regeneration. Accumulation of analyte and loss of
capacity. .
E Loss of ligand activity. Irreversible change.
B Acceptable regeneration.
D Incomplete regeneration. Accumulation of analyte.
F Loss of ligand. Irreversible change.
-
- MATERIALS AND METHODS -
- 23 -
condition following. For each condition the analyte was injected
and regenerated four to five
times.
The conditions are met if the response is recovered to
preferably 70 % from the first cycle and
the baseline is similar to the first cycle, a small constant
decrease in baseline may be acceptable as
long as the analyte response is repeatable. The condition that
gives the best regeneration is
verified by 20 or more cycles with the same condition. Further,
the injection time of regeneration
solution might be increased or decreased in order to give a
better regeneration.
The tested regeneration solutions were:
10 mM Glycine-HCl, pH 3.0 to 1.5
Ethylene glycol, 50% to 100%
Sodium hydroxide (NaOH), 1 mM to 75 mM
Magnesium chloride (MgCl2), 1 M to 4 M
Sodium chloride (NaCl), 0.5 M to 5 M
Sodium dodecyl sulphate (SDS), 0.02 % to 0.5 %
3.2.4. Biacore concentration assay development
There were three different methods of concentration
determination in Biacore to choose
from. First, the traditional method where the relative response
of the calibrator was plotted
against the concentration [23]. Second, a method where the
binding rate (RU/s) of the calibrator
was plotted against the concentration [31]. Third, a calibration
free concentration analysis
(CFCA) where a calibrator was not needed by using different
flow-rates [32-33]. With the plasma
and process samples that were analysed, the traditional relative
response method was chosen due
to large bulk responses and some non-specific binding during
injection of non-purified samples
interfering with the other methods.
The Biacore concentration assays that were developed in this
study had a number of
parameters that were optimized and thus leading to the assays
presented in section 4.1.2 for total
IgG, section 4.2.2 for IgG subclass distribution and section
4.3.2 for albumin. These parameters
and the criterions to determine them will be discussed here.
Biacore concentration assay parameters:
Ligand antibody
o Choice of antibody
The desired characteristics for an antibody to be used in a
concentration assay
was that it binds the analyte specific and with a high affinity
when immobilized
on the sensor chip. Antibodies were also necessary to be able to
regenerate
under known conditions without losing activity. Preferably
commercially
available monoclonal antibodies were chosen.
-
- CHAPTER 3 -
- 24 -
o Immobilization level: Buffer, injection time, flow-rate,
concentration
The aimed immobilization level for concentration assays is
generally around
10000 RU. A high immobilization level is necessary in order to
have mass-
transport limited interaction as discussed below. To reach a
certain level the
pH of the pre-concentration buffer had to be determined by a pH
scouting
(section 3.2.1). Also the injection time was evaluated to obtain
desired level.
Finally the concentration of the antibody diluted into the
pre-concentration
buffer was studied to determine a suitable concentration.
Several antibodies
were delivered in sodium-azide preservative and Tris-buffer and
had to be
diluted enough to avoid interfering with the immobilization. As
these
compounds contain a primary amine they would otherwise be
immobilized.
The flow-rate decreased in order to reduce consumption of
reagents.
Concentration assay
o Buffer
HBS-EP+ has in several previous studies been shown as an
appropriate buffer
for real-time interaction studies and was found to work well
also for these
assays.
o Choice of reagent
The reagent used as standard needs to interact with the antibody
in an identical
manner as the sample. The reagent should preferably be
commercially
available. Users of the assays can utilise their own standards
as long as it is
calibrated against the international reference material.
o Concentration range, injection time and flow-rate
Injection time and concentrations were varied to obtain an assay
where the
lowest point in the standard curve gave high enough response
while
maintaining sufficient sensitivity. At the same time the assays
were designed to
be as rapid as possible. The time could readily be shortened as
the sensitivity
was not the main focus, since the samples generally had high
concentrations.
The highest point in the standard curve was chosen so the
interaction would
be mass-transport limited and thus having a linear increase of
response during
the injection and avoiding the antibodies to approach
steady-state [31, 33].
This lead to linear standard curves without a plateau,
consequently giving a
higher resolution and precision. The dilutions of the standard
were typically
done by six serial 2- or 2.5-fold dilutions.
Even though the flow-rate might affect the response slightly
this parameter
was normally only set to reduce sample consumption.
-
- MATERIALS AND METHODS -
- 25 -
o Regeneration: Conditions, injection time, flow-rate
If not previously known, the regeneration conditions were found
by
regeneration scouting (section 3.2.3.1). The flow-rate was
typically set slightly
higher than the flow-rate for the analyte injection. The
injection time of
regeneration solution was also kept as short as possible to have
a short analysis
time but with a complete regeneration. A so called pre-dip was
used to avoid
dilution of the regeneration solution with running buffer during
analysis of
many samples.
An example of a sensorgram from the injection of a standard
curve is shown in Figure 3-3.
Some of the parameters discussed above are also illustrated in
the figure. The relative response
was found by subtracting the baseline response, before
injection, from the response after
injection as illustrated. Also seen in the figure is the short
sample injection time with almost
completely constant RU/s. The delay between end of injection and
regeneration was limited by
IFC washing in the instrument.
Figure 3-3: Example of sensorgrams from injection of standard
curve. Illustrating standard curves from six concentrations of
standard. In the example the injection and regeneration time are
illustrated with arrows.
x x
[standard] (µg/ml)
0
1000
2000
3000
0 20 40 60 80 100 120 140 160
Time (s)
x
x
x
x x
50
20
3.2
8
1.3 0.5
Injection Regeneration
Response (RU)
x
-
- CHAPTER 3 -
- 26 -
3.2.5. Activity and cross-reactivity experiment with capture
antibodies
The set-up for an experiment with a capture antibody was used to
ensure the mildest possible
treatment of the antibody by relieving it from the stress of
being immobilized by acidic
conditions during covalent coupling. The set-up is also
preferred if the regeneration conditions
for the antibody are not yet known.
It is possible to get false negative results if the immobilized
ligand binds to the binding
domain of the capture antibody making it unable to bind its
antigen. To eliminate false positive
results it is essential to also inject the analyte without the
capture antibody to ensure it does not
interact with the ligand alone.
The set-up is illustrated in Figure 3-4 supported with a
schematic sensorgram. As the capture
antibody is injected there is an increase in response. If there
is another increase in response as the
analyte is injected it is considered a positive interaction.
This is followed by regeneration of the
surface and a second capture antibody can be injected.
Figure 3-4: Set-up of method with capture antibody The first
antibody (blue) represents the ligand immobilized to the dextran
matrix on the sensor chip. The capture antibody (red) is injected
giving a response seen in the sensorgram. As the analyte (green) is
injected it gives a response in the sensorgram if the interaction
is positive.
Inject capture Ab
Inject sample Positive!
Regeneration
Immobilized ligand Capture antibody Analyte
Re
spo
nse
(RU
)
Time (s)
-
- MATERIALS AND METHODS -
- 27 -
3.2.6. Value transfer from international reference material to
calibrator
A protocol developed by the International Federation of Clinical
Chemistry and Laboratory
Medicine (IFCC) to transfer plasma protein concentration values
from an reference material
(RM), here ERM-DA470k/IFCC [14], to an internal calibrator,
called target material (TM), was
followed [16-17]. The procedure will be described here and
deviations from the protocol will be
accentuated. The value transfer was performed after the assays
were completed. While the
procedure is described here the results for each of the three
assays are presented in their
respective result section (4.1.3 for total IgG, 4.2.3 for IgG
subclass distribution and 4.3.3 for
albumin).
According to the protocol, the measurements were to be performed
three times a day on four
consecutive days, but due to time constraints the measurements
were only performed once a day
over three days but with duplicate measurements for both
calibration curve and samples. For
each day, new dilutions were made. Both the calibration curve
and the sample consisted of six
dilutions each. This yielded in 36 determinations each (6
dilutions * 2 replicates * 3 days = 36) for
the RM and the TM. An additional special dilution of RM was used
as control sample giving
another 6 determinations (2 replicates * 3 days = 6).
The RM was reconstituted according to the product sheet:
The vial was thawed in room temperature for one hour.
The vial was tapped gently to ensure all material settled on the
bottom.
Removing the screw cap.
The vial together with rubber stopper was weighed in gram with
four decimals.
1 mL of water was added, new weight recorded to acquire the
water mass.
The concentration after constitution was calculated with
Equation 3-1 below.
After one hour, the vial was inverted gently five times during
one hour.
Vail stored in room temperature overnight.
Equation 3-1
Where is the certified concentration and the actual
concentration after
reconstitution.
Six dilutions of the reconstituted RM served as standards for
the calibration curve. The
concentrations were evenly distributed over the measuring range
of the assay. To minimize the
sources of errors all volumes dispensed were controlled by
weighing and the actual dilutions with
four decimals were calculated. The densities of all liquids were
approximated to 1.
water
R
water
RR
m
C
m
mCC
0000.1''
'
RC RC
-
- CHAPTER 3 -
- 28 -
An example of the dilution scheme for human IgG is presented in
Table 3-1 below. In order
to get suitable volumes of RM and dilution buffer in the scheme,
a predilution of the RM was
performed. Additionally, to avoid pipetting small volumes,
Std.4, Std.5 and Std.6 were prepared
from Std.2, Std.3 and Std.5 respectively.
The relative concentration in percentage was calculated by
Equation 3-2 and these
concentrations represent the values on the x-axis in the
calibration curve. Another excel-
spreadsheet equivalent to that in Table 3-1 was filled in with
actual masses from pipetting where
actual dilutions and relative concentrations were calculated.
Hence, these were the values entered
into the method as concentration in percentage.
Equation 3-2
Where is the mass of reference material, the mass of dilution
buffer,
predilF is the predilution factor and Std.2 is the relative
concentration of Standard 2.
Human IgG - Intended predilution of the Reference Material
Dilution buffer, (g) 0,9500
Reference Material, (g) 0,0500
Predilution factor, FPredil 0,050000 RM reconst. conc. (CR)
Human IgG - Intended dilutions of the Reference Material 9259,82
μg/ml
Std. 1 Std. 2 Std. 3 Std. 4 Std. 5 Std. 6 Control
Dilution buffer, (g) 0,8000 1,2000 1,3000 0,4000 0,7000 0,6000
1,1000
Reference Material, (g) 0,1000 0,1000 0,0750 0,3000 0,3000
0,1000 0,0500
of predil of predil of predil of Std. 2 of Std. 3 of Std. 5 of
predil
Total mass, (g) 0,9000 1,3000 1,3750 0,7000 1,0000 0,7000
1,1500
Dilution factor 0,1111 0,0769 0,0545 0,4286 0,3000 0,1429
0,0435
Predilution factor 0,050000 0,050000 0,050000 0,003846 0,002727
0,000818 0,050000
Relative concentration, (%) 0,5556 0,3846 0,2727 0,1648 0,0818
0,0117 0,2174
Concentration, (μg/ml) 51,44 35,61 25,25 15,26 7,58 1,08
20,13
Aimed target conc. (μg/ml) 50,00 35,00 25,00 15,00 7,50 1,00
20,00
Table 3-1: Example of intended dilution scheme of reference
material for value transfer Predilutions for Std.4, Std.5 and Std.6
were calculated from the dilution of the standard it was prepared
from. All weights were recorded in grams with four decimals and
predilution factors with six decimals. Volumes were chosen to give
a concentration in μg/ml (calculated from CR and the relative
concentration) close to the aimed target concentration, which was
based on the concentration range of the assay. This dilution scheme
was followed and an equivalent excel-spread sheet was filled in
with actual masses where actual dilutions and actual relative
concentrations were calculated.
RM DilM
(when Std.4 is prepared
from Std.2)
or
DilR
Rpredil
MM
MFStdconcrel
1002..
DilR
R
MM
MStdStdconcrel
2.4..
-
- MATERIALS AND METHODS -
- 29 -
The target material was also diluted in six dilutions. As these
six dilutions were samples they
were aimed to all fall within the calibration curve generated. A
predilution was performed to a
concentration in the upper quarter of the standard curve. The
prediluted TM was added in
decreasing volumes to correspondingly increasing volumes of
dilution buffer yielding in a
constant total volume, with all weights recorded. An example of
dilution scheme of target
material for IgG is shown in Table 3-2. The dilution factor
denoted FT2 was calculated by
Equation 3-3 and was used as the x-value in the upcoming plot
and calculations. The actual
masses were weighed and entered in an excel-spreadsheet giving
the actual values of FT2.
Equation 3-3
Where is the mass of prediluted target material and the mass of
dilution buffer.
Human IgG - intended predilution of the Target Material
Dilution buffer, (g) 2,5000
Target Material, (g) 0,0400
Predilution factor, FT1 0,015748
TM estimated conc.
Human IgG - intended dilutions of the Target Material 2000,00
μg/ml
Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6
Dilution buffer, (g) 0,0000 0,0500 0,1000 0,2000 0,3000
0,3750
Target Material, (g) 0,5000 0,4500 0,4000 0,3000 0,2000
0,1250
Dilution factor, FT2 1,0000 0,9000 0,8000 0,6000 0,4000
0,2500
Predilution factor, FT1 0,015748 0,015748 0,015748 0,015748
0,015748 0,015748
Relative concentration, (%) 1,5748 1,4173 1,2598 0,9449 0,6299
0,3937
Concentration, (μg/ml) 31,50 28,35 25,20 18,90 12,60 7,87
Table 3-2: Example of intended dilution scheme of target
material for value transfer The highlighted values are dilution
factor FT2. All weights were recorded in grams with four decimals.
The actual values from the corresponding excel-spreadsheet will be
used as x-values in the upcoming plot and calculations.
When all dilutions were made, the assay was executed with one
set of standards, duplicates of
all samples, a duplicate of control sample and finally another
set of standards. The outcome was a
standard curve similar to the illustration in Figure 3-5, with
relative concentration in percentage
versus response in RU. In the evaluation software, the responses
of the samples were
interpolated on the standard curve giving a relative
concentration in percentage.
Figure 3-5: Schematic standard curve value transfer The standard
curve with relative concentration in percentage on the x-axis and
response in RU on the y-axis.
DilT
TT
MM
MF
2
TM DilM
Response, (RU)
Relative concentration, (%)
-
- CHAPTER 3 -
- 30 -
The average concentration for each sample was transformed with
Equation 3-4 to become a
y-value, or relative concentration factor FR, that was
comparable with the x-values, FT2. The
control samples were compared to expected relative
concentrations to give an indication of assay
performance. This was done by dividing the interpolated relative
concentration with the expected
relative concentration; a value of 1.0 equals a control with 100
% compared to expected.
Equation 3-4
Where FR is the relative concentration factor for sample i and
FT1 is the predilution factor.
The dilution factor FT2 was then plotted against the measured
relative concentration factor FR
for all six samples. This gave a plot similar to Figure 3-6. A
linear regression ( ) was
performed and if a zero intercept was within the confidence
interval a new regression was
performed with the intercept set to zero ( ). A zero intercept
means that there was no
matrix effects in the assay, i.e. the buffer conditions were the
same for TM and RM after
dilutions.
Figure 3-6: schematic plot of FT2 versus FR for value transfer
Linear regression with intercept set to zero for dilution factors
plotted against relative concentration factors. The slope is equal
to the ration of target material and reference material
concentrations as shown in Equation 3-5.
According to Blirup-Jensen et al [16], derivation not shown
here, the slope of the line is equal
to the ratio between the target and reference material
concentrations as in Equation 3-5.
Accordingly, the actual target material concentration CT was
calculated by the right side of
Equation 3-5.
Equation 3-5
Where CT is the actual target material concentration, CR the
reconstituted reference material concentration calculated in
Equation 2-1 and β the slope from the linear regression.
1001
)(
T
RF
isampleaverageiF
XY
0
RT
R
T CCC
C
y-value: relative concentration factor, FR
x-value: Dilution factor, FT2
-
- MATERIALS AND METHODS -
- 31 -
A transfer factor (TF) was calculated with Equation 3-6 for
transformation of results done
prior to this reference calibration. Previous results were
multiplied with the TF giving the correct
concentrations.
Equation 3-6
Where CT is the actual target material concentration after value
transfer and is
the previously known concentration of the target material (if
applicable).
This whole procedure was performed on three consecutive days
giving three linear regressions
with one slope each, giving three values on the actual target
material concentration CT from
which the average target material concentration was calculated.
The results for each assay are
presented in section 4.1.3 for total IgG, 4.2.3 for IgG subclass
distribution and 4.3.3 for albumin.
3.2.7. Biuret, total protein concentration assay
The total protein concentration was determined with the biuret
assay as described in the
European Pharmacopoeia 2.55.3 [2]. The assay involves a reaction
in alkali solution between
cupric ions and peptide bonds to form a complex with absorbance
at 546 nm. Preparation of the
biuret solution is described below. A protein standard
containing both HSA and hIgG was used
(80mg/ml, Sigma). Two different standard curves and protocols
were used depending on the
estimated samples concentration; see table x. The standard
curves included 5 points in 2 times
serial dilutions. The larger volume of the standards/samples for
the low calibration curve was due
to that the absorbance should be between OD 0.1-1 to be optimal
in measurement.
Volume Standard
5-80 mg/ml Standard 0.5-6 mg/ml
Standard 10µl 100µl
Sample 10µl 100µl
Biuret solution 200µl 100µl
Table 3-3: Volumes used for two standard curves.
Standards and samples were added in duplicates into a flat
bottom 96-well microplate. Biuret
solution was added and the microplate was incubated for 30
minutes on a shaker. The
absorbance was measured in a plate reader at 541 nm and the
concentration calculated by
constructing a linear calibration curve.
Preparations of biuret solution:
3.0 g CuSO4x5 H20 + 9.0 g C4H4KNaO6x4 H20 + 5.0 g KI
Add 800 mL milli-Q water, stir until dissolved
Add 100 mL 6.0 M NaOH
Fill up to 1000 mL with milli-Q water
old
T
T
C
CTF
old
TC
-
- CHAPTER 3 -
- 32 -
3.2.8. SDS-PAGE
In order to calculate the specific IgG or albumin concentration
with traditional methods, the
purity in percentage was estimated by SDS-PAGE and then
multiplied with the total protein
concentration from biuret.
The samples were diluted with water according to Table 3-4 to
get suitable amounts of protein
on the gel. To reduce the proteins prior to loading the samples
were mixed with sample loading
buffer (10 µl sample + 10 μl 4X NuPAGE sample buffer with 20%
β-mercapthoethanol) and
heated at 70°C for 10 minutes.
Sample concentration (µg/µl) Dilution
1-2 2x
2-9 5x
10-19 10x
20-70 50x
70-90 100x
>100 200x
Table 3-4: Sample dilutions for SDS-PAGE Simplified dilution
scheme of samples in order to load an appropriate amount of protein
onto the gel to avoid over-load or not having enough protein.
The gel was docked to the electrophoresis system and running
buffer (NuPAGE MOPS SDS
running buffer, Invitrogen) was added. 5 μl of molecular weight
marker (Precision Plus Protein
Dual Color Standard, Bio-Rad Laboratories AB) was added to the
first lane and 10 μl of sample
mixture to all other lanes. The gel was run for 10 minutes with
60 V to gather the protein bands
below the wells followed by 70 minutes with 150 V. The protein
bands were stained using
GelCode blue staining kit (Thermo Fisher Scientific) over night
while shaking, then destained in
water for another 24 hours.
Dyed gels were scanned on an ImageScanner III using Labscan 6.0
(GE Healthcare) and
analysed in ImageQuant TL 6.0 (GE Healthcare). The peaks on each
lane were identified and
cut-offs determined. The known proteins in the samples, such as
IgG, albumin and transferrin,
were recognized. An example gel is shown below in Figure 3-7
with the “pixelogram” analysis of
lane 10 in Figure 3-8. From the area under the curves the
relative quantity of that specific protein
was calculated in relation to the total curve area in the lane.
As seen in Figure 3-8 the relative
quantity of each protein was an estimation and therefore not
highly accurate. The relative
quantity was also interpreted as the purity of the proteins in
the sample.
-
- MATERIALS AND METHODS -
- 33 -
Figure 3-7 (left): Scanned picture of a SDS-PAGE gel Lane 1
contains the molecular weight marker and lane 2 to 12 samples from
different steps in the plasma fractionation process.
Figure 3-8 (right): Analysis of lane 10 from SDS-PAGE Example
showing the analysis of lane 10 from the gel in Figure 3-7. Peak 5
at 70 kDa was believed to corresponds to transferrin, peak 6 and 7
at 62 and 57 kDa to albumin and finally peak 9 and 12 at 50 and 21
kDa to IgG. The relative quantity, thereby also the purity, was
calculated to 51 % for albumin and 18 % for IgG in this sample.
As IgG consists of several chains linked with disulphide bonds
more than one band will
appear on a reduced gel as illustrated in Figure 3-9. If
detected they were added for the full IgG
composition.
Figure 3-9: IgG bands on reduced SDS-PAGE gel Due to complete or
incomplete reduction of disulphide bonds in IgG up to five
detectable bands occur. The highest relative quantity is that of
completely reduced heavy chain at ~55 kDa and light chain at ~22
kDa.
~155
~130
~75
~55
~22
250
150
100
75
50
37
25
20
IgG kDa #1 #2 Mw kDa
Lane: 1 2 3 4 5 6 7 8 9 10 11 12
Lane 10
-
- CHAPTER 3 -
- 34 -
3.2.9. ELISA
Enzyme-linked immunosorbent assay, or ELISA, was used to analyse
the IgG subclass
concentrations in samples to compare with the developed Biacore
IgG subclass distribution
assay. Peliclass human IgG subclass kit (Sanquin) was used for
the measurements. The kit
contained strips of wells, pre-coated with specific monoclonal
anti-human subclass antibodies.
Six strips of eight wells existed for every subclass. For each
experiment three strips for each
subclass was used giving 96 wells in total. The calibration
curve had five points in duplicates, two
blanks and one control in duplicate leaving space for five
samples in duplicates as illustrated in
Figure 3-10. If all six strips were used in one analysis this
would leave space for 17 samples.
Figure 3-10: Illustration of human IgG subclass ELISA kit
set-up. a) Calibration curve. b) Blanks. c) Control sample. Five
samples in duplicates can be analysed at once.
The product protocol for the kit was followed. Due to the
different abundances of the
different subclasses different ranges of standards were used.
Following a dilution scheme the
calibrator was diluted to eight points, ranging from 10000 times
to 1280000 times dilution. The
five most diluted were used for IgG1 and the five least diluted
were used for IgG2-4, see Table
3-5.
# Dilution IgG1 IgG2 IgG3 IgG4
ng/ml ng/ml ng/ml ng/ml
1 1:10000 - 368 45 59
2 1:20000 - 184 22 29
3 1:40000 - 92 11 15
4 1:80000 81 46 6 7
5 1:160000 41 23 3 4
6 1:320000 20 - - -
7 1:640000 10 - - -
8 1:1280000 5 - - -
Table 3-5: IgG subclass concentrations in ELISA calibrator
Concentrations of IgG1-4 in diluted standards for IgG1-4 in
Peliclass human IgG subclass kit according to manufacturer.
a a a a a a a a
a a a a a a a a
a a a a a a a a
a a a a a a a a
a a a a a a a a
b b b b b b b b
c c c c c c c c
α-hIgG1 α-hIgG2 α-hIgG3 α-hIgG4
-
- MATERIALS AND METHODS -
- 35 -
Samples and control sample were diluted 240000 times for IgG1
strips and 30000 times for
IgG2-4 strips. As the kit was only intended for plasma samples
and not purified IgG the purified
process samples were also prediluted 1, 2 and 4 times in order
to not exceed the calibration
curve. Also the HRP-conjugated secondary antibody had individual
dilutions for each IgG
subclass strip: 1:500, 1:3000, 1:2000 and 1:1000 for IgG1 to
IgG4 respectively.
Wells were washed four times with wash buffer.
100 μl of calibrators, control sample and samples were added to
their intended wells.
Incubated for 1 hour at 37°C.
Wells were washed four times with wash buffer.
100 μl of specifically diluted HRP-conjugated antibodies were
added.
Incubated for 1 hour at 37°C.
Wells were washed four times with wash buffer.
100 μl of ABTS-substrate diluted in substrate buffer were added
to all wells.
Incubated for 30 minutes at room temperature.
50 μl stop solution were added to all wells.
Plates were read in a SPECTRA Max PLUS 384 plate reader at 414
nm.
4-parameter standard curves were plotted and fitted. Individual
evaluation files were created
for each IgG subclass with one standard curve each. The software
calculated concentrations,
taking dilutions into consideration, giving the individual
subclass concentrations for the control
sample and process samples.
-
- RESULTS -
- 37 -
Chapter 4
4Results
4.1. Total IgG concentration assay
4.1.1. Evaluations of reagents for total IgG concentration
assay
The conditions used for the immobilization and regeneration of
the antibody used was
already optimized and performed according to instructions from
the manufacturer. The antibody
α-hIgG was diluted to 20 μg/ml in 10 mM sodium acetate pH 5.0
pre-concentration buffer and
injected for 6 minutes, typically resulting in an immobilization
level of 10000 RU. Regeneration
was performed with 3M MgCl2 for 30 seconds according to the
product protocol.
4.1.2. Assay development total IgG concentration
Initially the method to use a slope (RU/s) instead of the
relative response (RU) as a measure
of the signal was evaluated. Due to bulk effects from high
protein concentrations (e.g. when
detecting IgG losses in discarded samples with high albumin
level) and variable NaCl levels this
approach was not suitable for the assay. In order to keep the
analysis time to a minimum, the
injection time was kept to only 20 seconds. Owing to the great
performance of the monoclonal
anti-human IgG antibody it was possible to use a master standard
curve for at least one week of
measurements with over 1000 process samples. After samples
analysis the result-file was
appended in the evaluation software with a result-file
containing the standard curve, it required to
be with the same method and from the same chip, flow-cell and
immobilization.
-
- CHAPTER 4 -
- 38 -
At least one start-up cycle was necessary to condition the
surface for the analysis, especially if
a master standard curve from a prior measurement was used. One
or two control samples were
evenly distributed during the analysis. For example one with a
high concentration and one with a
low concentration on the standard curve.
4.1.2.1. Standard curve
The standard curve was set to start at 50 μg/ml with six
2.5-fold dilutions to approximately
0.5 μg/ml. The lower point was chosen to get a high sensitivity
of the assay and the higher point
to avoid the antibodies to be saturated and thereby reducing the
resolution on the standard curve
for higher concentrations, as discussed in 3.2.4. The samples
were then diluted to fit on the
standard curve and at the same time eliminate pH, buffer and
NaCl effects by dilution. The
standard curve for IgG can be seen in Figure 4-6.
4.1.2.2. Sample preparations
Samples were first diluted using a dilution factor based on the
estimated concentration,
followed by two two-fold dilutions to increase the number of
measurement points and to ensure
the sample concentrations fall within the standard curve.
Samples expected to contain IgG were
diluted 200 times, samples close to the final product with
estimated concentrations above 10
mg/ml were diluted 1000 times and samples expected not to
contain IgG were diluted at a
minimum 10 times to detect losses and to avoid pH, buffer and
NaCl effects. The effect of
dilution on samples with a high (500 mM) and low (0 mM) NaCl
level is exemplified in Table 4-1
concluding that the critical samples containing IgG with
dilutions above 200 times were
completely diminished from NaCl effects.
Dilutions in 150 mM NaCl (HBS-EP+)
Sample 10X 20X 40X 200X 400X 800X
500 mM NaCl 185 167 159 152 151 150
0 mM NaCl 135 142 146 150 150 150
Table 4-1: Calculated NaCl levels in diluted samples Dilutions
of samples with high and low NaCl level into HBS-EP+ with 150 mM
NaCl. Samples diluted 10 and 20 times have a moderately increased
or decreased level from optimal which might give a positive or
negative bulk, as described in section 2.3. Although for samples
diluted 200 times and more all effects are diminished, these are
also the sample with the most critical concentrations.
-
- RESULTS -
- 39 -
4.1.2.3. Assay procedure
Immobilization of 20 μg/ml α-hIgG for 6 minutes resulted in
approximately 10000 RU ligand.
After conditioning start-up cycles the calibrant was injected in
increasing concentrations, if a
master standard curve was not employed. Thereafter samples with
increasing concentrations
within the three dilutions were injected in duplicates. Also
evenly distributed control-samples
were injected. Regeneration was performed with 30 second
injection of 3M MgCl2 after each
cycle. When evaluating the results one dilution giving either a
too low concentration or a too high
concentration on the standard curve was excluded giving four
determinations for each sample
(n=4). For samples without expected IgG, i.e. 10 times diluted
samples, if possible the two most
diluted samples were chosen to avoid buffer, pH and NaCl
effects. F