Stefanie Schindler
Optimization and novel applications of the in vitro
pyrogen test (IPT) using human whole blood
Dissertation
Universität Konstanz
November 2005
Optimization and novel applications of the in vitro
pyrogen test (IPT) using human whole blood
Dissertation
zur Erlangung des akademischen Grades
des Doktors der Naturwissenschaften
an der Universität Konstanz (Fachbereich Biologie)
vorgelegt von
Stefanie Schindler
Tag der mündlichen Prüfung: 23. 01. 2006
Referenten: Prof. Dr. Dr. T. Hartung
Prof. Dr. A. Wendel
PD Dr. Bert-André Zucker
LIST OF PUBLICATIONS
List of publications:
Major parts of this thesis are published or submitted for publication:
Schindler, S., and Hartung, T. (2005).
Development, validation and applications of the in vitro pyrogen test (IPT)
based on human whole blood (submitted to J Clin Immunol)
Hoffmann, S., Peterbauer., A., Schindler, S., Fennrich, S., Poole, S., Mistry,
Y., Montag-Lessing, T., Spreitzer, I., Loschner, B., van Aalderen, M., Bos, R.,
Gommer, M., Nibbeling, R., Werner-Felmayer, G., Loitzl, P., Jungi, T., Brcic,
M., Brugger, P., Frey, E., Bowe, G., Casado, J., Coecke, S., de Lange, J.,
Mogster, B., Naess, L. M., Aaberge, I. S., Wendel, A., and Hartung, T., 2005.
International validation of novel pyrogen tests based on human monocytoid
cells. J Immunol Methods 298:161-73
Schindler, S.; Asmus, S., von Aulock, S., Wendel, A., Hartung, T., and
Fennrich, S., 2004. Cryopreservation of human whole blood for pyrogenicity
testing. J Immunol Methods 294: 89-100
Schindler, S., Rosenberg, U., Schlote, D., Panse, K., Kempe, A., Fennrich,
S., and Hartung, T., 2005. Application of the InVitro Pyrogen Test (IPT) based
on cryopreserved human whole blood for lipidic parenterals (Pharmeuropa, in
print)
Schindler, S., Spreitzer, I., Hoffmann, S., Hennes, K., Halder, M.,
Brügger, B., Frey, E., Montag-Lessing, T., Löschner, B., Poole, S., and
Hartung T., 2005. International validation of pyrogen tests based
on cryopreserved human primary blood cells (J Immunol Methods, in print)
LIST OF PUBLICATIONS
Further publications, not integrated into this thesis:
Schindler, S., Bristow, A., Cartmell, T., Hartung, T., and Fennrich, S. 2003.
Comparison of the reactivity of human and rabbit blood towards pyrogenic
stimuli. ALTEX 20: 59-63
Schindler, S., and Hartung, T., 2002. Comparison and validation of novel
pyrogen tests based on the human fever reaction.
Dev Biol 111: 181-6
Mazzotti, F., Beuttler, J., Zeller, R., Fink, U., Schindler, S., Wendel, A.,
Hartung, T. and von Aulock, S. 2006. In vitro Pyrogen Test - a new test
method for solid medical devices. J Biomed Mat Res: Part A, in print
Spreer, A., Gerber, J., Hanssen, M., Schindler, S., Hermann, C., Lange, P.,
Eiffert, H., and Nau, R. 2006. Dexamethasone increases hippocampal
neuronal apoptosis in a rabbit model of Escherichia coli meningitis. Pediatric
Research, in print
ACKNOWLEDGEMENTS
Acknowledgements
The work presented here was carried out between January 2002 and March
2005 at the chair of Biochemical Pharmacology at the University of Konstanz
under the supervision of Prof. Dr. Dr. Thomas Hartung.
I especially want to thank my supervisor Thomas Hartung for his advice, his
confidence, and for the excellent working facilities.
Also special thanks to Dr. Stefan Fennrich for his encouragement and
friendship.
Many thanks go to Prof. Dr. Albrecht Wendel for welcoming me in his
department.
Special thanks to the pyrogen team; Ina Seuffert for the introduction into the
topic, Gregor Pinski, Ilona Kindinger and Silvia Asmus for their friendship and
never tiring support, their valuable ideas and their threats of physical violence
should I ever give up on this project.
I want to thank all my lab colleagues for their continuous help. My special
thanks in this respect goes to Dr. Sonja von Aulock and to Dr. Sebastian
Hoffmann.
I thank all members of the “Lehrstuhl Wendel” for their support, their
contribution to the outstanding working atmosphere and for five wonderful
years.
And last but not least I thank my parents Klaus and Birgit Schindler.
ABBREVIATIONS
Abbreviations
AAMI American Association of Medical
Instrumentation
AWIPT absorb and wash in vitro pyrogen test
BAL bronchoalveolar lavage
BET bacterial endotoxin test
cAMP cyclic adenosinmonophospate
CD cluster of differentiation
CFU colony forming unit
COX cyclooxygenase
CRP C-reactive protein
CV coefficient of variation
DIC disseminated intravascular
coagulation
DL developing laboratory
DMSO dimethyl sulfoxide
ECVAM European Center for the Validation of
Alternative Methods
ELC endotoxin limit concentration
ELISA enzyme-linked immunosorbent assay
EU endotoxin unit
GLP good laboratory practice
HD hemodialysis
HSA human serum albumin
ICE interleukin-1 converting enzyme
IL interleukin
IPT In Vitro Pyrogen Test
IU International Unit
NIBSC National Institute for Biological
Standards and Controls
LAL Limulus Amoebocyte Lysate
LBP LPS-binding protein
ABBREVIATIONS
LoD Limit of Detection
LPS lipopolysaccharide
LTA lipoteichoic acid
LVP large volume parenteral
MID minimum interference dilution
MM-6 Monomac-6
MVD maximum valid dilution
PAMPs pathogen-associated molecular
patterns
PBMCs peripheral blood mononuclear cells
PBS phosphate buffered saline
PEI Paul Ehrlich Institute
PG prostaglandin
POD peroxidase
PPC positive product control
NFКB nuclear factor kappa B
NL naive laboratory
NPC negative product control
NSAID non-steroidal anti-inflammatory drug
OD optical density
OVLT organum vasculosum laminae
terminalis
PM prediction model
PMN polymorphonuclear
PTFE polytetrafluorethylene
RT room temperature
SOP standard operating procedure
SVP small volume parenteral
TLR toll-like receptor
TMB tetramethylbenzoate
TNF Tumor Necrosis Factor
USP United States Pharmacopoeia
WBT whole blood test
ABBREVIATIONS
WHO World Health Organisation
TABLE OF CONTENTS
Table of Contents
1. Introduction 1
1.1. Pyrogens 1
1.1.1. Lipopolysaccharide (endotoxin) 1
1.1.2. Non-endotoxin pyrogens 2
1.2. Traditional pyrogen tests 3
1.2.1. Rabbit pyrogen test 3
1.2. 2. Limulus amoebocyte lysate test (LAL) 4
1.3. Mechanism of fever 4
1.4. Cell-based pyrogen tests 5
1.5. Human whole blood test (IPT) 6
2. Aims of the study 7
3. Development, validation and applications of the in vitro
pyrogen test (IPT) based on human whole blood 8
3.1. Abstract 8
3.2. Introduction 9
3.3. Basic principle of the whole blood test 11
3.4. Comparison of the in vitro reaction of the human whole
blood test to rabbit whole blood 13
3.5. Establishment of the IPT as a test for biologicals 13
3.6. Validation
3.7. Development of the commercially available IPT kit 15
3.8. Special adaptations 18
3.9. Conclusion 28
3.10. Appendix
3.11. Acknowledgements 28
TABLE OF CONTENTS
4. International validation of novel pyrogen tests based
on human monocytoid cells 29
4.1. Abstract 30
4.2. Introduction 31
4.3. Materials and Methods 33
4.4. Results 41
4.5. Discussion 50
4.6. Acknowledgements 52
5. Cryopreservation of human whole blood for
pyrogenicity testing 53
5.1. Abstract 53
5.2. Introduction 53
5.3. Materials and Methods 54
5.4. Results 58
5.5. Discussion 71
5.6. Acknowledgements 73
6. International validation of pyrogen tests based on
cryopreserved human primary blood cells 74
6.1. Abstract 75
6.2. Introduction 76
6.3. Materials and Methods 76
6.4. Results 83
6.5. Discussion 92
TABLE OF CONTENTS
7. Pyrogen testing of lipidic parenterals with a novel
in vitro test 95
7.1. Abstract 96
7.2. Introduction 97
7.3. Materials and Methods 98
7.4. Results 101
7.5. Discussion 110
7.6. Conclusion 112
8. Summarizing discussion 113
9. Summary 118
10. Zusammenfassung 119
11. References 120
INTRODUCTION
1
1 Introduction
1.1 Pyrogens
The term “pyrogen” derives from the greek word “pyros” (fire). Pyrogens are
therefore substances that have been recognized to cause fever in the
organism. The relation of bacteria and fever was first recognized by
Semmelweis (1) and Lister (2). The association of fever and intravenous
injection, on the other hand, dates back to the eighteenth century, when van
Haller noticed that the injection of putrid materials caused severe fever
reactions (3). Panum, with the help of Virchow, was the first to state that the
substance responsible was heat-stable, water-soluble, alcohol-insoluble, and
independent of the presence of living bacteria (4). The term “pyrogen” was
apparently used first by Billbroth (5).
At the end of the 19th century, Centanni first reproducibly isolated a
substance from a variety of Gram-negative bacteria which he called
pyrotoxina, which was most probably the first purified endotoxin in history (6).
Injection fevers associated with intravenously applied parenterals were first
systematically investigated by Hort and Penfold in 1912 (7), who injected
them intravenously into the rabbit, measured the fever reaction, and classified
the bacteria into pyrogenic and non-pyrogenic. Basically, these were the first
rabbit pyrogen tests. Seibert then proved that the fever reactions were
caused by filterable, heat-stable pyrogens from Gram-negative bacteria, a
finding which was later confirmed by Rademaker, who already stressed the
importance of avoiding contaminations in parenterals and differentiated
between the terms “sterile” and “pyrogen-free” (8, 9). World War II then
brought the development of large volume parenterals as volume substitution
for injured soldiers. The occurrence of severe fever reactions resulted in a
collaborative study establishing the rabbit pyrogen test (10, 11) and its
incorporation into the US Pharmcopoeia in 1942.
1.1.1. Lipopolysaccharide (endotoxin)
Endotoxin as a component of the cell walls of Gram-negative bacteria is the
INTRODUCTION
2
most potent and the most extensively studied pyrogen. Due to the fact that
Gram-negative bacteria are ubiquitous, contaminations of parenterals with
endotoxin pose a constant threat to the health of patients. Endotoxins are
released from the cell not only after lysis, but are shed constantly from the
living bacterium as well (12). Lipopolysaccharide (LPS) is a highly purified
(protein-free) form of endotoxin. Chemically, they are heat-stable substances
with three distinct regions: the lipid A portion, which has been shown to be
responsible for the pyrogenic activity (13, 14), the core polysaccharide, and
the antigenic O-specific side chain. The biological activities of endotoxins do
not restrict themselves to causing fever and other inflammatory reactions, but
also include complement activation, hypotension, and activation of the
coagulation system, all of which can lead to severe complications, up to
hypovolemic shock, disseminated intravascular coagulation (DIC) and death.
A maximum endotoxin contamination of 50 pg/ml (0.5 ng/kg) was first
published by the Bureau of Drugs in 1980 (15), apparently with no scientific
study having been performed to confirm this very restrictive threshold. In
2005, a study at the Paul-Ehrlich Institute (PEI) in Germany fully confirmed
this limit (16).
1.1.2. Non-endotoxin pyrogens
Substances that have pyrogenic properties but are not of an endotoxin nature
include enterotoxins (17, 18), exotoxins, (19), viruses (20), peptidoglycan (21-
23) and fungi (24, 25). Since Gram-positive bacteria are as frequent as
Gram-negative bacteria, the pyrogens of the former can be a serious health
hazard as well. A major component of the Gram-positive cell wall is the
peptidoglycan, which consists of β-1,4 linked N-acetyl-D- glucosamin and N-
acetyl muramic acid, and was shown to have pyrogenic properties similar to
those of endotoxin (21). The other prominent pyrogen of Gram-positive
bacteria called lipoteichoic acid (LTA) was successfully purified in an
endotoxin-free and biologically active manner in 2001 (26).
INTRODUCTION
3
1.2. Traditional pyrogen tests
1.2.1. Rabbit pyrogen test
The rabbit pyrogen test has been the gold standard in pyrogen testing since
1942, when it was introduced into the USP (United States Pharmacopoeia).
The rabbit species was chosen by Seibert, who also discovered the
pyrogenic principle (8). In 1941, the need for pyrogen testing of LVP (large
volume parenterals) due to World War II caused the Committee of Revision of
the USP to authorize the first USP collaborative study of pyrogens with
pyrogen filtrates of Pseudomonas aeruginosa. The results of this study led to
the incorporation of the rabbit test in the 12th edition of the USP in 1942. In its
simplest form, the test involves measuring a rise in body temperature for 3
hours following intravenous injection of a test solution into the marginal ear
vein at a volume of not more than 10 ml/kg. Temperature is to be measured
by a clinical thermometer inserted into the rectum of the rabbit at a depth of
not less than 7.5 cm. Rabbit breeds used for testing are New Zealand Whites,
Belgium Whites, Chinchillas and Dutch Belts. Differences in sensitivities of
various strains have been investigated by van Dijck et al (27). Animals of one
single sex are preferred, and there have been reports about male rabbits
being more sensitive to pyrogens than females (28).
The test is positive if the sum of the rises in three rabbits exceeds 2.65 °C.
The rabbit has a labile thermoregulation and tends to give false-positive
results. Also, the very rigid fixation and the handling (injection procedure) can
cause a hyperthermia due to excitement. On the other hand, it has been
reported that the fixation and lack of movement can cause a hypothermia
yielding false-negative results (29). Comparisons between the reactivity of
humans and rabbits in vivo by Greisman 1969 showed that the threshold
towards three endotoxin preparations was comparable, but that the humans
respond more vigorously than the rabbits (30).
INTRODUCTION
4
1.2.1. Limulus amoebocyte lysate test (LAL)
When in contact with the lipid A portion of endotoxin, the amoebocytes from
Limulus polyphemus (horseshoe crab) coagulate due to an enzymatic
reaction (31, 32). In the presence of calcium, the clotting enzyme zymogen is
activated by a serin protease and acts on coagulogen, a clottable protein in
the lysate, producing a smaller clot protein. The clotting can be observed by
turning the tube with the lysate 180° (clot end point LAL) or, in a more
quantitative way, by the turbidimetric LAL, which measures kinetically ranges
of the clotting. The basic principle has been improved on and modified in
many ways (33). A sensitivity of 0.0005 µg/ml was determined by the
developers.
The lysate is prepared by placing the crabs in restraining racks and inserting
a needle through the muscular hinge between the cephalothorax and the
abdominal region. Hemolymph is then drawn from the cardiac chamber into a
container with anticoagulant. After collection, the amoebocytes are
centrifuged and the supernatant is discarded. After 2-3 washing steps, the
cells can be subjected to osmotic shock by adding distilled water and the
intracellular lysate is released. The bled crabs are then thrown back into the
sea, and their survival rate is unknown. In some countries (e.g. Japan) the
crabs are squeezed in a mill.
One of the drawbacks of the LAL is that it only detects endotoxin (34, 35).
The pyrogenic potency of non-endotoxin substances has been recognized
since the 1960s, leaving a safety gap when performing pyrogen tests with the
LAL. Contaminations of drugs with Gram-positive bacteria, fungi or their
fragments/toxins are not an unlikely event.
1.3. Mechanism of fever
The concept of a substance produced in the mammalian organism in
response to pyrogens which is causative in the genesis of fever dates back to
1948 (36). This substance, which was produced by immune cells evoked
fever when injected into healthy rabbits and was then called endogenous (or
INTRODUCTION
5
leukocytic) pyrogen. Dinarello et al. could demonstrate, that this endogenous
pyrogen consisted of two distinct proteins (37), probably pro-Interleukin-1 and
Interleukin-1 (IL-1). Other, similar mediators of fever were found later and
were termed Interleukin-6 (IL-6) and Tumor Necrosis Factor-α (TNF-α).
During a response to pyrogens, they are secreted by a subfraction of the
white blood cells, the monocytes, and are called proinflammatory cytokines. It
is of considerable interest that the receptors recognizing pathogen-associated
molecular patterns (PAMPs) of bacteria, the so-called toll-like receptors (tlr)
shares in its cytoplasmic domain the signaling areas with the IL-1 receptor
(38). Additionally, all pyrogenic cytokines share a common intracellular
pathway which results in the activation of the nuclear factor-κB (NF- κB). The
current understanding of the mechanism of fever in the mammal is that this
transcription factor results in the expression of the enzyme cyclooxygenase-2
(COX-2) which results in prostaglandin (PG) E2 synthesis. Mice deficient in
COX-2 did not develop fevers in response to LPS, IL-1, IL-6 or TNF (39-42).
Specifically one of altogether four PGE2 receptors in the brain, the EPR-3, is
required to develop fever (43, 44), probably via the induction of a second
messenger such as cyclic adenosinmonophosphate (cAMP) (45). That IL-1β
is the most potent fever inducer compared to IL-6 and TNF-α when injected
intravenously into rabbits could be demonstrated (46, 47). These findings
formed the basis for the development of cell-based in vitro assays which are
described in the next chapter.
1.4. Cell-based pyrogen tests
The discovery that white blood cells produce cytokines in a dose-dependent
manner in response to pyrogens led to the development of altogether six in
vitro assays based on primary human blood cells or cell lines. All of them
have the same basic concept of incubating the substance in question at 37°C
with the cells, and, as a second step, measuring the cytokine production (or,
in one case, nitric oxide) by an enzyme-linked immunosorbent assay (ELISA).
Four assays have been successfully validated in an international
collaborative study and are described in detail in the publication of Hoffmann
INTRODUCTION
6
et al. 2005 (48). One of these assays was the human whole blood test (IPT)
whose further development is described here.
1.5. Human whole blood test (IPT)
A new way of measuring pyrogens has been introduced in 1995 by Hartung
and Wendel (49). Basically, fresh heparinized human whole blood is diluted in
physiological saline and brought together with the sample. In the case of
pyrogenicity, the monocytes produce IL-1β in vitro over a period of 10-24
hours at 37°C which can be measured by a specific ELISA the next day. The
test has a detection limit of 0.25 EU/ml and has the advantage that it is
performed with the cells of the relevant species, that is, the human reaction is
tested.
The ELISA (Enzyme-Linked-Immunosorbent Assay) is an assay based on the
reaction of specific antibodies towards an antigen, in this case IL-1β. An
antibody is bound to a microtiter plate with high protein binding capacity; the
pyrogen-stimulated cell supernatant is added to the antibody and the cytokine
binds. After a washing step, a second, labeled detection antibody is added
which also binds to the antigen; the label is in this case biotin, which binds
with high affinity to avidin coupled to POD (horseradish peroxidase). After a
second washing step, substrate, in this case TMB (Tetramethylbenzidine) is
added. The enzymatic reaction of the POD with the TMB changes the color of
the latter from colorless to blue and the antibody-antigen reactions are made
visible.
AIMS OF THE STUDY
7
2 Aims of the study
Pyrogen testing of parenterals has been performed routinely in vivo in the
rabbit since the early 1940s. Recently, a cell-based in vitro alternative has
been developed which aims to replace the rabbit pyrogen test as an
alternative method. The European legislation clearly states that animal testing
is forbidden if there is a viable and validated in vitro alternative available.
Making the human whole blood test (IPT) a standardized and commercially
available alternative to the rabbit was the goal of the following work.
• The first part of this thesis validated the human whole blood test in an
international collaborative study including laboratories from England,
Switzerland, Norway, the Netherlands and Germany and control institutions
such as the Paul-Ehrlich Institute, Germany, and the European Centre for the
Validation of Alternative methods.
• The second part of this thesis standardized the most critical and the
most crucial reagent: the human whole blood. In order to make this highly
varying and perishable component of the assay more reliable and available, a
method for cryopreserving the blood was developed, and a pooling protocol
was found which levels out the interindividual differences of the human
donors.
• As a third step, the whole blood test using the newly developed
cryopreserved blood was validated in an international collaborative study
including three different laboratories.
• The last part extended the application possibilities of the new test
towards testing not hydrophilic, but lipophilic substances in order to avoid
large numbers of animal experiments. The testing for pyrogens in so-called
small volume perenterals, e.g. lipophilic drugs, is obligatory since January
2004 due to a change in European Pharmacopoeia.
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
8
3 Development, validation and applications of the in vitro
pyrogen test (IPT) based on human whole blood
Stefanie Schindler*, Sonja von Aulock* and Thomas Hartung+ *
* Biochemical Pharmacology, University of Konstanz, Universitätsstr. 10,
D-78457 Konstanz + ECVAM, Institute for Health and Consumer Protection, Joint Research Centre,
European Commission, I-21020 Ispra (VA)
Corresponding author
Thomas Hartung, MD, PhD
ECVAM
Institute for Health and Consumer Protection
Joint Research Centre
European Commission
I-21020 Ispra (VA)
e-mail: [email protected]
Tel: +39-0332-785939
Fax: +39-0332-786297
3.1. Abstract
Microorganisms such as Gram-negative or Gram-positive bacteria, viruses and
fungi contain components that activate the innate immune system. These
components, called pyrogens (Greek: pyros = fire), can occur independently of
viable microorganisms and are a major safety concern in parenterally
administered drugs, since they can cause severe reactions such as fever, organ
failure and shock in the recipient. So far, these drugs have been tested by
injecting them intravenously into rabbits and measuring their fever reaction or
alternatively by the Limulus Amoebocyte Lysate (LAL) test, employing the
coagulation of the hemolymph lysate of Limulus polyphemus. Both tests have
inherent limitations. A new in vitro test based on human whole blood, capable of
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
9
measuring all pyrogens relevant to the human patient was introduced and
validated recently. This review describes its principle, development, validation
and the wide spectrum of applications, such as for testing of medical devices,
blood products, lipidic parenterals and air quality. This alternative method aims
to replace fully the rabbit pyrogen test.
Key words
In vitro pyrogen test; interleukin-1β; validation study; alternatives to animals
3.2. Introduction
Pyrogens, as fever-inducing substances of microbial origin, can derive from
dead or viable bacteria, viruses or fungi. Therefore, they can occur even in
sterile environments. Contaminations of parenterals with such substances can
induce local or systemic inflammatory reactions in the recipient, intended to
eliminate an invading pathogen, including a rise in body temperature, but also
more severe adverse reactions such as shock, disseminated coagulation, organ
failure and even death. Therefore, the testing of parenterals prior to batch
release is obligatory for manufacturers.
The best-known fever-inducing contaminant is a component of the cell wall of
Gram-negative bacteria, i.e. endotoxin or lipopolysaccharide (LPS). Pyrogenic
components of Gram-positive bacteria are equally important and include
lipoteichoic acid (LTA) (26) and peptidoglycan (21 - 23). Further possible
pyrogenic contaminants are exotoxins (19), enterotoxins (17, 18) , viruses (20),
and fungal components (24, 25).
Classical pyrogen tests
Testing for pyrogens has been a major issue since the appearance of large
volume parenterals in the 1930s. These bore a label claim of being pyrogen-
free as asserted by the rabbit pyrogen test. This drew attention to the need for
an official test procedure for non-pyrogenicity, which was strengthened by the
heavy demand for large volume parenterals in World War II. A collaborative
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
10
study was initiated to develop the rabbit pyrogen test (10, 11), which led to the
incorporation of the rabbit pyrogen test into the pharmacopoeias. Since then, all
parenterals must be tested for pyrogens. This involves the measurement of the
rabbit’s body temperature after the application of not more than 10 ml/kg
bodyweight of the substance to be tested. The very rigid fixation of the rabbit
and the handling (injection procedure) can cause hyperthermia due to
excitement and therefore lead to false-positive results. On the other hand, it has
been reported that the fixation and lack of movement can cause a hypothermia
yielding false-negative results (29).
In 1964, Levin and Bang published that the hemolymph of the horseshoe crab
Limulus polyphemus coagulates upon contact with endotoxin. This led to the
development of the Limulus amoebocyte lysate (LAL) test, which is employed to
exclude endotoxin contamination in parenteral drugs (31, 32). The Limulus is
collected from beaches, its hemolymph is drawn out by puncture and the
animals are then thrown back into the sea. 10 to 20 percent do not survive the
bleeding procedure (50-52). The mortality associated with collecting, shipping
and handling the animals remains unknown. The LAL has not been able to
replace fully the rabbit test, since it is defined not as a pyrogen test, but as an
endotoxin test, which fails to recognize e.g. Gram-positive or fungal
contaminants, toxoids, or viral antigens. Due to the crucial role of Gram-
negative endotoxin, it was nevertheless possible to substitute most
pharmacopoeial pyrogen testing with a mere endotoxin test. Additionally, the
LAL does not reveal the biological potency of a given endotoxin in the mammal,
which can differ between bacterial strains by a factor of up to 10’000 (53). Most
importantly, however, certain products tested in rabbits cannot be tested in the
LAL, e.g. various biologicals and vaccines, due to interference.
Fever reaction in the mammal
The finding that mammalian immune cells produce an endogenous pyrogen
when in contact with pyrogenic materials dates back to 1948 (36). Bennett et al.
could identify leukocytes as the source of this factor in 1953 (54). The nature of
this substance was further elucidated by Dinarello et al. (37), who identified two
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
11
distinct proteins, probably the pro- and the mature form of interleukin 1β (IL-1β).
The pyrogenicity of IL-1β, when injected at very low doses into rabbits, was
proven by Dinarello et al. 1991 (46). IL-6 and TNF-α, which were isolated later,
were found to be pyrogenic cytokines as well, though only at higher doses (46,
47).
The current understanding of the mechanism of fever in the mammal, as
reviewed by Dinarello 2004 (55), is that these proinflammatory cytokines bind to
receptors on the blood side of the organum vasculosum laminae terminalis
(OVLT) and initiate the expression of the enzyme cyclooxygenase-2 (COX-2),
which mediates prostaglandin (PG) E2 synthesis. Mice deficient in COX-2 do
not develop fever in response to injection with LPS, IL-1β or IL-6 (39-41).
Specifically one of altogether four PGE2 receptors in the brain, the EPR-3, is
required to develop fever (43), probably via the induction of a second
messenger such as cAMP (45). Thus, the pyrogenic cytokines cause a change
in the set-point of body temperature in the hypothalamus and are therefore the
mediators responsible for initiating the fever reaction. The finding that
monocytes, a subfraction of the white blood cells, secrete proinflammatory
cytokines such as IL-1β upon contact with pyrogenic material was the basis for
the development of the whole blood test as a pyrogen test (49).
3.3. Basic principle of the whole blood test
Blood incubation
The procedure is described in detail by Hoffmann et al. (48). Briefly, freshly
drawn, heparinized human whole blood from a healthy donor is diluted in
physiological, pyrogen-free clinical grade saline and brought together with the
test sample. In response to pyrogens, the monocytes contained in the blood
sample produce proinflammatory cytokines in a dose-dependent manner. The
proinflammatory cytokine IL-1β is measured by ELISA.
ELISA procedure
The IL-1β or IL-6 in the sample is sandwiched between a monoclonal coat
antibody and a polyclonal peroxidase-labeled detection antibody. Unbound
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
12
material is removed by washing. The peroxidase metabolizes e.g.
tetramethylbenzidine. The reaction is stopped with acid and the optical density
(OD) is measured at 450 nm.
Controls
As an assay control, a dose-response curve of an LPS from E. coli O111: B4 is
performed in parallel in each assay. This LPS is calibrated to the international
WHO reference standard from E. coli O113: H10 (56). The dose-response
curve must contain the concentration 0.5 EU/ml and a negative control. The IL-
1β released in response to the concentration of 0.5 EU/ml must test positive
when compared to a negative control for the experiment to be valid. 0.5 EU/ml
corresponds to 50 pg/ml of the international reference standard and is
considered the threshold endotoxin concentration that causes fever in the most
sensitive rabbit strains. This threshold was confirmed by a study performed at
the Paul-Ehrlich Institute in 2005, which analyzed 171 rabbits (16).
Testing for interference
In order to test for a given substance’s interference with the activity of the
monocytes, samples (pure or diluted) are incubated together with a 0.5 EU/ml
concentration of the LPS dose response curve. The mean OD of the spiked
sample must be within a 50-200% range of the 0.5 EU/ml concentration of the
dose response curve. If this is not the case, the sample has to be diluted until
the interference criteria are met.
Development of the Gram-positive standard lipoteichoic acid (LTA)
LTA from Staphylococcus aureus was first purified in a biologically active and
endotoxin-free quality by Morath et al. (26). Later, the improved purification
procedure was applied to produce LTA from Bacillus subtilis (57). The
successful identification of the purified LTA as a pyrogenic substance, which is
negative in the LAL (57) and therefore represents a pyrogenic principle that is
only recognized by the rabbit pyrogen test and the cell-based assays, led to the
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
13
inclusion of the Gram-positive standard LTA derived from B. subtilis into the IPT
procedure.
In order to make this method commercially available and replace the rabbit
pyrogen test, the following steps were taken:
3.4. Comparison of the in vitro reaction of human whole blood with
that of rabbit whole blood
Since the human whole blood test (WBT) aims to replace the rabbit pyrogen
test, the sensitivity of both species towards different pyrogenic stimuli was
compared using human and rabbit whole blood. For this, a rabbit whole blood
test was developed which followed the procedure of the human whole blood
incubation in every detail (58). Overall, the IL-1β response of the rabbits
towards different pyrogenic stimuli was comparable to that of humans. In the
case of the Gram-positive stimulus, LTA, the rabbit blood was less sensitive
than human blood, thus confirming the human whole blood test as an equal or
even superior test system to reflect the human response.
3.5. Establishment of the IPT as a test for biologicals
Biologicals, such as protein solutions, cytokines, antibodies, heat shock
proteins, blood coagulation factors and vaccines for intravenous use, pose a
particular problem in pyrogen testing. They can influence the LAL results due to
their characteristics, such as color and viscosity, and they are potentially
immunogenic in the rabbit, causing fever reactions that are independent of
contaminations. In any case, if immunogenic substances are tested, the animals
may only be used once, which results in extremely high costs for the
manufacturers. The IPT does not pose such problems. Some examples of the
application of the IPT for pyrogen testing of such samples are given below.
Control of vaccines
In 2001, new batches of vaccines against early summer meningoencephalitis
were released that caused severe fever reactions in some recipients. Although
they were negative in the LAL, these batches gave a high signal in the WBT.
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
14
This phenomenon could be shown to be due to the removal of the mercury-
containing additive thiomersal. The additive suppressed the IL-1β response in
the WBT, making it likely that it also suppressed the pyrogenic property of the
attenuated virus in the vaccine in earlier batches (59).
In 2003, Carlin and Viitanen demonstrated that trivalent vaccines (diphtheria,
tetanus and polio), which tested negative in the LAL, were powerful inducers of
IL-6 in 4 out of 8 donors in the whole blood incubation (60). This difference
between the LAL and the WBT could be attributed to the toxoid of
Corynebacterium diphteriae, and, to a lesser extent, to that of Clostridium tetani,
both non-endotoxin pyrogens. (61). Additionally, in both studies, the authors
found pronounced differences in the IL-6 and IL-1β response of different donors
towards the vaccines and their components, although they displayed highly
conserved LPS reactivity. This indicated a more variable interindividual
sensitivity of human donors towards these non-endotoxin stimuli. Nonetheless,
it was demonstrated that pyrogenic reactions towards non-endotoxin stimuli can
be just as vigorous as those towards endotoxin. These results show that the
rabbit pyrogen test cannot be replaced by the LAL for vaccines, but that only the
measurement of the cytokine response of primary human cells, e.g. the WBT,
represents an adequate alternative.
Measurement of albumins
Pyrogenic reactions of human patients after the administration of human serum
albumin, which had tested negative in the rabbit, were observed in 1978 (62). In
this study, the LAL yielded positive results without perceivable patient reactions.
Pool et al. (63) tested 22 batches of human serum albumin (HSA), fibronectin
and stabilized human serum solutions using artificial contaminations of
endotoxin and LTA from B. subtilis. None of these products interfered with the
production of IL-6 by whole blood, whereas one batch of artificially
contaminated albumins tested false-negative in the LAL. Another study using
the WBT performed with albumins, coagulation factor, vaccines and
immunoglobulins indicated a high sensitivity and reliability of the WBT for these
substances (19).
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
15
A comparison between the rabbit and the human whole blood test for the
detection of pyrogens in albumins was performed by Spreitzer et al. (64) with 29
batches of human serum albumin. The WBT was clearly superior to the rabbit
test, especially at the limit of detection of 5 EU/kg (0.5 EU/10ml/kg), with the
WBT retrieving all 29 spiked samples as positive compared to only 5 positive
rabbit tests and 23 temperature rises, which would have required a repetition of
the test. This limit of detection represents the 0.5 EU/ml pyrogenic threshold.
3.6. Validation
Six cell-based assays, including two variants of the WBT measuring IL-1 and IL-
6, respectively, were validated in an international collaborative study including
laboratories from Austria, Germany, Switzerland, England, Norway and Italy
and the participation of control institutions. The study validated assays such as
the cell line THP-1 with the endpoint TNF-α (65) or with the endpoint neopterin
(66, 67), the cell line Monomac-6 measuring IL-6 (68), isolated peripheral blood
mononuclear cells (PBMCs) with endpoint IL-6, and the human whole blood test
(49), using blinded endotoxin stimuli and altogether 13 intravenously applied
drugs. Sensitivities ranged between 73-96% and specificities between 90-97%.
The WBT measuring IL-1 achieved 73 and 93%, and the WBT measuring IL-6
88.9 and 96.6%, respectively. The development and outcome of this study is
described in detail elsewhere (48, 49, 69- 71)
3.7. Development of the commercially available IPT kit
The established WBT procedure was adapted to materials provided by Charles
River Endosafe and a commercial kit was developed, which was named In Vitro
Pyrogen Test (IPT). This kit contains all the reagents necessary for the
incubation and ELISA procedure except for the human whole blood.
Development of cryopreserved blood
Fresh human whole blood is a highly perishable item that cannot be stored
longer than 4 hours at room temperature without loss of sensitivity. Additionally,
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
16
it is not easily available, a potential hazard due to unrecognized infections (HIV,
hepatitis) and, due to donor differences, cannot be standardized. In order to
overcome these difficulties a procedure was developed to successfully freeze
and store whole blood. The protocol closely followed the method of de Boer,
1981, who had already successfully frozen isolated monocytes (72). Blood from
five healthy donors is mixed with 10% endotoxin-free dimethylsulfoxide (DMSO)
(v/v ratio) and left to stand for 15 minutes. The blood is then pooled (Fig. 1) and
frozen in a computer-controlled freezing process to –120°C. The blood is stored
in the vapor phase of liquid nitrogen and, after thawing, can be used like fresh
blood without any washing steps. The cryopreserved pooled blood renders
highly reproducible results and is at least equal to fresh blood concerning a
wide variety of applications and stimuli (73).
0.5 EU/ml
1.0 EU/ml
0
2500
5000
7500
0.5 EU/ml
1.0 EU/ml
saline control0.5 EU/ml1.0 EU/ml
1 2 3 4 5 6 7 8 9 10 pool 1 pool 2
donor number
IL-1
ββ ββ±± ±±
SD
Fig. 1: Comparison of the reactivity of frozen blood from 10 individual
donors and that of pooled blood from the same donors.
The calculated means of the response of the five individual donors towards the
0.5 EU/ml LPS corresponds to the response of the pooled blood. The higher
response of donor 4 is therefore leveled out.
Pool 1: The blood was pooled before freezing
Pool 2: The blood was pooled after thawing
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
17
Since blood frozen using the described method could only be stored and
shipped in the vapor phase of liquid nitrogen, a reagent that is not available to
all laboratories, an alternative freezing method was developed by the Paul-
Ehrlich Institute, Langen, Germany. The method is described in detail by
Schindler et al, 2006 (74). The alternative cryopreservation method provided
blood that could be stored (and shipped) at -80°C, therefore making the blood
available for users without liquid nitrogen infrastructure.
Validation of the cryopreserved blood
In an additional validation process, which followed the exact procedure of the
former process described above, both methods of cryopreservation were
validated (74). Furthermore, the IPT incubation steps, which had been developed
and validated in pyrogen-free reaction tubes, had in the meantime been
successfully transferred to the 96-well microtiter plate by reducing the volumes
used and adapting the protocol accordingly. Therefore, the fresh blood incubation
in the microtitre plate was validated as well as the cryopreserved blood both in
the 96-well microtiter plate and in the pyrogen-free reaction tubes. The overall
performance of all approaches was very good, with sensitivities of over 90% and
specificities around 80%. Remarkably, these excellent performance
characteristics were achieved although the spike concentrations chosen were at
or below the defined pyrogenicity threshold of 0.5 EU/ml (48). Indeed, the few
misclassifications only occurred for these borderline cases. Therefore, the IPT
could be improved concerning its availability, its performance and its handling
(Table I).
Test Inter-laboratory
reproducibility
(%)
Sample size:
sensitivity
Sensitivity
(%)
Sample size:
specificity
Specificity
(%)
WBT
Fresh blood
Reaction tubes
DL-NL1: 72.9
DL-NL2: 81.6
NL1-NL2: 70.2
88 72.7 59 93.2
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
18
IPT
Cryopreserved
blood (-80°C)
DL-NL 1: 86.7
DL-NL 2: 87.5
NL 1-NL 2: 100
77
97.4
45
82.2
IPT
Cryopreserved
blood
(nitrogen)
DL-NL 1: 66.0
DL-NL 2: 63.3
NL 1-NL 2: 83.3
74
82.4
46
89.1
IPT
Fresh blood
plate
DL-NL 1: 88.1
DL-NL 2: 89.7
NL 1-NL 2: 91.5
84
98.8
55
83.6
Table I: Outcome of the validation of the basic WBT procedure using
reaction tubes and fresh blood and of the IPT methods using
cryopreserved or fresh blood in a microtiter plate.
3.8. Special adaptations
Medical devices
Due to manufacturing and handling, medical devices can bear pyrogens on their
surface which, when brought into the human organism may lead to
inflammatory reactions and reduced biocompatibility. Recognizing this problem,
the Medical Device Directive 93/42 EEC states that medical devices must be
designed and manufactured in such a way that they will not compromise the
clinical condition or the safety of the patients. The Association for the
Advancement of Medical Instrumentation (AAMI) stated in 2001 that products
with direct or indirect contact with the circulatory system or the lymph or
products that interact systemically with the body should be tested for pyrogens
(75).
Products in direct (blood bags, needles) and indirect (swabs, gloves) contact to
the blood circulation can have a serious impact on the organism, as
contaminations induce systemic reactions. A severe case of contact dermatitis
due to endotoxin contamination of surgical gloves was described in 1984 by
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
19
Shmunes and Darby (76). After eight pyrogenic reactions in 69 patients
undergoing heart catheterization, Kure et al. described endotoxin contamination
of extracts of the hospital’s surgeon’s latex gloves, which evoked fever in
rabbits and could be successfully transmitted to cardiac catheters (77). Grötsch
et al. were able to evoke fever reactions in rabbits with an eluate of gloves
found to contain up to 2560 EU (78).
Medical devices pose a particular problem for pyrogen testing, since they
cannot be examined directly with the rabbit or the LAL test. Their diversity with
regard to size, form, material and form of application challenges the existing
assays, demanding individual approaches. In order to judge a possible
contamination, an eluate of the respective material must be either injected into
the rabbit or used in the LAL. However, it is unclear, how well rinsing a medical
device in water can release pyrogens from its surface and the dilution of such
released pyrogens in a large volume of rinsing water reduces the limit of
detection. The alternative of transplanting the questionable material directly into
the rabbit is highly invasive, causing possible reactions not associated with
pyrogenic contaminations but rather with tissue damage and is therefore
questionable in its ethical and scientific implications. The obvious advantage of
the IPT over the classical test methods is that the whole blood comes into direct
contact with the respective device and no preparation of an eluate is required.
This has been demonstrated using aneurysm clips as proof of principle (79).
Additionally, unlike the LAL, the IPT detects all pyrogens relevant to humans,
not only endotoxin.
Testing for the inflammation-inducing potential of implant surfaces for the
judgment of biocompatibility is a relatively new field. In the early 1980s, it was
noted that the monocyte is one of the first cells to arrive at an implant site and
displays manifold functions (for review see 80, 81). Its specific preference for
rough and hydrophobic surfaces differs from that of fibroblasts (82). The role of
cytokine production of the monocytes/macrophages in the early stages of
implant insertion is poorly understood. The fact that some materials are
obviously capable of modulating the cytokine response (83, 84) makes it difficult
to distinguish a genuine pyrogenic contamination from an unspecific activation
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
20
and poses the problem of adequate negative controls. For this purpose, a
process was developed for the thorough depyrogenation and a device was
developed for the testing of metallic or plastic surfaces with the IPT in order to
gain information about possible inherent activating or inhibiting characteristics of
materials (85). The device was made up of a perforated metal plate pressed
onto the sample surface by screws through a metal frame. The resulting wells
were watertight due to the use of washers. The blood was incubated directly in
the wells of the depyrogenized device contacting the surface to be tested. The
study showed that pyrogenic contaminations on surfaces could be reliably
removed only when heated for 5 h at 300° C. This applied to titanium, titanium
alloy (TiAl6V4) and steel material for implants. Artificial contaminations were
detected in a dose-dependent manner.
Some medical devices are absorbed completely by the body, as are any
contained pyrogenic contaminations. Examples are liposomes and alginate
microcapsules used as drug carriers. The detection of pyrogenic contaminations
in alginate microcapsules is illustrated in Fig. 2.
A B C D E 0 25 100 500 10000
2500
5000
7500
10000
12500
15000
17500
alginate samples LPS (E. coli O-113) (pg/ml)
IL-1
ß (
pg
/ml)
Fig. 2: IL-1β production of fresh blood upon stimulation with different
alginate solution samples (A-E).
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
21
Cellular therapeutics
Cellular therapeutics are defined as living cells that are transferred into the
intact organism in order to introduce a new or restitute a defective function (86).
This includes a wide variety of cells such as chondrocytes, stem cells, bone
marrow cells, and blood cells such as lymphocytes, erythrocytes, thrombocytes.
The latter pose a particular problem, since they are stored at room temperature
and are therefore easily subject to extensive bacterial growth. Transfusion
reactions may range from shivering, fever and chills all the way to septic shock.
The problem is rather under- than overrated, since numerous clinical events are
not recognized as being transfusion-associated, but are often rather attributed
to the underlying disease. Additionally, medication and immunosuppression
might mask an existing septic/pyrogenic event which likely contributes
significantly to the patient’s overall morbidity. Recently attention has focused on
viral infections, although the incidence of viral contaminations of blood products
is less than 1 in 1.000.000 per unit for HIV in comparison to 1 in 3000 for
bacterial contaminations (87).
Two large studies in France (BACTHEM study, 88) and the USA (BaCon study,
89) revealed that platelets hold a significantly higher risk of bacterial
contamination than red blood cells, irrespective of whether they were single-
donor or pooled preparations. Pathogens associated with bacteremia in the US
study were 59% Gram-positive (mainly skin contaminants such as
staphylococci, streptococci and propionibacteria) and 41 % Gram-negative (coli,
serratia, enterobacter). Gram-negative Yersinia enterocolitica was not found in
that study, although it occurs frequently in transfusion-related sepsis and was
responsible for 7 of the 8 fatalities recorded in the US between 1986 and 91.
Incidences of microbial contamination increased with prolonged storage, and
both studies linked fatalities to the occurrence of Gram-negative bacteria. The
US study also determined endotoxin levels (up to 273,500 EU/ml, according to
LAL). The authors estimated rates of transfusion-transmitted bacterial infections
of 1:100’000 for platelets and 1 in 5 million for red blood cells, with fatalities of 1
in 500.000 and 1 in 8 million, respectively. Overall, Gram-negative bacteria
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
22
tended to occur more frequently in red blood cells, probably due to the storage
at lower temperatures.
In 2004, a roundtable meeting on bacterial detection took place during the
Annual Congress of the International Society of Blood Transfusion in Edinburgh
to discuss the limitations of current testing methods. Currently, culturing
methods, such as the very sensitive BacT/ALERT method, are considered the
best, though they are very time-consuming (12 h to 7 days). Platelet
concentrates are released on a “negative-to-date” base and recalled if
necessary. The panel reported occurrences in the Netherlands, where platelet
concentrates containing skin bacteria were tested positive only after 48 h. By
this time, the batch had been released and about 50% of the units had already
been transfused. Very similar events were described by Belgian blood centers
(87). Anaerobic bacteria, e.g. Corynebacterium spp., are picked up even later
and there is extra cost involved. However, anaerobic bacteria have been linked
to fatal septic transfusion incidents (90). Altogether, culture methods are
incapable of providing complete safety, and other, especially quicker methods,
are sought.
A method to inactive contaminating bacteria in transfusion products by
photochemical treatment (PCT) (91) has been developed. Still, it must be kept
in mind that although this inactivation may inhibit growth, it will have no
influence on the already existing pyrogenic content. Therefore, the testing of
these cellular products and their suspension materials is an interesting future
challenge for the IPT. Pretesting of clinical grade erythrocytes and
thrombocytes intended for transfusion indicated interference-free retrieval of an
artificial endotoxin spike (Fig. 3) when compared to the saline control.
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
23
0 0,125 0,25 0,5 0,75 1 2 40.0
0.6
1.2
saline controlthrombocytes
erythrocytes
O113 (EU/ml)
OD
450
Fig. 3: Retrieval of endotoxin spikes in red blood cell and platelet
concentrates
Dialysis
Pyrogenic reactions in hemodialysis patients at the end of a session were first
associated with high bacterial and endotoxin levels by Raij et al. (92) and
Favero et al. (93). Since then, contaminations have been found in the pure
water (94-97), the machines, especially in areas with low circulation or dead
spaces which serve as a reservoir for bacteria (98), filter materials (99) and
bicarbonate concentrates (95).
In 1993, the AAMI released recommendations for the quality of treated water
and dialysate, which restricted the content of heterotrophic bacteria to 200 and
2000 cfu/ml, respectively. Studies in Germany (97), Greece (100), the USA
(94), and Canada (101) revealed that even these moderate standards are not
met, which is even more critical considering that a patient with chronic renal
failure receives up to 400 l of dialysis fluid a week. Next to Gram-negative
bacteria, cocci (micrococci, staphylococci and streptococci) were found in the
dialysate of 83, 70, and 10% of the centers, respectively, indicating the
importance of Gram-positive contaminations. That this might indeed be crucial
for judging the pyrogenic exposure of a dialysis patient was assessed by
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
24
Marion-Ferey et al., who tested scrapings of bacterial biofilms in dialysis tubes
and found a 20-fold higher response in the IPT than in the LAL (102). The
passage of cytokine-inducing substances, not only endotoxins, but exotoxins
and peptidoglycans as well, through the dialysis membrane has been
demonstrated (103- 106).
The chronic exposure of dialysis patients to even low concentrations of
pyrogens is thought to contribute to inflammatory processes in the joints and
bones and therefore to the carpal tunnel syndrome and arthropathy associated
with long-term hemodialysis (107, 108). In 1991, Baz et al. showed that the use
of ultrapure water delays the onset of the carpal tunnel syndrome (109). The
group of Schwalbe (110) showed in a retrospective study that the incidence of
amyloidosis decreased between 1988 and 1996 along with the introduction of
reverse osmosis, a very effective method for purifying water. A connection
between other phenomena, such as malnutrition, poor immune responses and
high incidence of malignant tumors in long-term HD patients with pyrogen
exposure has yet to be established.
In all, the testing for pyrogens in dialysis fluids is a crucial issue for the safety of
the patients. Since the fluids themselves are either highly hyper- or hypotonic, a
variant for testing dialysis fluids in the IPT established the percentages of
diluents and samples that can be tested (own unpublished results). Still, the
problem remains that the patients receive very high volumes of fluid in one
session, and therefore pyrogens must be detected at very low concentrations. A
promising possibility is a modification of the basic IPT protocol, the so-called
adsorb and wash IPT (AWIPT), discussed later, which can concentrate
pyrogens on the surface of albumin-coated macroporous Matisse™beads, thus
enhancing the sensitivity by a factor of 250 (111).
Airborne pyrogens
Inhalable whole or fragments of microorganisms have long been recognized as
causes of airway hyperreactivity. Monday sickness with its typical symptoms
(chest tightness, respiratory distress and coughing) was described as early as
1936 (112). In 1942, rural mattress makers experienced headache, nausea,
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
25
chills and fever about 6 hours after exposure to low grade cotton. Neal et al.
associated these phenomena with high numbers of a Gram-negative bacterium
in the material (113). Additionally, milder symptoms occurring 8 hours after
exposure could also be evoked by sterilized cotton, which was thought due to
remaining endotoxin. A highly significant (r > 0.95) dose-response relationship
between Gram-negative bacterial count and symptoms of byssinosis such as
chest tightness, airway inflammation and coughs was established by Cinkotai et
al. (114). In the same study, a good correlation existed between symptoms and
mainly Gram-positive bacteria, whereas one to fungal spores could not be
established. Acute bronchoconstriction as well as chronic airway irritation with
bronchitis and decrements in airflow over the work day have been reported for
personnel working in animal confinement buildings (115 - 117). Long term
consequences are of allergic, inflammatory and immunostimulatory nature, e.g.
organic dust toxic syndrome (ODTS) and chronic bronchitis. The LAL test for
these contaminants has the drawback that it does not reflect the biological
potency of a given LPS in the mammal (53) and the LAL test can only be
performed with an eluate of a filter or by impingement, i.e. the air to be tested is
led through pyrogen-free water which is then tested in the LAL. The higher
pyrogen retrieval by impingement when compared to filtration, possibly due to
the incomplete eluation of the sample from the filters, was demonstrated by
Zucker et al. (118).
A new approach of measuring the integral inflammatory activity in air samples in
different environments by IPT was reported by Kindinger et al., 2005 (119). A
defined amount of air is drawn through a filter in a sealable plastic monitor. The
blood incubation is performed directly on the filter inside the monitor, thus
making any handling of the filter unnecessary. When compared to the LAL, a 2-
25 fold higher pyrogenic load was found in the IPT in samples drawn in parallel.
Epidemiological studies will show what levels of exposure to inflammatory
stimuli in the air eventually lead to the above-mentioned lung diseases.
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
26
Lipidic formulations
In January 2004, pyrogen testing of so-called small volume parenterals (< 15ml)
became obligatory in Europe. This concerns many formulations that had not
been subjected to pyrogen testing before, such as vitamin preparations and
steroids. Many of these are applied intramuscularly or subcutaneously and are
therefore not necessarily of a hydrophilic nature. This poses a completely new
challenge to all methods of pyrogen testing, since a lipophilic substance on the
one hand cannot be injected intravenously into the rabbit due to the danger of
clogging small vessels with lipid drops and severely damaging vital organs and
will, on the other hand, influence the optical density measured as the readout in
the LAL due to the formation of oil droplets. Furthermore, the pyrogenic portion
of LPS, lipid A (for review see Rietschel et al., 1993 (120) has been reported to
be masked by lipoproteins (121) and lipophilic parenterals (122) in the LAL.
Therefore, the IPT procedure was adapted to suit lipophilic substances. As a
first step, interference-free oils, such as sesame oil, were identified by
comparing an LPS dose response curve in these oils with a similar curve done
in saline. Surprisingly, many oils (sesame oil, peanut oil, paraffin, miglyol) were
interference-free while others interfered strongly by suppressing the endotoxin
stimulus added. Oils that proved interference-free were then used as diluents
for interfering end-products. It was possible to dilute their interference to non-
detectable limits with full recovery of an artificial endotoxin spike. From this
minimum valid dilution a possibly detectable endotoxin concentration could be
calculated, which was 20 EU/ml for the respective end-products. Since these
products are applied at a very small volume (1 ml per person), a relatively high
endotoxin concentration can be tolerated. The established protocol leaves a
broad safety margin, especially since the strict criteria for intravenous drugs
were applied to this situation (123).
AWIPT (absorb and wash IPT)
Another interesting development is the so-called absorb and wash IPT
(AWIPT). It uses porous acrylic beads with immobilized albumin, which has a
higher affinity than native plasma albumin to endotoxin (124), to separate the
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
27
pyrogenic contamination from the sample. These beads were originally
developed as LPS adsorbers (Matissebeads™) to be applied in sepsis patients.
The AWIPT uses this material to collect the endotoxin contained in a sample
after an absorption phase in the substance to be tested. The beads are then
washed in order to remove the unbound material and can then be used directly
in the IPT incubation. It could be shown that this works also for LTA of
Staphylococcus aureus and zymosan, a yeast extract. This procedure has
already brought promising results with substances that interfere with the
standard IPT procedure, i.e. toxic or immunomodulatory drugs (111). Another
possibility is the testing of high-volume parenterals such as dialysis fluids, which
contain endotoxin concentrations below the detection limit of other pyrogen
tests. The beads could be used to concentrate the endotoxin on their surface
from a large sample volume. Using this procedure, the detection limit of the IPT
could be lowered from 0.25 EU/ml of E. coli endotoxin down to 1 x 10-5 EU/ml
(Fig. 4).
0.000 0.001 0.0100
50
100
150
200
250
300
350
400
O-113 LPS [EU/ml]
IL-1
ββ ββ [
pg
/ml]
Fig. 4: Limit of detection in the AWIPT
In all, the further development of the IPT into the modified form of the AWIPT
promises to overcome shortcomings for special applications caused by
interferences of certain drugs or substances with the classical IPT procedure. It
allows lowering of the detection limit, and provides a useful tool for the testing of
toxic or strongly interfering substances, even those that suppress the immune
system and therefore cytokine production.
DEVELOPMENT, VALIDATION AND APPLICATIONS OF THE IN VITRO
PYROGEN TEST (IPT) BASED ON HUMAN WHOLE BLOOD
28
3.9. Conclusions
Pyrogens (fever-inducing substances) from microorganisms can occur as
contaminations of parenterals. Until now, the safety of injectable drugs has
been assessed by the in vivo rabbit pyrogen test and the in vitro Limulus
amoebocyte lysate test (LAL).
The new cell-based in vitro pyrogen test based on fresh or cryopreserved
human whole blood (IPT) has been successfully validated and has proven to be
a reliable and useful tool for a wide spectrum of applications, ensuring patient
safety in many medical fields such as hydrophilic and lipophilic drugs, dialysis
fluids, airborne pyrogens, medical devices and biologicals. It is capable of
measuring all known pyrogens relevant for the human and is highly reliable,
robust and easy to perform.
3.10. Appendix
Abbreviations: AAMI, Association for the Advancement of Medical
Instrumentation; AWIPT, absorb and wash in vitro pyrogen test; cAMP, cyclic
adenosinmonophosphate; DMSO, dimethylsulfoxide; ECVAM, European Centre
for the Validation of Alternative Methods; ELISA, enzyme-linked immunosorbent
assay; ELC, endotoxin limit concentration; EU, endotoxin unit; HD,
hemodialysis; HSA, human serum albumin; IL, interleukin; IPT, in vitro pyrogen
test; LAL, Limulus amoebocyte lysate; LPS, lipopolysaccharide; LTA,
lipoteichoic acid; NIH, National Institutes of Health; NIBSC, National Institute of
Biological Standards and Controls; OD, optical density; OVLT, organum
vasculosum laminae terminalis; PBMCs, peripheral blood mononuclear cells;
PEI, Paul-Ehrlich Institute; PG, prostaglandin; POD, peroxidase; RNA,
ribonucleic acid; TMB, tetramethylbenzidine; TNF, tumor necrosis factor; USP,
United States Pharmacopoeia; WBT whole blood test
3.11. Acknowledgements
The validation study was funded by the European Union [QLRT-1999-00811].
The authors would like to thank the numerous scientific and industrial
supporters who provided sample materials.
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
29
4 International validation of novel pyrogen tests based on
human monocytoid cells
Sebastian Hoffmanna;h§; Anja Peterbauere*§; Stefanie Schindlera§; Stefan
Fennricha; Stephen Pooleb; Yogesh Mistryb; Thomas Montag-Lessingc; Ingo
Spreitzerc; Bettina Löschnerc; Mirjam van Aalderend; Rogier Bosd; Martin
Gommerd; Ria Nibbelingd; Gabriele Werner-Felmayere; Petra Loitzle; Thomas
Jungif; Marija Brcicf; Peter Brüggerg; Esther Freyg; Gerard Boweh; Juan
Casadoh; Sandra Coeckeh; Jan de Langeh; Bente Mogsteri; Lisbeth M. Næssi;
Ingeborg S. Aabergei; Albrecht Wendela; Thomas Hartunga;h#
a Institute of Biochemical Pharmacology and Steinbeis Center InPuT, University
of Konstanz, Universitätsstrasse 10, D-78457 Konstanz, Germany b NIBSC, National Institute for Biological Standards and Control, Blanche Lane,
South Mimms, Potters Bar, Herts EN6 3QG, England, UK c Paul Ehrlich Institute, Paul-Ehrlich Strasse 51-59, D-63225 Langen, Germany d RIVM, National Institute of Public Health and the Environment, A. van
Leeuwenhoeklaan 9, P.O.Box 1, 3720 BA Bilthoven, The Netherlands e Institute of Medical Chemistry and Biochemistry, Fritz-Pregl-Strasse 3, A-6020
Innsbruck, Austria f Institute of Veterinary Virology, Länggass-Strasse 122, University of Bern, CH-
3012 Bern, Switzerland g Biological Analytics, Novartis Pharma AG, CH-4002 Basel, Switzerland h European Centre for the Validation of Alternative Methods (ECVAM), Institute
for Health & Consumer Protection, European Commission Joint Research
Centre, Via Fermi 1, I-21020 Ispra, Italy i Division of Infectious Disease Control, Norwegian Institute of Public Health,
P.O. Box 4404 Nydalen, NO-0403 Oslo, Norway
* Present address: Ludwig Boltzmann Institute for Experimental and Clinical
Traumatology/Blood Transfusion Service for Upper Austria, Blumauerstr. 3-5,
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
30
A-4020 Linz, Austria § S. Hoffmann, A. Peterbauer, S. Schindler contributed equally to the work
presented here
4.1. Abstract
Parenteral medicines are required to be tested for pyrogens (fever-causing
agents) in one of two animal-based tests: the rabbit pyrogen test and the
bacterial endotoxin test. Understanding of the human fever reaction has led to
novel non-animal alternative tests based on in vitro activation of human
monocytoid cells in response to pyrogens. Using 13 prototypic drugs, clean or
contaminated with pyrogens, we have validated blindly six novel pyrogen tests
in ten laboratories. Compared with the rabbit test, the new tests have a lower
limit of detection and are more accurate as well as cost and time efficient. In
contrast to the bacterial endotoxin test, all tests are able to detect Gram-
positive pyrogens. The validation process showed that at least four of the tests
meet quality criteria for pyrogen detection. The here validated in vitro pyrogen
tests overcome several shortcomings of animal-based pyrogen tests. Our data
suggest that animal testing could be completely replaced by
these evidence-based pyrogen tests and highlight their potential to further
improve drug safety.
Keywords: Pyrogens; validation study; cytokines; monocytes; alternatives to
animals; cell culture
Abbreviations: BET, bacterial endotoxins test; CI, confidence interval; DL,
developing laboratory; ELC, endotoxin limit concentration; ELISA, enzyme-
# Corresponding author: Prof. Thomas Hartung, MD, PhD; European
Commission, Joint Research Centre; Institute for Health and Consumer
Protection; ECVAM; 21020 Ispra; Italy; Tel.: +39 0332 785939; Fax: +39 0332
786297; e-mail: [email protected]
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
31
linked immunosorbent assay; EU, endotoxin units; IFNγ, interferon γ; IL,
interleukin; LAL; Limulus amoebocyte lysate; LPS, lipopolysaccharide; LTA,
lipoteichoic acid; LoD, limit of detection; MM6, MONO MAC 6; MVD, maximum
valid dilution; NL, naive laboratory; PBMC, peripheral mononuclear blood cell;
PBS, phosphate buffered saline; TLR, toll-like receptor; TNFα, tumor necrosis
factor α; WBT, whole blood test
4.2. Introduction
Pyrogens, a chemically heterogeneous group of fever-inducing compounds, are
derived from bacteria, viruses, fungi or the host himself.
Monocytes/macrophages react to microbial products during an immune
response by producing endogenous pyrogens such as prostaglandins and the
pro-inflammatory cytokines interleukin-1 (IL-1), interleukin-6 (IL-6) and tumor
necrosis factor-α (TNFα) (132). Depending on the type and amount of pyrogen
challenge and the sensitivity of an individual, life-threatening shock-like
conditions can be provoked. Consequently, to assure the quality and safety of
any pharmaceutical product for parenteral application in humans, pyrogen
testing is mandatory.
Depending on the drug, one of two animal-based pyrogen tests is currently
prescribed by the health authorities and Pharmacopoeias, i.e., for more than
sixty years, the rabbit pyrogen test or the bacterial endotoxins test (BET), often
referred to as Limulus amebocyte lysate test (LAL). For the rabbit pyrogen test,
sterile test substances are injected intravenously into rabbits and any rise in
body temperature is measured. This in vivo test detects various pyrogens but
alone the fact that large numbers of animals are required to identify the rare
pyrogen-containing samples in routine practice argues against its use if valid
alternatives are available. In the past two decades, the declared intention to
refine, reduce and replace animal testing, the 3Rs concept (125) that was
implemented e.g. into European legislation in 1986 (126), has led to a reduction
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
32
in rabbit pyrogen testing by 80 % by allowing the BET as an in vitro alternative
pyrogen test for many parenteral products.
Bacterial endotoxin comprised largely of lipopolysaccharide (LPS) from the cell
wall of Gram-negative bacteria that stimulates monocytes/macrophages via
interaction with CD14 and toll-like receptor 4 (TLR4) (127) is the pyrogen of
major concern to the pharmaceutical industry due to its ubiquitous sources, its
stability and its high pyrogenicity (128-130). With the BET, endotoxin is
detected by its capacity to coagulate the amoebocyte lysate from the
haemolymph of the American horseshoe crab, Limulus polyphemus, or the
Japanese horseshoe crab, Tachypleus tridentatus, a principle recognized some
40 years ago (31). In the United States, Limulus crabs are generally released
into nature after drawing about 20 % of their blood and therefore most of these
animals survive. However, the procedure still causes mortality of about 30.000
horseshoe crabs per year, which adds to the even more severe threats of the
horseshoe crab population such as its use as bait for fisheries, habitat loss and
pollution (http://www.horseshoecrab.org). As with the rabbit test the general
problem of translation of the test results to the human fever reaction persists.
Moreover, although it is highly sensitive, the failure of the BET to detect non-
endotoxin pyrogens as well as its susceptibility to interference by, for example,
high protein levels of test substances or by glucans impedes full replacement of
the rabbit pyrogen test (53,131). Hence, an estimated 200.000 rabbits per year
are still used for pyrogen testing in the European Union.
A test system that combines the high sensitivity and in vitro performance of the
BET with the wide range of pyrogens detectable by the rabbit pyrogen test is
therefore required in order to close the current testing gap for pyrogens and to
avoid animal-based tests. With this intention and due to improved
understanding of the human fever reaction (132), test systems based on in vitro
activation of human monocytoid cells have been developed. First efforts
date back about 20 years, when peripheral blood mononuclear cells (PBMC)
were used to detect endotoxin by monitoring the release of pyrogenic
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
33
cytokines (133, 134). Subsequently, a number of different test systems, using
either whole blood, PBMC or the monocytoid cell lines MONO MAC 6 (MM6)
(68) or THP-1 (135) as a source for human monocytes and various read-outs
have been established and were recently reviewed (136). Here, the six most
prominent of these test systems were formally validated with the aim of
developing an evidence-based tool for safer, animal-free and more efficient
pyrogen detection and allowing their regulatory acceptance. Formal validation
of in vitro methods, i.e. the evaluation of reliability and relevance of a method,
was developed by the European Centre for the Validation of Advanced and
Alternative Methods (ECVAM) and is now internationally accepted (137-139).
4.3. Methods
Rabbit pyrogen test
For this study data from 171 rabbits (kindly provided by Dr. U. Lüderitz-Püchel)
accumulated over several years at the Paul Ehrlich Institute, the German
Federal Agency for Sera and Vaccines in Langen, were used for analysis. For
these experiments, Chinchilla Bastards (Charles River) were injected with 0, 5,
10, 15, 20 EU in 1 ml/kg of E. coli LPS (EC5) (140) or EC6 (56) in saline
(corresponding to 0, 0.5, 1.0, 1.5 and 2.0 EU/kg in 10 ml, the largest volume
allowed for injection in rabbits). The fever threshold in rabbits was defined as a
body temperature increase of 0.55 °C during 180 min after injection. This value
represents the mean individual rabbit value at the threshold of 6.6 °C of the EP
when the maximum of twelve animals is tested (141).
In vitro monocyte-based tests
Good laboratory practice concordant Standard Operating Procedures of the
various methods were made available by ECVAM (www.ecvam.jrc.it). The
test systems are summarized by Hartung et al. (70) and detailed in previous
work (49, 66, 67, 69, 142, 143).
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
34
Reagents and consumables for all methods
The 2nd International WHO Standard for endotoxin (from E. coli O113:H10:
K(-) (94/580), which is identical to FDA/USP standard EC6/Lot G was used as
the standard endotoxin (56). Test materials for validation are specified in the
Results section. All consumables were purchased as sterile and pyrogen-free
and not specified reagents were pro analysis grade.
PBMC-IL6
Blood Collection and preparation of PBMC
Blood donors had to describe themselves as being in good health, not suffering
from any bacterial or viral infections for at least one week prior to the donation
of blood and not to be taking drugs known to influence the production of
cytokines. Using a heparinized (50 µl Fragmin at 10000 IU, Dalteparin,
Pharmacia) syringe, 30 ml blood were collected. Within two hours, PBMCs
were isolated from 20 ml Lymphoprep (Nycomed, Oslo, Norway), 15 ml PBS
and 15 ml of heparinized whole blood by centrifuging at 340 x g for 45 min at
room temperature. The PBMC-layer was washed twice with PBS centrifuging at
340 x g for 15 min. The sediment was suspended with RPMI-C (RPMI 1640,
Life Technologies, Paisley, Scotland) with 10 ml/l human serum AB from
clotted human male whole blood (Sigma), 10 ml/l L-Glutamine
(Life Technologies), 200 mM, and 20 ml/l Penicillin/Streptomycin solution
(Seromed, Vienna, Austria)) after counting in a Neubauer haemocytometer to 1
mio cells/ml. The cells shall be incubated with samples within four hours after
blood withdrawal.
Protocol for PBMC-IL6
In quadruplicate per each of four blood donors, 100 µl of RPMI-C, 50 µl of
samples/controls and 100 µl of gently swirled PBMC were incubated in a 96-
well tissue culture plate (Falcon Microtest, Becton Dickinson Labware) at
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
35
37°C for 16- 24 hours in an atmosphere of 5% CO2 in humidified air. After
incubation, 50 µl of supernatant from each of the wells was transferred on the
ELISA plate ensuring that cells are not aspirated by angling the assay plate.
ELISA for PBMC-IL6
2.5 µg/ml coating mouse monoclonal anti-IL-6 antibody (Novartis in-house
Clone 16) was added at 200 µl to each well of a 96-well microtitre plate (Nunc-
Immuno 96-well plate MaxiSorp, F96; Life Technologies) at 15 - 25 °C for 16 -
24 hours. The washed plate was coated with 200 µl blocking buffer (24.2 g/l
Tris(hydroxymethyl)aminomethane, 0.2 ml/l Kathon MW/WT (Christ Chemie
AG, Reinach, Switzerland) and 10.0 g/l bovine serum albumine). Plates were
incubated with 200 µg/ml horseradish peroxidase conjugated to sheep anti-IL-6
antibodies (Novartis, in-house) for 2-3 hours at 20-25°C. Shortly before use, 90
ml substrate buffer and 4.5 ml TMB solution (240 mg 3,3',5,5'Tetramethyl-
benzidine in 5 ml acetone, 45 ml ethanol and 0.3 ml Perhydrol (30 % H2O2))
were mixed and 200 µl pipetted into each well. After 10-15 minutes, the
enzyme reaction was stopped by 50 µl of 5.4% H2SO4 per well. The
absorbance was measured at 450 nm using 540-590 nm as reference
wavelength.
WBT-IL1
Blood Collection for WBT-IL1
Blood donors should show no evidence of disease or need of medication during
the last two weeks. Blood was collected into heparinized tubes (Sarstedt S-
MONOVETTE 7.5 ml, 15 IU/ml Li-Heparin) and used within four hours (144).
Protocol for WBT-IL1
In this order and in quadruplicates per single blood donor, 1000 µl saline, 100
µl sample/control and 100 µl blood were added to pyrogen-free reaction tubes
(Greiner Bio-one tubes, 1.2 ml (polystyrene) or 1.5 ml (polypropylene),
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
36
Frickenhausen, Germany). Closed tubes were mixed gently, inverted once or
twice and then incubated in an incubator or a heating block at 37°C ± 1°C for
10-24 hours. The incubation tubes were mixed thoroughly by inverting them.
Incubations were centrifuged for 2 minutes at 10.000 g and the clear
supernatant, taking aliquots of ≥ 150 µl, was used for the ELISA (ENDOSAFE-
IPT, Charles-River Endosafe, Charleston, USA) following the manufacturer’s
procedure.
WBT-IL6
Blood Collection for WBT-IL6
Blood donors were selected as described for PBMC-IL6. 30 ml blood were
drawn and immediately transferred into a 50 ml sterile centrifuge tube
containing 300 IU heparin (Fragmin, Pharmacia, diluted 1/10 with saline). The
closed tubes were inverted slowly five times to ensure thorough mixing without
vortexing and used within four hours (174).
Protocol for WBT-IL6
In quadruplicate per each of four blood donors, 50 µl of saline, 50 µl of gently
mixed blood, 50 µl of samples/controls and 100 µl of saline were incubated in a
96-well tissue culture plate (Falcon Microtest, Becton Dickinson Labware) at
37°C for 16-24 hours in a humid atmosphere of 5% CO2. After incubation, 50 µl
of supernatant from each of the wells was transferred on the ELISA plate
ensuring that cells are not aspirated by angling the assay plate. The
same IL-6 ELISA as for PBMC-IL6 was used.
MM6-IL6
Cell culture for MM6-IL6
The human monocytoid cell line MonoMac-6 was obtained from Prof. H.W.L.
Ziegler-Heitbrock (Institute for Immunology, University of Munich, Munich,
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
37
Germany). Frozen cells from liquid nitrogen were thawed on ice. Cells were
transferred to a 50 ml centrifuge tube, 10 ml RPMI (+4°C) (e.g.
Life Technologies) added and then centrifuged at 100 x g for 5 min at +4°C.
Afterwards the cells were resuspended in 10 ml RPMI-M (containing 10% ml
heat-inactivated low-pyrogen foetal calf serum, 2 mM L-Glutamine, 0.1 mM
MEM non-essential amino acid, 0.23 IU/ml Bovine insulin, 1 mM Oxaloacetic
acid, 1 mM Sodium pyruvate, 20 mM HEPES). After a wash step, cells were
transferred to a 25 cm2 tissue culture flask and incubated at 37°C, with 5% CO2
and high humidity. The number of viable cells was determined by Trypan blue
exclusion using a haematocytometer. The cells were passaged with 2 x 105
cells/ml twice a week.
Protocol for MM6-IL6
To pre-incubate the cells for a test, 30-50 ml of cell suspension were
centrifuged at 100 x g for 8 min at room temperature and resuspended in
RPMI-C (as RPMI-M, but only 2% heat-inactivated foetal calf serum) at a final
concentration of 4x105 cells/ml. The cells were incubated approximately 24
hours at 37°C, 5% CO2 and high humidity. Cells were washed and counted as
above, diluting to 2.5 x 106 viable cells/ml, just prior to addition to the culture
plate. In quadruplicates, 50 µl of samples/controls, 100 µl of RPMI-C and 100 µl
of gently swirled MM6 were incubated in 96-wells tissue culture plates at 37°C
for 16-24 hours with 5% CO2 and humidified air. After incubation, 50 µl of
supernatant from each of the wells was transferred on the
ELISA plate ensuring that cells are not aspirated by angling the assay plate.
The same IL-6 ELISA as for PBMC-IL6 was used.
THP-Neo
Cell culture for THP-Neo
THP-1 cells were obtained from the American Type Culture Collection (ATCC,
TIB-202). 6 x 106 cells were seeded in 60 ml medium (RPMI 1640
supplemented with 10 % (v/v) FCS (high-quality lots with the lowest endotoxin
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
38
content available (< 30 pg/ml) were chosen, e.g. Biochrom, Berlin, Germany) in
75 cm2 culture flasks. Flasks were incubated in upright position at 37° C with
5% CO2 and humidified air. On the fourth day of culture, further 30 to 60 ml
(depending on the culture doubling time) of culture medium were added and
cells were incubated for another three days. If cells from freshly thawed stocks
are used, they have to be grown for two to three weeks in order to ensure that
they divide properly before using them for tests. Furthermore, cells should not
be kept in culture for more than four months but new cultures should be started
from frozen stocks at regular intervals. Cells were counted with a
hemocytometer and cell viability by trypan blue exclusion was ≥ 90%. Tubes
with 2.5 x 107 cells (for one plate) were centrifuged at 400 x g and 20° C for 7
min and resuspended in 20 ml medium, 2 mM L-glutamine and 50 µM 2-
mercaptoethanol.
Protocol for THP-Neo
100 µl IFNγ (human, recombinant, endotoxin content < 0.1 EU/mg;
Gammaferon 50, Rentschler Biotechnologie, Laupheim, Germany) stock
solution (6250 U in 100 µl medium, 110 µl aliquots) were added to 20 ml of cell
suspension and mixed well. 200 µl/well of mixed cell suspension were added to
a 96-well cell culture microtiter plate. After incubation for 30 min, 50 µl of
vortexed samples/controls (in quadruplicate) were added and put on an orbital
plate shaker for 2 min at room temperature and 500 rpm. After 18-22
hours of incubation, 150 µl of supernatant were collected and frozen and/or
directly processed with the neopterin ELISA (Elitest Screening, Brahms
Diagnostica, Berlin, Germany) according to the manufacturer’s protocol.
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
39
THP-TNF
Protocol for THP-TNF
THP-1 cells (obtained from ATCC or ECACC) were used. Subclones from this
cell line prepared in-house showed a higher sensitivity towards LPS.
Cells were cultured in RPMI (1% L-glutamine, 1% HEPES, 1%
Penicillin/streptomycin solution, 1% Sodium pyruvate, all from Biochrom (Berlin,
Germany), 1% nonessential aminoacids for MEM, 0.4% MEM vitamin solution,
0.5% β-mercaproethanol (10 mM), all from Invitrogen (Basle, Switzerland), and
12% heat-inactivated low-pyrogen FCS in 6-well plates or T25 flasks at 37°C in
a humidified 5% CO2 incubator. They were passaged once weekly. When new
cells are required for an assay, cells from a cryovial were thawed two to three
weeks before use. For the last passage prior to the test, terminal differentiation
was induced by cultivating the cells in the presence of sterile-filtered calcitriol
(1,25-dihydroxy vitamin D3, Sigma or Hoffmann-La Roche, Basle, Switzerland)
(10 µg/ml) for 44-48 hours. Cells were collected, centrifuged and resuspended
in culture medium containing calcitriol (final concentration 100 ng/ml). They
were counted and adjusted to 1 to 1.25x106 cells/ml. Cells were cultured for 44-
48 hours in T25 flasks. Then, terminally differentiated cells were harvested and
counted using a haematocytometer and trypan blue. Cells were diluted to
1.25x106 cells/ml and 200 µl of suspension were dispensed into each well of
the above 96-well cell culture plate containing already 50 µl of sample/control in
quadruplicates. Plates were incubated for 16-24 hours at 37°C and 5%CO2.
TNFα ELISA for THP-TNF
Non-sterile plates Dynex PF microtiter ‘flat bottom’ styrene 96-well plates
(Dynex Tech., Worthing, UK) were rinsed extensively with pyrogen-free PBS.
The plates were coated with 1 µg/ml monoclonal antibody 101-4 against human
TNFα (a generous gift from Dr. T Meager, Division of Immunobiology, NIBSC,
UK) at 100 µl/well and 4 ºC overnight. 50 µl of sample/control (in
quadruplicates) or duplicates of TNFα standards (250, 62.5, 15.6, 3.9, 0.98,
0.24, 0 U/ml, NIBSC) were added for 16-24 hours at 37°C and 5% CO2. An
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
40
aliquot of the detecting antibody (biotinylated goat-anti-human TNF-α from the
Duoset kit, R&D) was diluted 180-fold, using dilution buffer (0.1 % bovine
serum albumin, 0.1% Tween 20, in 20 mM Tris, 100 mM NaCl, pH 7.2-7.4). 100
µl were dispensed to each well for two hours at room temperatue. After
washing, 100 µl Streptavidin-peroxidase conjugate (R&D) was added for 20
min. After washing, 100 µl of TMB (Sigma) were dispensed and incubated in
the dark before reading at 650 nm. Incubation time was chosen so that 250
U/ml TNFα value had an OD > 1.5.
Data analysis
The rabbit fever reaction was modeled by regression techniques applied to the
logarithmically transformed data. The within- and between-laboratory
reproducibility were assessed comparing the resulting classifications by means
of simple matching, i.e. the proportions of identically classified samples, as a
measure of similarity. In case of the within-laboratory reproducibility, where
three independent but identical runs were performed, the mean similarity was
calculated.
A one-sided t-test, assuming hazard and thus designed to proof safety of a
tested compound, was employed as a so-called prediction model (PM) to
dichotomize the test results into a classification of either ‘pyrogenic’ or ‘non-
pyrogenic’. The t-test compares the data of a given sample against the data of
the standard positive control of 0.5 EU/ml, which is performed in parallel. It is
calculated with the log-transformed data and a local significance level of 1%
was chosen in order to increase safety. If this test resulted in a significant
p-value, i.e. smaller than 1 %, then the considered sample was classified as
non-pyrogenic, and as pyrogenic otherwise. This means that a negative sample
had to be significantly lower than 0.5 EU/ml. The levels of contaminations
chosen were 0, 0.25, 0.5 (twice) and 1 EU/ml. According to the rabbit model, 0
and 0.25 EU/ml were considered as non-pyrogenic samples and 0.5 and 1
EU/ml as pyrogenic samples. Having thus defined the reference standard, i.e.
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
41
the ‘true’ contamination level, we calculated via 2x2-contingency tables the
performance parameters sensitivity, i.e. the probability of a correct positive
classification, and specificity, i.e. the probability of a correct negative
classification. Confidence intervals for these parameters were calculated with
the Clopper and Pearson method based on the F distribution (145).
4.4. Results
The limit of endotoxin detection in rabbits
Employing regression techniques, the temperature data from 171 rabbits could
be modeled by the equation y = 0.217 * (EU + 1)0.508, where y is the expected
temperature increase for a given concentration EU/ml (Fig. 1). This approach
was recently described in more detail and further exploited (16). The model
indicated that 50 % of the animals develop fever, i.e. showing a 0.55 °C rise of
body temperature within 180 min after injection, in response to 5.22 EU per kg
body weight of endotoxin with a 95 %-confidence interval of 4.24 to 6.21 EU/ml.
Only at 20 EU per kg of body weight, all animals showed an increase in
temperature of 0.55 °C or more. We deduced from these data that a sample
concentration of 0.5 EU/ml represents the required limit of detection (LoD) that
alternative pyrogen tests must meet. This assumption takes into account the
fact that the largest volume allowed for injection into rabbits is 10 ml per kg,
corresponding to 0.5 EU/ml for injections at 10 ml/kg.
Thus, the concentration of 0.5 EU/ml was defined as the threshold between
pyrogenic and non-pyrogenic samples.
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
42
0 5 10 15 200.0
0.5
1.0
1.5
2.0
endotoxin units/ml per kg bodyweight
tem
pera
ture
incr
ease
[°C
]
Fig. 1. Temperature increase of 171 rabbits upon endotoxin injection with
a fitted regression line
The maximum temperature increase in ºC within 180 minutes after endotoxin
injection of 171 rabbits is presented. The mean temperature increase, modeled
with regression techniques, is indicated by the dotted line.
Prevalidation of the novel in vitro pyrogen tests
Before prevalidation, the test-developing laboratories that took part in the study
compiled standard operating procedures for the alternative tests. This required
an intensive phase of test optimization and standardization in order to allow the
transfer of the tests. A standard curve of endotoxin in saline including the 0.5
EU/ml concentration as the threshold for pyrogenicity was included in all tests.
Only if the 0.5 EU/ml endotoxin standard was detectable, did the test run qualify
for analysis. Before prevalidation was started, the naive laboratories proved
evidence of successful transfer of the respective test systems (data not shown).
Prevalidation was then carried out with twelve
blinded samples. These consisted of three drugs spiked with either pyrogen-
free saline (clinical grade 0.9 % NaCl) or with reference endotoxin. Two
negative, i.e. pyrogen-free samples, and two LPS-containing, i.e. pyrogenic
samples (0.5 EU/ml and 1.0 EU/ml sample concentration, respectively) were
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
43
tested. The concentration of 0.5 EU/ml was the limit of detection defined for the
rabbit pyrogen test (see above). The drugs used were Gelafundin, a volume-
replacement therapy for transfusion with high protein (gelatine) content (B.
Braun Melsungen AG, Melsungen, Germany), Jonosteril, an electrolyte
infusion (Fresenius AG, Bad Homburg, Germany) and Haemate, a factor VIII
preparation (Aventis Behring GmbH, Marburg, Germany). In addition, a positive
control (0.5 EU/ml LPS in saline) and a negative control (endotoxin-free saline)
were included. Each test was performed three times in the respective
developing laboratory (DL) as well as in two naive laboratories (NL).
Test
System
Readout
Ref.
Within-
laboratory
reproduci-
bility (%)
Between-
laboratory
reproduci-
bility (%)
Sensi-
tivity
(%)
Speci-
ficity
(%)
WBT-
IL6
whole
blood
IL-6
136
DL: 83.3
NL1: 94.4
NL2: 100
DL-NL1: 72.2
DL-NL2: 72.2
NL1-NL2: 96.3
72.2
92.6
WBT-
IL1
whole
blood
IL-1β
49
DL: 88. 9
NL1: 95.8
NL2: 94.4
DL-NL1: 91.7
DL-NL2: 76.8
NL1-NL2: 67.8
72.0
100.0
PBMC-
IL6
PBMC
IL-6
143
DL: 94.4
NL1: 100
NL2: 94.4
DL-NL1: 80.6
DL-NL2: 86.1
NL1-NL2: 88.9
87.0
98.1
MM6-
IL6
MM6
(68)
IL-6
136
DL: 100
NL1: 94.4
NL2: 94.4
DL-NL1: 97.2
DL-NL2: 88.9
NL1-NL2: 86.1
72.2
100.0
THP-
TNF
THP-1
clone
TNFα
65
DL: 94.4
NL1: 83.3
NL2: 55.5
DL-NL1: 90.7
DL-NL2: 67.6
NL1-NL2: 65.7
66.7
88.9
THP-
Neo
THP-1
parental
(135)
neo-
pterin
65
DL: 100
NL1: 94.4
NL2: 77.7
DL-NL1: 97.2
DL-NL2: 50.0
NL1-NL2: 51.8
88.9
72.2
Table 1: Novel pyrogen tests and their performance in prevalidation
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
44
Protocols for all methods are listed in Poole et al. (136) and in the Methods
section. All tests include dilution of the sample by 1:5 with the exception of the
IPT-IL-1 test that requires a 1:12 dilution of the sample. The WBT-IL6 and the
PBMC-IL6 tests combine data from three respectively four blood-donors per
run, the WBT-IL1 from one donor per run. Samples and controls were tested in
quadruplicate in each of the tests. DL denotes developing laboratory, NL1 and
NL2 the two naive laboratories. The sample size analyzed for sensitivity and
specificity was 108 for all tests besides WBT-IL1 (100 samples). Sensitivity
describes the probability to correctly classify positive samples and specificity
describes the probability to correctly classify negative samples.
Table 1 summarizes the six novel test systems used, their major
characteristics, their performance regarding reproducibility, which was
assessed before the blinding code was broken, as well as sensitivity and
specificity. As can be seen, the predictive capabilities of the various tests were
encouraging, particularly in the light of the restricted stability of endotoxin
spikes at the borderline concentration of 0.5 EU/ml. Although all tests were
successfully transferred to the naive laboratories during the preparatory phase
of prevalidation, this optimal performance could not be maintained for the two
test systems using THP-1 cells, as is reflected by the comparatively low
between-laboratory reproducibility between the developing laboratory and one
of the naive laboratories for each. The lower specificity of the THP-Neo test
was entirely caused by misclassification in NL2. Furthermore, prevalidation also
revealed that, despite preceding interference testing and diluting of the drugs
accordingly, interference/recovery problems persisted in some cases, as is
reflected by the values for sensitivity.
Validation phase
For the validation phase,10 drugs with five blinded spikes each (0 (i.e. pyrogen-
free), 0.25, 0.5 (twice) and 1 EU/ml) were tested, again in three laboratories,
i.e. the DL of a test and the two NLs, respectively. To avoid the possibility that
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
45
different dilutions of the drugs were tested depending on their different
interference with different test systems, all drugs were tested at their maximum
valid dilution (MVD), thus adopting the rationale of the pharmacopoeial BET
reference (limit) test. The MVD is calculated from the endotoxin limit
concentration (ELC in EU/ml) defined for a drug by the European
Pharmacopoeia (146), divided by the threshold of pyrogenicity as the limit of
detection (LoD), i.e. 0.5 EU/ml. Drugs, sources, ELCs and MVDs (= ELCs/LoD,
where LoD=0.5) are summarized in Table 2.
Drug Source Agent Indication ELC
(EU/ml)
MVD
(-fold)
Glucose 5
% (w/v)
Eifelfango
GmbH
glucose nutrition 35 70
Ethanol
13 % (w/v)
B.Braun AG ethanol diluent 17.5 35
MCP Hexal AG metoclo-
pramid
antiemetic 175 350
Orasthin Aventis
Pharma GmbH
oxytocin initiation of
delivery
350 700
Binotal
Aventis
Pharma GmbH
ampicillin antibiotic 70 140
Fenistil
Novartis
Consumer
Health GmbH
dimetinden-
maleat
antiallergic 87.5 175
Sostril
GlaxoSmithKli
ne GmbH
ranitidine antiacidic 70 140
Beloc
Astra Zeneca
GmbH
metoprolol
tartrate
heart
dysfunction
70 140
Drug A* 0.9 % NaCl 17.5 35
Drug B* 0.9 % NaCl 35 70
Table 2: Test substances for the validation phase
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
46
* Drugs were selected by a selection committee which excluded the developing
laboratories and included experts. Drugs A and B which were saline only were
included as further controls using notional ELCs.
Drugs were obtained from Eifelfango GmbH (Bad Neuenahr-Ahrweiler,
Germany), B. Braun AG (Melsungen, Germany), Hexal AG (Holzkirchen,
Germany), Aventis GmbH (Bad Soden, Germany), Novartis GmbH (München,
Germany), GlaxoSmithKline GmbH (München, Germany) and
Astra Zeneca GmbH (Wedel, Germany). ELCs of drugs were calculated
according to European Pharmacopoeia (146).
While the tests using whole blood, PBMC and MM6 cells performed well in all
three test laboratories in terms of reproducibility (Table 3), technical problems
with the two tests using THP-1 cells were obvious. For the THP-TNF test this
was caused by a batch of TNFα-ELISA plates sent out to the two NLs that did
not satisfy the quality criteria with regard to detection limit when used with cells.
For the THP-Neo test, the technical problems in NL2 persisted such that the
quality criteria defined in the SOP were not met. The tests could not be
repeated due to the limited time frame of validation and for logistical reasons.
Therefore, for the THP-TNF assay only the data from the DL and for the THP-
Neo assay only the data from the DL and from NL1 could be analyzed.
Sensitivity and specificity were 76.7 % and 78.9 % for the THP-TNF assay
(sample size = 40) and 93.3 % and 47.5 % for the THP-Neo assay (sample size
= 100). The data for the other four tests are summarized in Table 3. Almost all
misclassifications, either false negatives or false positives, occurred around or
at the defined classification threshold, i.e. for the contaminations of 0.25 and
0.5 EU/ml. Confidence intervals (CI) with a significance level of 5 % were
calculated for sensitivity and specificity. By focusing on the lower bounds of CI
(Fig. 2), a worst-case scenario can be conducted by which the likelihood of
underestimation of pyrogen content is maximized and thus possible negative
consequences for health can be estimated.
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
47
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0 0.1 0.2 0.3 0.4 0.5
1-specificity
sen
sit
ivit
y
MM6-IL6
PBMC-IL6
WBT-IL6
WBT-IL1
rabbit
Fig. 2. Sensitivity and specificity of four in vitro assays in the validation
study and modeled rabbit test performance with 95%-confidence intervals
The sensitivity and specificity resulting from the pre-defined prediction model
and considering samples with 0 and 0.25 EU/ml as non-pyrogenic and with 0.5
and 1 EU/ml as pyrogenic are presented with their corresponding 95 %
confidence intervals for four validated tests. Similarly, the respective
parameters were calculated with the rabbit model. As performance improves
towards the upper left of the graph, all validated tests outperform the rabbit test.
The lower predictive capability of the WBT-IL1 test as compared to the WBT-
IL6 and the PBMC-IL6 test can be explained by the one-donor approach used
for the WBT-IL1 test and the multiple-donor approach used for the other tests
that is more conservative and laborious, but decreases the probability for false-
negative classification. For the THP-TNF assay, the lower bounds of CI for
sensitivity and specificity were 60.6 % and 55.2 %, respectively. For the THP-
Neo assay, the respective lower bounds were 78.0 % and 38.7 %.
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
48
Applying this kind of analysis to the in vivo assay employing the regression
model based on the data from the rabbit pyrogen test yields a sensitivity of 57.8
% and a specificity of 88.3 % (Table 3) with confidence intervals also presented
in Fig. 2. Thus, the novel pyrogen tests listed in Table 3 show parameters of
performance outperforming the rabbit pyrogen test.
Test Between-
laboratory
reproducibility
Sample
size:
sensitivity#
Sensitivity
(%)
Sample
size:
specificity
Specificity
(%)
WBT-IL6
DL-NL1: 85.4
DL-NL2: 85.4
NL1-NL2:92.0
89 88.9 59 96.6
WBT-IL1
DL-NL1: 72.9
DL-NL2: 81.6
NL1-NL2:70.2
88 72.7 59 93.2
PBMC-IL6
DL-NL1: 84.0
DL-NL2: 86.0
NL1-NL2: 90.0
90 92.2 60 95.0
MM6-IL6
DL-NL1: 90.0
DL-NL2: 89.6
NL1-NL2: 83.3
89 95.5 59 89.8
Rabbit† - - 57.9 - 88.3
Table 3: Validation of the predictive capability of novel pyrogen tests
# sample sizes are reduced by outlier exclusion defined in the study protocol † parameters calculated by the fitted regression model
An additional analysis, which could be conducted with the available data,
supports this conclusion. According to their SOPs, the four systems included an
uncontaminated negative control, i.e. saline, and another positive control of 1
EU/ml. For each of these two controls we adapted the prediction model
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
49
described above: First, we compared the blinded samples against the
response of 1 EU/ml control. Therefore, we constructed a modified prediction
model using the 1 EU/ml control response instead of the positive control of 0.5
EU/ml, which is denoted in the following by PM1. In doing so, the true
classification of the samples changed, as now only the samples spiked with 1
EU/ml were considered as pyrogenic and the other samples as non-pyrogenic.
Second, with a modified prediction model, denoted as PM0, classifying a
sample as pyrogenic when the response was significantly larger than the
negative control response (significance level 1 %), we compared all spikes
against this control. Again, the true classification of the samples needed to be
adjusted considering the contaminated samples (0.25, 0.5, 1.0 EU/ml) as
pyrogenic and the unspiked samples as non-pyrogenic. The resulting
sensitivities and specificities are summarized together with the results from the
original PM for the four test systems in Fig. 3. All tests performed best for PM0,
where the sum of these two parameters was at least 1.90, while WBT-NI even
resulted in the maximum sum of 2.
0.95 0.96 0.97 0.980.92
0.87
1.000.89
0.69
0.94
0.73 0.72
0.58
0.900.98 0.97
0.95 0.98
1.00
0.97
0.99
0.96
0.93 0.97
0.88
1.00
0
1
2
PM0 PM PM1 PM0 PM PM1 PM0 PM PM1 PM0 PM PM1
MM6-IL6 PBMC-IL6 WBT-IL6 WBT-IL1 rabbit
su
m o
f sen
sit
ivit
y a
nd
sp
ecif
icit
y
specificity
sensitivity
Fig. 3. Sum of sensitivity and specificity resulting from three prediction
models for four in vitro assays in the validation study
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
50
The validation data of four tests were analysed with three prediction model
employing different controls for comparison and thus defining the true
classification of the samples (non-pyrogenic vs. pyrogenic) accordingly. The
test accuracy is described for each test and prediction model by the sum of
specificity and sensitivity allowing also for individual parameter assessment.
For comparison, the rabbit test performance according to the pre-defined
prediction model is added. The DLs also tested lipoteichoic acid (LTA) from
Bacillus subtilis, a BET-negative Gram-positive compound that activates
cytokine release from human monocytes (26, 57) prepared according to Morath
et al. (57), which was clearly detectable by the novel tests.
4.5. Discussion
Previous work (49, 65, 67, 70, 136, 143, 147, 150, 151) had established that
different sources of human monocytoid cells are valuable tools for mimicing the
human fever reaction in vitro. Not only can these cells detect the important
pyrogen LPS from E. coli and other Gram-negative bacteria but also a number
of compounds involved in the immune response to Gram-positive bacteria such
as LTA (58), exotoxins (67, 148), cell wall components like muramyl dipeptide
(148) or peptidoglycan (149), S. aureus Cowan (SAC) (147) or DNA (67) as
well as poly (I:C) (147), a synthetic double-stranded RNA used as a virus model
compound in fever research. It was also established that these novel test
systems overcome limitations of the BET and yield results comparable to the
rabbit pyrogen test (64, 67, 147, 148,). For the first time, six of these
monocytoid-cell based in vitro pyrogen tests were formally validated in the
present study. For this purpose, a harmonized analysis procedure was
established that allowed the direct comparison of the different tests and
incorporated various safety aspects. A conservative statistical approach
showed that four test systems met the criteria for safe detection of pyrogens.
The two test systems based on the use of THP-1 cells posed problems in
performance. These were related to insufficient transfer to one naive laboratory
(THP-Neo) and to use of an ELISA batch for the one-plate assay format (THP-
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
51
TNF) that, although qualifying for the detection of TNFα did not qualify for the
use with cells and caused their prestimulation. Both problems became obvious
only during validation and could not be overcome within the tight schedule of
validation. Thus, for these two systems additional validation processes would
be required. However, the data obtained for the other four test systems clearly
suggest that these have reached a stage of development that makes them
suitable for use in pyrogen testing as replacements for the rabbit pyrogen test.
For the purpose of this study, a threshold value of 0.5 EU/ml was chosen on the
basis of historical data from rabbit tests carried out in a national control
authority. This approach was conservative as only 50% of animals of the very
sensitive strain used showed a febrile reaction at this concentration.
Additionally, in order to be classified negative, the samples had to be,
according to the PM, significantly lower than 0.5 EU/ml. On the one hand, the
enormous challenge to the models by placing two samples at the threshold of
0.5 EU/ml, which had to be classified positive, resulted in reduced sensitivities.
On the other hand including a sample with 0.25 EU/ml, which had to be
identified as negative, was the reason for almost all false-positive classifications
resulting in the reduced specificities without representing any safety concern.
When tested against the negative control (PM0), i.e. when samples which are
not significantly different from the negative control were considered as pyrogen-
free, the tests performed even better, i.e. with increased sensitivity. However,
this approach increases consumers’ safety on the cost of rejecting drugs,
whose minor pyrogenic contamination would not induce adverse health effects
in humans. At the same time, this reflects the fact that the study design put
main emphasis on the threshold of 0.5 EU/ml. Similarly, decreased sensitivity
when given the task of identifying 1 EU/ml as threshold value shows that the
tests were especially designed for the threshold of 0.5 EU/ml. Since the test
performance when changing the threshold is still acceptable or even better, the
robustness of the alternative tests is underlined.
In summary, this study provides thus evidence of the validity of these tests and
should facilitate the regulatory acceptance of these novel tests and lead to their
introduction into Pharmacopoeias.
INTERNATIONAL VALIDATION OF NOVEL PYROGEN TESTS BASED ON
HUMAN MONOCYTOID CELLS
52
Conflict of interest
S. Poole is named as an inventor in Patent Number US 6,696,261 B2 , Feb 24,
2004: 'Pyrogenicity test for use with automated immunoassay systems'.
T. Hartung and A. Wendel are named as inventors in Patent Number US
5,891,728 , Apr 6, 1999: 'Test for determining pyrogenic effect of a material'.
4.6. Acknowledgements
We thank U. Lüderitz-Püchel from the Paul Ehrlich Institute, Langen,
Germany, for providing rabbit pyrogen test data.
This work was supported by the European Union [QLRT-1999-00811].
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
53
5 Cryopreservation of human whole blood for pyrogenici-
ty testing
Stefanie Schindler, Silvia Asmus, Sonja von Aulock, Albrecht Wendel,
Thomas Hartung*, Stefan Fennrich
Biochemical Pharmacology, University of Konstanz, D-78457 Konstanz,
Germany
Corresponding address:
Thomas Hartung, PhD, MD
University of Konstanz
D-78457 Konstanz
Tel: +49-7531-88-4116
Fax: +49-7531-88-4117
e-mail: [email protected]
Abbreviations:
DMSO, dimethylsulfoxide; ELC, endotoxin limit concentration; EU, endotoxin
equivalent units; IL-1β, interleukin-1β; LPS, lipopolysaccharide, endotoxin;
LTA, lipoteichoic acid; MVD, maximal valid dilution; RT, room temperature;
WHO, World Health Organization
5.1. Abstract
Human whole blood assays are increasingly employed to test immune functions
or detect pyrogenic contaminations, since they offer advantages such as easy
performance, few preparation artifacts and physiological cell environment. The
approach, however, is often limited by availability of freshly drawn blood,
putative safety concerns in case of infected donors and interindividual donor
differences. To overcome these limitations, a method was developed and
optimized to produce batches of cryopreserved blood that can be used directly
after thawing without any washing steps. Mononuclear cells remained intact as
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
54
shown by FACS analysis. Cytokine release could be induced by a variety of
immunological stimuli. The cell preparation released higher amounts of IL-1β
and IL-6 compared to fresh blood, which could be attributed to the presence of
the cryoprotectant DMSO. Large batches of cryopreserved blood could be
produced by mixing blood donations of up to ten donors, independent of
differing blood groups. The detection limit for the WHO LPS reference
preparation (EC-6) with regard to induction of IL-1β release was at least 0.5
EU/ml. Endotoxin spikes at the limit concentrations prescribed by European
Pharmacopoeia could be detected in a series of drugs, showing that the In vitro
Pyrogen Test (IPT) can also be run with cryopreserved blood. Further possible
applications include high-throughput screening for immunomodulators or toxins
as well as preservation of patient samples for later analysis of cell functions.
Key words: Cryopreservation, Blood, Endotoxin, Interleukin-1β, In vitro
Pyrogen Test (IPT)
5.2. Introduction
Cryopreservation of cells represents a standard procedure in cell culture.
Human primary leukocytes are cryopreserved on a routine basis, for example to
store human bone marrow cells (151). Further cryopreservation protocols have
been established for various isolated blood cell populations including
lymphocytes and mononuclear cells and the retention of various cell functions
after thawing has been investigated (72, 152-155).
Although it is popular to isolate the respective immune cells from blood, it is
evident that such isolated cells do not reflect the in vivo situation: the cells are
often stimulated during the isolation procedure as indicated by basal mediator
release or adherence of the cells, interaction between different cell types
cannot take place and plasma components that often play an important role in
immune recognition are no longer present.
Methods employing whole blood have been developed to detect pyrogenic
(fever-inducing) contaminations, e.g. of batches of injectable drugs (49). This
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
55
application has been successfully validated in a collaborative European study
and awaits incorporation into the pharmacopoeias (71). Further, we have
suggested the study of cytokine, histamine or eicosanoid release with this
method to allow the characterization of putative drugs or immunotoxins (156,
157). These methods can also be used ex vivo on treated volunteers or
patients to monitor the course and effects of treatment (158-160).
Many of these procedures could be simplified or optimized by the availability of
cryopreserved whole blood. The blood could be supplied in the form of a
standardized test reagent which could be stored until needed and be certified
free of infectious agents. A method to preserve and store cells from treated
patients might allow performance of the often laborious cell assays on a series
of collected samples in parallel or at a distant laboratory, thus reducing
variability and logistical problems.
We sought to develop a protocol which would allow the use of the thawed
whole blood samples directly without any washing steps to remove the
cryoprotectant, as such a step would eliminate essential advantages of the
human whole blood assay, i.e. the ease of performance which allows a high
degree of standardization as shown for various applications (161). Furthermore,
beside stress and handling artifacts, the cells would lose their autologous
plasma that enables a number of physiological responses, e.g. the sensitive
response to lipopolysaccharides (endotoxin, LPS) via lipopolysaccharide
binding protein (LBP) (162, 163).
In this report we describe the development of a protocol to freeze human whole
blood and demonstrate retention of sensitivity and functionality regarding
stimulation of cytokine release in response to inflammatory agents.
5.3. Materials and methods
Freezing procedure
Blood was drawn from healthy volunteers into tubes containing 15 IU/ml Li-
Heparin (Sarstedt, Nürnbrecht, Germany) and differential blood cell counts
were performed on each sample to rule out active infections (Pentra 60, ABX
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
56
Diagnostics, Montpellier, France). In order to rule out pyrogenic contaminations
of any component used in the incubations, negative saline controls were
included in each experiment The heparinized blood was pre-cooled in ice water
for 15 minutes. Clinical grade dimethylsulfoxide (DMSO, Waco Chemicals,
Dessau-Thornau, Germany) was added to the blood in 50 ml centrifugation
tubes (Greiner bio-one, Frickenhausen, Germany) in small amounts to a final
concentration of 10% (vol/vol ratio) under constant gentle agitation to avoid cell
damage. Pooling was performed in 50 ml centrifugation tubes after the addition
of DMSO to the blood of the individual donors. Blood was pipetted as 1, 3 or 4
ml aliquots into pre-cooled cryotubes (1.8, 3.6 or 4.5 ml, Nunc, Wiesbaden,
Germany) and put into the rack of a programmable freezer with a TP type
nitrogen container (Nicool Plus PC, Air Liquide, Marne-la-Vallée Cedex 3,
France), pre-cooled to 4°C. A temperature probe was inserted into an extra
aliquot containing the same volume of blood to follow the freezing process. The
freezing program was started 5 min after closing the freezer. The blood was
cooled down to – 5°C at a rate of 1°C/minute. In order to compensate the latent
fusion heat generated by the blood when changing from the liquid to the solid
state, the temperature Tx in the freezing chamber was set to – 30°C. The
crystallization temperature was –12°C. When this temperature was reached,
the blood was cooled down to – 40°C at a rate of 2°C/min. The blood was given
120 seconds to stabilize before being cooled down to –120°C at a rate of
10°C/minute. After freezing, the tubes were removed from the freezer and put
immediately into the vapor phase of liquid nitrogen (nitrogen tank, Air Liquide,
Kryotechnik, Düsseldorf, Germany).
Thawing procedure
The closed tubes were left in an incubator at 37°C until completely thawed.
The aliquots of single donors were either pooled or the blood was pipetted
individually from each aliquot. Pooling of the blood of different donors could be
performed after thawing as an alternative to the procedure described above.
The whole blood incubation was started not more than 30 minutes after
complete thawing.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
57
Whole blood incubation
Human whole blood incubations were performed according to the protocol of
the in vitro Pyrogen Test (IPT; 49, 71). Briefly, 100 µl of fresh or cryopreserved
human blood were added to 1 ml physiological, clinical grade, pyrogen-free
saline in polypropylene reaction tubes (Eppendorf, Hamburg, Germany). After
the addition of stimuli, the tubes were closed, shaken gently and incubated
overnight (16-24 hours) at 37°C. The cells were resuspended and assayed
immediately or frozen until measurement. Within each experiment performed,
all samples were incubated and measured in parallel. When all samples of an
experiment were measured on the same ELISA plate, absorbance (A 450) was
given as the unit of measurement according to the IPT protocol. When the
samples could not be measured on the same ELISA plate, a recombinant
standard curve was run on each ELISA plate to allow interpolate comparison.
Endotoxin stimuli were LPS from Escherichia coli O113 (WHO standard
material), kindly provided by Dr. Stephen Poole, NIBSC, Hertfordshire, GB, or
LPS from E. coli O111 (IPT Kit, Charles River Endosafe, Charleston, SC, USA)
calibrated to the WHO standard material. One important criterion for the In Vitro
Pyrogen Test (IPT) is the ability to reproducibly detect the presence of 0.5 EU
(endotoxin equivalent units) per ml, equivalent to 50 pg/ml of the WHO
reference endotoxin standard or to 100 pg/ml of the LPS from E. coli O111,
respectively, in a sample solution, this being the fever threshold of the most
sensitive rabbit strain if applied at a dose of 10 ml/kg. Therefore, this LPS
concentration was included in every assay.
Non-endotoxin stimuli were lipoteichoic acid (LTA) from Bacillus subtilis (IPT
Kit, Charles River Endosafe) (49), glucan standard (Charles River Endosafe),
glucan from barley (Sigma, Munich, Germany), lectin from Phaseolus vulgaris
(PHA-L and PHA-E, Sigma), curdlan (Sigma) and zymosan A (Fluka, Buchs,
Switzerland).
Substances tested at MVD were furosemid (Lasix®), ampicillin (Binotal®),
Articain/Epinephrin (Ultracain®) (Aventis, Germany), Theophyllin
(Bronchoparat®), (Fujisawa, Munich, Germany), dimethindenmaleat
(Fenistil®) (Novartis, Munich, Germany), ranitidin (Sostril®) (Glaxo Smith
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
58
Kline, Munich, Germany) and metoprolotartrat (Beloc®) (Astra Zeneca, Wedel,
Germany).
Cytokine ELISAs were based on commercially available antibody pairs against
IL-1β or TNFα (Endogen, Biozol, Eching, Germany), and IL-6 (R&D,
Wiesbaden, Germany). Binding of biotinylated antibody was quantified using
streptavidin-peroxidase (Biosource, Camarillo, CA, USA) and the substrate
TMB (3,3’,5,5’-tetramethylbenzidine, Sigma). Recombinant cytokines serving as
standards were gifts from Dr. S. Poole, NIBSC.
FACS Analysis
25 µl of fresh or cryopreserved blood was stained with 5 µl each of anti-CD45-
APC and anti-CD14-FITC antibodies (BD Biosciences, Heidelberg, Germany)
for 30 min at room temperature in the dark. 1 ml Cell Wash and propidium
iodide in a final concentration of 500ng/ml were added directly, immediately
before measurement in a FACSCalibur (all BD Biosciences). A live gate was
set on CD45-positive cells and 3000 leukocytes were counted. Whole blood
counts were determined by Türks staining and counting in a Neubauer
chamber.
Statistics
Statistics were performed with GraphPad InStat 3.0 (GraphPad Software, San
Diego, USA). Significance was tested by one-way ANOVA and Dunnetts post-
test/Dunn´s multiple comparison and with t-test, followed by Mann-Whitney
post-test.
5.4. Results
Freezing procedure
Different concentrations of the cryoprotectant DMSO were tested to determine
a concentration that would protect the cells and leave them functional after
thawing but which would be sufficiently low to have no toxic effects in the
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
59
incubation (Fig. 1). While 1% DMSO was insufficient to protect the cells during
freezing and 20% DMSO had toxic effects in the subsequent incubation period,
cell preparations frozen with either 5 or 10% DMSO responded to stimulation
with LPS in a concentration-dependent manner after thawing.
0.0
0.5
1.0
1.5con0.5 EU/ml1.0 EU/ml
1% 5% 10% 20%DMSO (v/v)
ns
*
OD
[450n
m]
±± ±± S
D
Fig. 1. Optimization of the final DMSO concentration in cryopreserved
blood.
Blood was frozen with different concentrations of DMSO as shown and
stimulated with LPS from E. coli O113 after thawing (representative experiment
of 2). Blood from one donor in 4 replicates is shown, *, p<0.05 (one-way-
ANOVA, post test: Dunn´s multiple comparison). Incubation supernatants were
measured by ELISA technique and the endpoint IL-1β was
given as OD.
We compared whether the reactivity of the cryopreserved blood measured as
IL-1β response to endotoxin stimulation was affected by the blood temperature
(room temperature or 4°C) at which the DMSO was added and whether DMSO
should be added as a bolus or in several aliquots. The addition of DMSO at
room temperature seemed to cause an increase in reactivity rather than a
decline and addition of DMSO in several aliquots was preferable to the bolus.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
60
The mean OD of the aliquots with DMSO added at room temperature and
stimulated with 0.5 EU/ml O113 was 0.596 OD (SD 0.049) when the DMSO
was added as a bolus and 0.728 OD (SD 0.051) when the DMSO was added in
three aliquots. The cv was 8.17 and 7.02%, respectively. When the DMSO was
added at 4°C, the response was 0.404 OD (SD 0.056) when the DMSO was
added as a bolus and 0.547 (SD 0.034) when added in three aliquots (cv 13.82
and 6.29%).
Next, we determined how long the blood could be kept after addition of DMSO
before freezing and whether room temperature or 4°C is preferable. For this
purpose, DMSO was added to the blood and an aliquot was frozen immediately
while other aliquots were stored at room temperature or at 4°C for up to 200
minutes and then frozen and tested in parallel (Fig. 2). These data suggest that
storage at room temperature for up to 2 hours is tolerable and that storage at
4°C is beneficial when the blood is stored for longer.
0
2500
5000
7500
1.0 EU, RT
frozen immediately (25min)
110min 200min
storage time of blood + DMSO before freezing
0.5 EU, RT 0.5 EU, 4°C
1.0 EU, 4°C
IL1
ββ ββ [
pg
/ml]
±± ±± S
D
Fig. 2: Comparison of different storage temperatures and durations
before freezing.
Blood with 10% DMSO was frozen immediately or stored as indicated before
freezing, then stimulated with LPS from E. coli O111 after thawing
(representative experiment of two). Control values were < 6 pg/ml IL-1β for
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
61
each condition. Four replicates of blood from one donor are shown (0.5 EU/ml:
n=5).
Comparison of different volumes
Different volumes of blood (1 ml, 3 ml and 4 ml aliquots) were frozen and
stimulated with endotoxin after thawing. The reactivity of the blood did not
depend on the volume of the frozen aliquots. The mean OD of the 1, 3, and 4
ml aliquots when stimulated with 1.0 EU/ml of O113 was 2.824 (SD 0.066),
2.463 (0.058) and 2.6 (0.087) OD, respectively, with a coefficient of variation of
2.32 , 2.35 and 3.35%.
Thawing procedure
A thawing protocol was developed in order to optimize the handling of the blood
aiming at maximum reactivity and viability. Aliquots of blood from the same
donors were thawed under different conditions, i.e. on ice, at room temperature
(20°C) and in an incubator (37°C) until completely thawed before stimulation
with endotoxin (Fig. 3). Quick thawing at 37°C resulted in the best response.
0.0
0.1
0.2
0.3
0.4
0.5
4°C RT (20°C) 37°C
***
***
OD
[450n
m]
±± ±± S
D
Fig. 3: Determination of a suitable thawing temperature.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
62
Cryopreserved blood with 10% DMSO was thawed at different temperatures
(4°C, room temperature, 37°C), then stimulated with 0.5 EU/ml (hatched bars)
and 1.0 EU/ml (black bars) LPS from E. coli O113 or saline (white bars)
(representative experiment of 4). Blood from one donor in 8 replicates is shown.
***, p<0.001 against the respective stimulation after thawing at 4°C or 20°C
(one-way ANOVA with Dunnett’s post-test). Incubation supernatants were
measured by ELISA technique and the endpoint IL-1β was given as OD.
An important issue was the potential cytotoxicity of the remaining
cryoprotectant DMSO after thawing and before dilution with saline. Therefore,
we tested how long the thawed blood samples could be kept at 37°C before
dilution and stimulation (Fig. 4). The reactivity of the blood towards the
endotoxin stimulation decreased after 45 minutes of thawing time. Therefore,
the blood was used within 30 minutes after thawing at 37°C in all subsequent
experiments.
Fig. 4: Effect of time between thawing of blood and incubation.
Blood was thawed and stored at 37°C for the times indicated, then stimulated
with 0.5 EU/ml LPS from E. coli O113 (representative experiment of 4). Blood
from one donor in five replicates is shown. ** = p<0.01 vs. the values at 15 min
15 30 45 600.0
0.1
0.2
0.3
0.4
0.5
control
0.5 EU/ml
****
ns
time between thawing and incubation [min]
OD
[450n
m]
±± ±± S
D
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
63
(one-way ANOVA with Dunnetts post test). Incubation supernatants were
measured by ELISA technique and the endpoint IL-1β was given as OD.
FACS Analysis
Differential blood cell counts were done in parallel samples of fresh and
cryopreserved blood from 5 donors. Although the whole blood cell counts of the
cryopreserved blood did not differ from those of the fresh blood samples,
the differential blood cell count revealed that the neutrophilic granulocytes had
lost their surface markers and could no longer be identified as live, CD45
positive cells. The ratio of monocytes to lymphocytes in the differential blood
cell count was the same in the fresh and the cryopreserved blood (1 : 6.7 ± 0.9
versus 1 : 8.1 ± 1.6, n.s.) with a viability of these two populations of 99 vs. 90%
as shown by propidium iodide exclusion.
Establishment of a pooling protocol
Blood samples from five different donors were compared with each other and
with pools of the blood from the same donors combined either directly after
addition of the DMSO or after thawing of frozen blood. Establishing a pooling
protocol with blood from different donors with different blood groups proved
easier than anticipated (Fig. 5). There was no difference in the reactivity of the
blood pools, whether they were made before or after freezing. Also, the
reaction of the pooled blood was equal to the mean of the reaction of the
individual donors.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
64
0.5 EU
1.0 EU
0.0
0.5
1.0
1.5
0.5 EU
1.0 EU
1 2 3 4 5 pool 1 pool 2
OD
[450n
m]
±± ±± S
D
Fig. 5: Comparison of the reactivity of frozen blood from 5 donors and
their pooled blood.
Blood from five separate donors as well as a pool of their blood was frozen and
stimulated as shown (representative experiment of 4). The horizontal lines
indicate the calculated mean of the blood from the five donors to 0.5 or 1.0
EU/ml LPS from E. coli O113. Three replicates of all samples were measured
(0.5 EU/ml: 4 replicates). Pool 1, the blood of the single donors was pooled
after addition of DMSO; pool 2, the blood was pooled after thawing. Incubation
supernatants were measured by ELISA technique and
the endpoint IL-1β was given as OD.
Interlot variability
5 different pools of cryopreserved blood, each containing the blood of 5 donors,
were compared (Fig. 6). The interlot variability was very low, indicating that the
use of 5 donors in the pooling protocol is sufficient for producing highly similar
batches of blood.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
65
0.0
0.1
0.2
0.3
0.4
0.5
0.6
batch 142 batch 156 batch 161 batch 162
controlO-113; 0,25 EU/mlO-113; 0,5 EU/mlO-113; 1,0 EU/mlO-113; 2,0 EU/mln.s.
n.s.n.s.
OD
[450n
m]
±± ±± S
D
Fig. 6: Interlot variability of five different pools
Cryopreserved pools each consisting of five different donors and frozen over
a period of 23 weeks were thawed and stimulated on the same day (0.25
EU/ml, 2 EU/ml: 2 replicates each; control, 0.5 EU/ml, 1 EU/ml: 4 replicates
each) with LPS from E. coli O113. p > 0.05 of the 1.0 EU/ml value of batch 142-
161 against the 1.0 EU/ml value batch 162 (one-way ANOVA, Dunnett´s post-
test). Incubation supernatants were measured by ELISA technique and the
endpoint IL-1β was given as OD.
Stability
Numerous aliquots of a pool of blood from 5 donors were frozen and their
reactivity tested on different days over a period of 4 months. The IL-1β
response to 0.5 EU/ml endotoxin was significantly different from the saline
controls at each of the time points tested, indicating that the cryopreserved
blood remained stable over this time period and did not lose sensitivity
(Table I).
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
66
Day after
freezing
Mean OD
saline control
Mean OD
0.5 EU/ml
Reactivity
(% saline control)
0 0.045 ± 0 0.311 ± 0.06 691
40 0.068 ± 0.01 0.298 ± 0.02 439
118 0.077 ± 0.01 0.755 ± 0.05 980
Table I: Stability of pooled cryopreserved blood from 5 donors over a
period of four months.
Blood was stimulated with 0.5 EU/ml LPS from E. coli O113, n=4. Incubation
supernatants were measured by ELISA technique and the endpoint IL-1β was
given as OD.
To determine inter-aliquot variability of aliquots from the same blood donor,
eight replicates each from 3 aliquots of 1 ml thawed cryopreserved blood
were stimulated with 0.5 EU/ml endotoxin and eight replicates each from three
aliquots were left unstimulated (Table II). The stimulated samples had mean
ODs of 0.27 – 0.49 and the coefficient of variation (cv) was 12.3 – 26.1%, while
the unstimulated samples had mean values ranging from 0.047 to 0.054 OD
and a cv of 5.4 – 42.2%.
The inter-aliquot variability of the same experiment was 0.051 ± 0.004 (cv
7.3%) for unstimulated versus 0.37 ± 0.11 (cv 31.3%) for blood stimulated with
0.5 EU/ml LPS.
aliquot 1
saline
aliquot 2
saline
aliquot 3
saline
aliquot 4
0.5 EU/ml
aliquot 5
0.5 EU/ml
aliquot 6
0.5 EU/ml
Minimum 0.040 0.043 0.048 0.242 0.216 0.384
Median 0.046 0.047 0.053 0.259 0.317 0.495
Maximum 0.107 0.051 0.066 0.339 0.490 0.582
Mean 0.053 0.047 0.054 0.274 0.327 0.493
SD 0.022 0.003 0.006 0.036 0.085 0.061
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
67
SEM 0.008 0.001 0.002 0.012 0.030 0.021
cv (%) 42.17 5.4 10.66 12.83 26.08 12.26
Table II: Intra-aliquot variability of cryopreserved blood from one donor.
8 replicates from each aliquot of blood of the same donor were stimulated with
LPS from E. coli O113. Incubation supernatants were measured by ELISA
technique and the endpoint IL-1β was given as OD.
Comparison of the reactivity of cryopreserved with fresh whole blood
The reactivity of the cryopreserved blood to endotoxin stimulation was
compared to that of fresh blood of the same individual donors. As can be seen
in Figure 7, 0.5 EU/ml LPS induced significant IL-1β release both in the fresh
and the cryopreserved blood of every donor. This is the sensitivity limit
of the most sensitive rabbit strain for testing according to the European
Pharmacopoeia for injectable drugs.
0.0
0.1
0.2
0.3
0.4
0.5
0.6fresh blood, saline
fresh blood, 0.5 EU/ml
cryopreserved blood, saline
cryopreserved blood, 0.5 EU/ml
donor 1 donor 2 donor 3 donor 4 donor 5
*
***
*
***
****
* ***
*
***
OD
[450n
m]
±± ±± S
D
Fig. 7: Comparison of the reactivity of fresh and frozen blood of 5
separate donors.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
68
Fresh blood (4 replicates each) and thawed cryopreserved blood (6 replicates
each, saline control n=8) from the same five donors was stimulated with LPS
from E. coli O113. Cryopreserved blood was thawed immediately after
complete freezing and was incubated in parallel with the fresh blood of the
same donors (representative experiment of 3).
*, p<0.05, *** p<0.001, against the respective saline control (t-test and Mann-
Whitney post-test). Incubation supernatants were measured by ELISA
technique and the endpoint IL-1β was given as OD.
When the response to endotoxin stimulation of cryopreserved and fresh blood
from the same donors was compared in a kinetic study, a noticeable difference
in the kinetics of the LPS-inducible IL-1β release was observed (Figure 8, upper
panel). Measurable IL-1β release occurred with several hours delay in
cryopreserved compared to fresh blood. This could be attributed to the
presence of the cryoprotectant, since fresh blood containing 10% DMSO
showed the same delay. Furthermore, in both cases, the presence of DMSO
increased the maximum amount of IL-1β released 7-fold (fresh blood plus
DMSO) and 5-fold (cryopreserved blood). The same held true for IL-6 (Fig. 8,
center panel), though here the amount of IL-6 was increased nearly 20-fold.
TNFα release was no longer detectable after addition of DMSO (Fig. 8, lower
panel), both in fresh and cryopreserved blood.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
69
0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 40
1 0 0 0
2 0 0 0
f r e s h b lo o d
f r e s h b lo o d + D M S O
c r y o p r e s e r ve d b lo o d
2 5 0 0
5 0 0 0
7 5 0 0
1 0 0 0 0
in c u b a t io n t im e [ h ]
IL-1
ββ ββ [
pg
/ml]
±± ±± S
D
0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 40
1 0 0 0
2 0 0 0
3 0 0 0
4 0 0 0
5 0 0 0
1 0 0 0 02 0 0 0 03 0 0 0 04 0 0 0 05 0 0 0 06 0 0 0 0
in c u b a t io n t im e [ h ]
IL-6
[p
g/m
l]±± ±±
SD
0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 40
5 0 0
1 0 0 0
1 5 0 0
in c u b a t io n t im e [ h ]
TN
F [
pg
/ml]
±± ±± S
D
Fig. 8: Kinetics of cytokine response of fresh blood, fresh blood after
addition of 10% DMSO and cryopreserved blood.
Three replicates of blood samples pooled from five donors were challenged
with 1.0 EU/ml LPS from E. coli O111 for the times indicated.
Next, the reactivity of cryopreserved blood to a variety of immune stimuli was
tested in comparison to fresh blood. Different pyrogenic (fever-inducing) stimuli
including LPS, LTA, and phytohaemagglutinin-L (PHA-L) induced IL-1β release
in cryopreserved blood, but not the non-pyrogenic substances PHA-E, glucans
and monophosphoryl lipid A (data not shown). Differences were seen for
curdlan and zymosan A and, in terms of a higher sensitivity of the
cryopreserved blood, for endotoxin from Pseudomonas aeruginosa. Taken
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
70
together, the cryopreservation procedure did not alter the spectrum of pyrogens
or immune stimuli detected and did not cause the cryopreserved blood to react
to substances which fresh blood does not react to either.
To test whether cryopreservation might interfere with the detection of
contaminations in pharmaceuticals, e.g. as a result of hemolysis, the
freezing/thawing procedure or the DMSO, a series of drugs was tested with
fresh and cryopreserved blood as to their interference with a given LPS spike.
Results are summarized in Table III.
Trade
name
Drug ELC
(EU/ml)
MVD MID
cryoblood
MID
fresh blood
Lasix furosemid-sodium 15 30 1 :10 1 :30
Ultracain articain/epinephrine 75 150 1:150 Not test.
Binotal ampicillin-sodium 75 150 1:50 1:100
Broncho
-parat
theophylline 37.5 75 1:25 1:75
Fenistil dimethindenmaleate 93.75 187,5 1:150 1:180
Sostril ranitidine 75 150 1:20 1:100
Beloc metoprolotartrate 75 150 1:40 1 :50
Table III: Interference testing of clinical-grade parenterals in fresh and
cryopreserved blood.
Drug samples were added to fresh blood of 2 donors (n=4) or cryopreserved
blood of one donor (n=4) and spiked with 0.5 EU/ml LPS from E. coli O113.
For positive spike retrieval, 50-200% of the response to the LPS spike in
saline had to be found in the spiked drug sample. ELC, endotoxin limit
concentration according to European Pharmacopoeia; MVD, maximum valid
dilution (ELC/0.5 EU/ml); MID minimum interference dilution. Incubation
supernatants were measured by ELISA technique and the endpoint IL-1β was
given as OD.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
71
The interferences differed, though surprisingly, the cryopreserved blood
proved to be less prone to interference than the fresh blood: the minimal
interference dilutions were lower for the cryopreserved blood and always at or
below the maximal valid dilution (MVD). These data show that a broad variety
of drugs can be controlled by a pyrogen test based on cryopreserved blood,
maintaining the endotoxin limit concentration (ELC) according to the
pharmacopoeias established for the Limulus amoebocyte lysate assay
without that test’s restriction to Gram-negative LPS. However, interference
testing must be performed in every case for both fresh and cryopreserved
blood for any given drug.
5.5. Discussion
The utility of human whole blood assays has been demonstrated in a broad
variety of applications. All applications so far have to be carried out within a
few hours after blood withdrawal, which makes parallel processing difficult,
leading to a high variability in clinical samples, complicating donor pre-testing
and posing problems with regard to continuous availability of fresh blood
samples.
The procedure described here offers a continuous supply with a highly
homogenous batch of blood. A regular blood donation (500 ml) would suffice
for 5000 tests (tube format) or 25.000 tests (microtiter plate format), which
can be increased even further by pooling blood from several donors before
freezing. Additionally, blood batches can be pre-tested with regard to
sensitivity, and infections such as HIV or hepatitis can be excluded. The
latter, however, can also be achieved by prescreening donors following the
standard guidelines for blood donation.
The established procedure has been optimized with regard to the retention of
sensitive cytokine response to pyrogenic contamination. Both the freezing
and the thawing protocol were optimized and the reactivity of the
cryopreserved blood was compared to that of fresh blood. The inter- and
intra-aliquot variabilities were tested, as was the reaction of cryopreserved
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
72
blood to different pyrogenic and non-pyrogenic substances in comparison to
the reaction of fresh blood.
The differential blood cell counts of fresh versus cryopreserved blood showed
that monocytes and lymphocytes are still viable after the cryopreservation
procedure, thus implying that functional assays of these cells can still be
performed efficiently. Also the ratio of monocytes to lymphocytes remained
unchanged. However, the surface properties of neutrophilic granulocytes
were affected by the cryopreservation procedure, suggesting that neutrophil
function may be lost. The established pooling procedure allows the
preparation of large batches of cryopreserved blood and also reduces the risk
of possible abnormal individual reactions.
Apart from other application possibilities, the results presented indicate that
cryopreserved blood can be used as an alternative to fresh blood in the In
vitro Pyrogen Test to detect contaminations in batches of different drugs.
These first data show that a broad variety of drugs could be controlled using
the cryopreserved blood, maintaining the endotoxin limit concentration (ELC)
established for the Limulus amebocyte lysate assay without that test’s
restriction to Gram-negative LPS.
However, the hemolysis, dead PMN as well as the DMSO might result in
interferences with some drugs, e.g. drugs that bind to hemoglobin. It is also
possible that the synergy of LPS and hemoglobin influences the results.
Therefore, the suitability of cryopreserved blood for pyrogen testing of a given
product will have to be demonstrated by separate interference testing.
This approach may also find application in high-throughput screening. It is
tempting to base screening assays on homogenous preparations of human
primary cells, requiring no prior culture or isolation procedures. Given the
broad variety of immunomodulators in clinical use, several application
opportunities can be imagined. Since the whole blood model also allows the
determination of eicosanoid release (157), this might extend to modulators of
eicosanoid formation such as NSAID (non-steroidal anti-inflammatory drugs).
However, the feasibility and relevance of this approach will have to be
established.
CRYOPRESERVATION OF HUMAN WHOLE BLOOD FOR PYROGENICITY
TESTING
73
Taken together, a variety of immunological models and tests might benefit
from the availability of functional, cryopreserved blood. Efforts to make this
available on a large scale are ongoing. The advantages and respective
adaptations for different uses will have to be established. The approach
promises, however, to overcome problems of availability and standardization
of human primary blood leukocytes and to provide standardized blood as an
immunological reagent for a broad spectrum of applications.
5. 6. Acknowledgements
This work was supported by the Bundesministerium für Bildung und Forschung
(BMBF 11425A) and the Stiftung zur Förderung der Erforschung von Ersatz-
und Ergänzungsmethoden zur Einschränkung von Tierversuchen (set). The
procedure has been granted a European patent (97 122 072.8).
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
74
6 International validation of pyrogen tests based on
cryopreserved human primary blood cells
Authors: Stefanie Schindler a, Ingo Spreitzer b Bettina Löschner b, Sebastian
Hoffmann c, Kilian Hennes d, Marlies Halder c, Peter Brügger e, Esther Frey e,
Thomas Hartung c , Thomas Montag-Lessing b
a = Biochemical Pharmacology, University of Konstanz, Universitätsstr.10, D-
78457 Konstanz, Germany
b = Paul-Ehrlich Institute, Paul-Ehrlich-Strasse 51-59, D-63225 Langen,
Germany
c = European Centre for the Validation of Alternative Methods, Institute for
Health and Consumer Protection, European Commision, Joint Research
Centre, Via Fermi 1, I-21020 Ispra, Italy
d = Qualis Laboratories, Blarerstr. 56, D-78462 Konstanz (www.qualis-
laboratorium.com)
e = Biological Analytics, Novartis Pharma AG, CH-4002 Basel, Switzerland
Abbreviations: CV, coefficient of variation; DL, developing laboratory; DMSO,
dimethyl sulfoxide; ECVAM, European Center for the Validation of Alternative
Methods; ELC, endotoxin limit concentration; ELISA, enzyme-linked
immunosorbent assay; IU, international unit; GLP, good laboratory practice; h,
hours; IL, interleukin; IPT, in vitro pyrogen test; LoD, Limit of Detection; LPS,
lipopolysaccharide; LTA, lipoteichoic acid; MVD, maximum valid dilution; PEI,
Paul Ehrlich Institute; PM, prediction model; PPC, positive product control; NL,
naive laboratory; NPC, negative product control; OD, optical density; SOP,
standard operating procedure; WHO, World Health Organisation
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
75
Corresponding author:
Thomas Hartung, MD, PhD
European Commission
Joint Research Centre
Institute for Health & Consumer Protection
ECVAM, I-21020 Ispra (VA)
e-mail: [email protected]
Tel: +39-0332-786256/ Fax: +39-0332-786297
6.1. Abstract
Pyrogens as fever-inducing agents can be a major health hazard in parenterally
applied drugs. For the control of these contaminants, pyrogen testing for batch
release is required by Pharmacopoeias. This has been done either by the in
vivo rabbit pyrogen test (since 1942) or the limulus amoebocyte lysate test
(LAL), since 1976. A new approach are cell-based assays employing in vitro
cultivation of human immune cells which respond e.g. with cytokine production
(IL-1β; IL-6) upon contact to pyrogens. Six variants of these assays have
recently been validated in a collaborative international study. Recently, the
development of successful cryopreservation methods promises to make
standardized immunoreactive primary human blood cells available for
widespread use. Furthermore, the pretesting of donors for infectious agents
such as HIV or hepatitis has made it possible to develop a safe and
standardised reagent for pyrogen testing. Using altogether 13 drugs, we have
validated here the pyrogen test based on fresh and cryopreserved human whole
blood in four laboratories. The test reached > 90% sensitivity and specificity. In
contrast to the LAL, the test is capable of detecting non-endotoxin pyrogens
derived from Gram-positive bacteria or fungi.
Keywords: Pyrogen testing; validation study; IL-1β; cryopreservation
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
76
6.2. Introduction
Testing for pyrogens has employed animals, either the rabbit (since 1942) or
the horseshoe crab, Limulus polyphemus (since 1976). Pyrogens, especially
lipopolysaccharide (LPS) as part of the cell walls of Gram-negative bacteria, are
an ubiquitous threat to human health due to their stability. Other pyrogens are
LTA (Morath et al., 2001), exotoxins (149, 18, 164), muramyldipeptide (165) or
peptidoglycan (22, 23,149). Alternative pyrogen tests, altogether 6 cell-based in
vitro pyrogen tests using the production of inflammatory mediators, (i.e. IL-1β,
TNF-α or IL-6 as well as neopterin) (detailed by 49, 65, 67, 142, 143, 150) of
human blood cells have been developed and validated in an international
collaborative study (48).
Since then, two variants, one using freshly drawn human whole blood and
measuring the IL-1β response (48, 53) and another using freshly drawn isolated
PBMCs and measuring IL-6 (150), have undergone further development and
refinement by developing a standardized freezing procedure using DMSO as a
cryoprotective agent. This allows both methods to become more standardized
and more widely available. For the human whole blood test (IPT), even two
variations of freezing were developed and optimized. The newer one was
optimized for storage at -80°C, the other one making use of storage in liquid
nitrogen has been described previously (73).
In a second study identical to the one previously performed, the refined human
whole blood assay was validated. This study focused again on LPS as a
pyrogen, since it is the most frequent contamination, and the capability of the
cell-based assays of detecting non-endotoxin pyrogens had been demonstrated
in the previous study.
6.3. Materials and Methods
The test system was validated in the developing lab and in three different naïve
labs (Table 1) after a detailed SOP had been compiled by the developing lab
(DL) and made available for the naïve labs (NL) by ECVAM. The successful
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
77
technology transfer had been assessed in a prevalidation phase.
Since the developing lab of the IPT (endpoint IL-1β) cannot work under GLP, the
assay was performed by three naïve labs (PEI, NL 1, Qualis, NL 2 and Novartis,
NL 3). The IPT was performed in three variants, one involving fresh blood and
two involving frozen whole blood stored at -80°C and liquid nitrogen,
respectively. (* The results of this lab did not enter the formal evaluation since it
is not operating under good laboratory practice (GLP)).
Test
system
Developing
lab (DL)
1st lab
(NL 1)
2nd lab
(NL 2)
3rd lab
(NL 3)
IPT
IL-1β
University of
Konstanz *
Paul-Ehrlich
Institute
Qualis
laboratories
Novartis
Pharma
Table 1: Laboratories performing the assays
Endotoxin stimulus
The second international WHO standard for endotoxin 94/580 from E. coli O113:
H10, which was used in the previous validation study, served again as the
standard endotoxin (Poole et al., 1997). 100 pg/ml of this endotoxin are
equivalent to 1 IU (International Unit)/ml.
Fresh and cryopreserved human whole blood test IL-1β
Blood collection
Blood from healthy donors was collected into heparinized tubes (Li-Heparin, 15
IU/ml) (Sarstedt-monovette, 7.5ml, Nümbrecht, Germany) using a multifly
needle set and used within 4h. Additionally, for the preparation of
cryopreserved blood, a sample from each of the five donors was drawn into
Serum- and EDTA tubes.
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
78
(Sarstedt) for differential blood counts and infection serology. For the fresh
blood assay, one single donor was used and his blood was subjected to a cell
count (Pentra 60, ABX Diagnostics, Montpellier, France), in order to exclude
infections.
Testing for infectious agents in cryopreserved blood
The additionally drawn blood of each donor was tested by a qualified laboratory
(Dr. U. Brunner, Konstanz, Germany) for Hepatitis A, B, C and HIV according to
the standards for blood donations for transfusion purposes in Germany. In the
meantime, the freezing of the blood took place.
Freezing for storage at -80°C (Method A)
Endotoxin-free Soerensen Buffer (Acila GMNmbh, Mörfelden-Walldorf,
Germany) was mixed with 20% v/v endotoxin-free DMSO (Wak-Chemie
Medical GmbH, Steinbach, Germany). 1.8ml cryotubes (Nalge Nunc
International, Denmark) were screwed open under a laminar flow bench and
0.6ml of the pooled blood of five donors were pipetted into each one. Using a
multipette with a 5ml combitip, the cryoprotective solution was added in three
aliquots of 200µl each, gently swirling the blood in between. The tubes were
closed and placed in a storage box (Nalge Nunc), leaving a space of about 1
cm between each vial in order to ensure a homogenous freezing process. The
boxes were then placed in a freezer at –80°C and left to freeze. They were
stored at –80°C.
Freezing for storage in liquid nitrogen (Method B)
An alternative freezing procedure involved a controlled freezing process using
the vapour phase of liquid nitrogen and has been described earlier (73). For this,
the DMSO was added directly to the blood of the individual donors at 10% final
concentration (v/v). The blood was then pooled and aliquoted at 1.2 ml/vial and
frozen in a computer-controlled freezer (Nicool Plus PC, AirLiquide, Marne-la-
Vallée Cedex 3, France) until at -120°C. The aliquots were then taken out of the
machine and placed in the vapour phase of liquid nitrogen.
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
79
Shipment
The required number of aliquots of frozen blood was shipped to the other
laboratories using a voyager containing liquid nitrogen (Air Liquide Kryotechnik
GmbH, Düsseldorf, Germany). The temperature was monitored using a
computer-controlled temperature probe (Thermory Mobile, Air Liquide, France;
software Logiciel Recwin, Marne la vallée Cedex, France). After arrival, the
aliquots were kept either in the transportation vehicle itself or transferred to the
vapour phase of liquid nitrogen, if available. Alternatively, the aliquots that had
been frozen at -80°C (Method A) could be retransferred to the freezer at -80°C.
Thawing procedure
The vials were taken out of the voyager/ the nitrogen tank and placed
immediately in an incubator at 37°C. After 15 min, the blood was pooled in a
centrifuge tube if more than one aliquot was used, and gently swirled in order to
ensure complete mixing.
Pretesting of the aliquots
After all donors were clearly negative regarding the infectious agents in
question, the cryopreserved blood was pretested before sending it out to the
other laboratories by carrying out a dose-response curve using the WHO
standard or an endotoxin calibrated to it. The criteria were an absorbance for
the saline control of 0.1 OD or lower, and a response to the 0.5 IU/ml of at least
1.6 times the OD of the saline control.
Incubation procedure
Cryopreserved blood (Method A and B)
Method A: 180µl of RPMI (Charles River Endosafe), 20µl sample/control and
40µl of thawed blood were added to a pyrogen-free microtiter plate (Falcon).
After adding the blood, the contents of the wells were mixed by gently
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
80
aspiring/dispensing 5 times using a multichannel pipette and sterile, pyrogen-
free tips, changing the tips in-between the rows in order to avoid cross
contaminations. The plates were then covered with a lid and placed in an
incubator at 37°C and 5% CO2 for 10-24 h.
Method B: The alternative version using nitrogen-stored blood was handled the
same, except that the incubation involved 200µl of RPMI, 20µl of
samples/controls and 20µl of thawed blood, since the blood had not been
prediluted in the freezing process.
Fresh blood (Method C)
In order to allow direct comparisons, the method validated in 2005 (Hoffmann et
al., 2005a) was adapted from 1ml vial incubation to microtiter plates as used for
the cryopreserved blood. This variant was also included in the validation to
exclude an effect of this changed format.
200µl of saline (Charles River Endosafe, Charleston, South Carolina, USA),
20µl sample/control and 20µl of blood were added to a pyrogen-free microtiter
plate (Falcon 96well flatbottom tissue culture plate, Becton Dickinson Labware,
Meylan Cedex, France). After adding the blood, the contents of the wells were
mixed by gently aspiring/dispensing 5 times using a multichannel pipette and
sterile, pyrogen-free tips, changing the tips in between the rows in order to
avoid cross- contaminations. The plates were then covered with a lid and
placed in an incubator at 37°C for 10-24 h.
ELISA procedure
The IPT Kit was used (Charles River Endosafe). Aliquots of 100µl of each well
of the incubation plates were added to the wells of the ELISA plate. When
transferring the supernatants, they were mixed by aspiring and dispensing them
2-3 times using a multichannel pipette. The ELISA was done according to the
manufacturer´s instructions.
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
81
Data analysis
Data analysis was the same as in the previous validation. The quality criterion
for acceptable variability, i.e. allowing a maximum coefficient of variation (CV)
of 0.45 has been empirically established in the previous study for both assays
in order to ensure the interpretability of test results and was adopted here. The
data of those samples and control exceeding this CV-value were tested for
outliers by the Grubbs-test. If an outlying replicate caused the excessive
variation it was excluded and further analysis was performed with the remaining
three replicates. The samples and controls, whose large variation was not
caused by an outlier were excluded from further analysis. In the cases when
the positive product control (PPC) CV exceeded 45%, the corresponding 0.5
IU/ml in saline, which was part of the dose-response curve using the WHO
standard endotoxin, were used instead. If the CV of this standard also
exceeded 45%, the whole set of data was not considered for analysis.
Furthermore, the response of the 0.5 IU/ml had to be significantly higher, i.e. a
p-value below 0.01, than the respective response of the negative saline control.
Accepted data were analysed by a so-called prediction model (PM): the data of
a blinded sample were compared with the PPC data or, if the PPC did not fulfil
the qualitiy criterion, the 0.5 IU/ml control using a one-sided t-test with log-
transformed data and a local significance level of 1%.
Blinding procedure
All test items are registered medicinal products and were obtained from a
pharmaceutical supplier. For the validation, test items and endotoxin spiking
samples were prepared by the University of Konstanz and blinded/coded under
GLP by personnel (G. Bowe and J. de Lange) from ECVAM, Italy. These were
then shipped by the University of Konstanz to each of the appropriate test
facilities participating in the study.
Prevalidation
The drugs used were the same as in the previous validation study (Hoffmann et
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
82
al., 2005a), that is Gelafundin, a volume-replacement therapy for transfusion
with high protein content (B. Braun Melsungen AG, Melsungen, Germany),
Jonosteril, an electrolyte infusion (Fresenius AG, Bad Homburg, Germany) and
Haemate, a factor VIII preparation (Aventis Behring GmbH, Marburg,
Germany). Additionally, a negative and a positive control (0.5 IU/ml) were
included in each run.
This set was tested in the developing laboratory (DL) Konstanz as well as in
two naive laboratories PEI (NL 1) and Qualis (NL 2) with the three approaches
of the IPT (Method A-C) in order to prove successful transfer.
Prior to preparing the spikes, an interference test was performed with all three
substances by the DL. Although this had been done in the previous validation,
a shift in the interference due to the DMSO/freezing process could not be
excluded. Interferences differed indeed for the IPT (data not shown), and the
spikes were calculated according to the required dilution.
Sample preparation and blinding was done at the University of Konstanz using
pyrogen-free clinical grade saline and the WHO reference standard endotoxin.
Validation
For the validation phase, in order to maximise comparability with the previous
validation study, the same ten drugs were employed. The concentrations were
based on a recent in-depth analysis of the fever response of a sensitive rabbit
strain (Hoffmann et al., 2005b): Five blinded spikes, two of them defined as
non-pyrogenic, that is below 0.5 IU/ml (0 and 0.25 IU/ml), and three as
pyrogenic (2 x 0.5 and 1.0 IU/ml) were tested in the different laboratories. All
drugs were tested at their MVD (maximum valid dilution), thus adopting the
rationale of the pharmacopoeial LAL reference (limit) test. The MVD is
calculated from the endotoxin limit concentration (ELC) in IU/ml, defined for a
drug by the European Pharmacopoeia (146), divided by the threshold of
pyrogenicity as the limit of detection (LoD), in this case 0.5 IU/ml. Drugs,
sources, ELCs and MVDs are summarized in Table 2.
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
83
Drug code Source Agent Indication MVD
Glucose
5 % (w/v)
GL Eifel Glucose nutrition 70
Ethanol
13 % (w/w)
ET B. Braun Ethanol diluent 35
MCP ME Hexal Metoclopramid antiemetic 350
Syntocinon SY Aventis Oxytocin initiation
of delivery
700
Binotal BI Aventis Ampicillin antibiotic 140
Fenistil FE Novartis Dimetinden-
maleat
antiallergic 175
Sostril SO GlaxoSmith
Kline
Ranitidine antiacidic 140
Beloc BE Astra Zeneca Metoprolol
tartrate
heart
dysfunction
140
Drug A LO - 0.9 % NaCl - 35
Drug B MO - 0.9 % NaCl - 70
Table 2: Drugs employed in the validation
All ten clinical-grade drugs, that had been used in the previous validation using
freshly drawn cells, were used again. All drugs were used at their respective
MVD (maximum valid dilution).
6.4. Results
Pretesting of the cryopreserved blood
The blood was tested employing an E. coli O113: H10 dose-response curve
(Fig. 1). The blood was considered suitable since the OD of the saline control
was below 0.1 OD and the mean OD of the 0.5 IU/ml was 1.6 times the mean
OD of the saline or higher.
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
84
0
1
2
3
version - 80°C(Method A)
nitrogen version(Method B)
controlO-113; 0.25 EU/mlO-113; 0.5 EU/mlO-113; 1.0 EU/mlO-113; 2.0 EU/ml
OD
[450n
m]
Fig. 1: Pretesting of both versions of the cryopreserved blood prior to
shipping
Both versions of the cryopreserved blood (Method A and B) were pretested
measuring the release of IL-1β in response to a pyrogenic stimulus prior to
shipping them to the participating labs. For this purpose, a dose-response
curve (0.25-2 IU/ml) using the international WHO standard was done (n=4).
Prevalidation
Method A – Cryoblood, -80°C version
The data produced with the method based on cryoblood frozen at -80°C are
summarized in Figure 2 as an example for all three methods. It has to be noted
that for the NL 1 the level approximated the maximum response level of 4 OD.
This might cause problems for discriminating pyrogenic spikes, especially when
the positive control of 0.5 IU/ml produces such high-level responses. The
background OD-levels were small in the NL 1 and in the DL. NL 2 provided the
data with the background level subtracted. The three standard curves,
consisting of C-, (0 IU/ml), 0.25, 0.5 and 1 IU, indicate a typical monotone
increase in OD-response with increasing concentration.
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
85
J-0
J-0
J-0.5
J-1
G-0
G-0
G-0
.5G-1
H-0
H-0
H-0
.5H-1 C
-
0.25
EU
0.5 E
U (C+)
1 EU
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9 Konstanz (DL)
OD
J-0
J-0
J-0.5
J-1
G-0
G-0
G-0
.5G-1
H-0
H-0
H-0
.5H-1 C
-
0.25
EU
0.5 E
U (C+)
1 EU
0
1
2
3
4PEI (NL1)
OD
J-0
J-0
J-0.5
J-1
G-0
G-0
G-0
.5G-1
H-0
H-0
H-0
.5H-1 C
-
0.25
EU
0.5 E
U (C+)
1 EU
0.0
0.5
1.0
1.5
2.0
2.5
3.0Qualis (NL2)
OD
Fig. 2: Prevalidation data for Method A of the three involved laboratories.
The treatments and controls are abbreviated (J = Jonosteril: G = Gelafundin;
H = Haemate; C- = saline; C+ = positive control) indicating the endotoxin
contamination in IU (0, 0.5 and 1 IU/ml).
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
86
As Figure 2 only gives an indication about variability of replicates, the CVs were
calculated for all samples and controls for all laboratories. While the major part
of the total of 48 CV-values was smaller than 30%, four samples, all of which
were unspiked, showed a CV larger 45%. In all of these, one of the four
replicate values was much larger than the others, and thus caused the high
variability, and was excluded by outlier analysis.
Application of the prediction model (PM) to these data revealed that eleven out
of the twelve spikes were classified in the same way in all laboratories.
Comparing the laboratories pairwise showed that 34 of the total of 36 single
comparison, i.e. 94.4%, resulted in the same classification.
Assessing in the final step preliminarily the predictive capacity, revealed that all
negative samples were classified correctly and that one 0.5 IU spike (NL 1: H-
0.5) at the rabbit classification threshold was classified false-negative. In terms
of performance parameters, this resulted in a specificity of 18/18 = 100% and a
sensitivity of 17/18 = 94.4%.
Method B – Cryopreserved blood (liquid nitrogen version)
Background OD-levels were small in the DL and NL 1. NL 2 provided the data
with the background level subtracted. The three endotoxin standard curves
indicate a higher limit of detection as the 0.25 IU standards, and for NL 2 also
the 0.5 IU standard, gave low OD-responses (data not shown).
The CVs were calculated for each treatment or control for all laboratories. While
the major part of the CVs was smaller than 40%, six samples (mainly from DL)
and one standard showed a CV larger than 45%. Nine out of the twelve spikes
were classified in the same way in all laboratories. Comparing the laboratories
pairwise showed that 30 of the total of 36 single comparison, i.e. 83.3%,
resulted in the same classification.
Assessing in the final step preliminarily the predictive capacity, revealed that
one negative samples was classified wrongly (NL 2: J-0) due to one outlying
value, and that two times a Haemate 0.5 IU/ml sample (DL and NL 1) at the
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
87
rabbit classification threshold was classified false-negative. In terms of
performance parameters, this resulted in a specificity of 17/18 = 94.4% and a
sensitivity of 16/18 = 88.9%.
Method C – fresh whole blood
Background OD-levels were small in the DL and NL 1. NL 2 provided the data
with the background level subtracted. The three standard curves, consisting of
the negative control C- (0 IU/ml), 0.25, 0.5 and 1 IU/ml, showed a typical
monotonous increase in OD-response with increasing concentration.
The CVs were calculated for each treatment or control for all laboratories. In
general, the CVs were smaller than 30%. Only two samples resulted in a CV
larger than 45%. These two samples were an H-0 tested at NL 2, which was
caused by an aberrant value, and a G-0 tested in DL with a CV of 48.8%.
Furthermore, a tendency for larger CV of endotoxin-free samples/treatments
was observed, as the background OD-level was lower compared to comparable
assays, e.g. in the main validation study. Ten out of the twelve spikes were
classified in the same way in all laboratories. Comparing the laboratories
pairwise, showed that 32 of the total of 36 single comparison, i.e. 89.9%,
resulted in the same classification.
Assessing in the final step preliminarily the predictive capacity, revealed that all
negative samples were classified correctly and that two 0.5 IU/ml spikes (DL: J-
0.5; NL 1: H-0.5), which are at the rabbit classification threshold, were classified
false-negative.
In terms of performance parameters, this resulted in a specificity of 18/18 =
100% and a sensitivity of 16/18 = 88.9%.
Validation
Inter-laboratory reproducibility
As within-laboratory reproducibility was generally successfully shown in
prevalidation, only inter-laboratory reproducibility, based on three laboratories
per test method, was assessed in the validation. The similarity of laboratories
was based on the classification resulting from the PM (Table 3) to compare the
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
88
laboratories with each other with respect to concordance, i.e. without taking the
true classifications of the samples into account. The results are presented in
Table 4. The overall inter-laboratory reproducibilities of IPT Method A and C are
consistently high. Regarding IPT Method B, only about 65% of the samples
were classified the same way due to NL 3 where four drugs caused problems.
IPT Method A IPT Method B IPT Method C
drug
spike
(IU/ml) truth NL 1 NL 2 NL 3 NL 1 NL 2 NL 3 NL 1 NL 2 NL 3
0.0 0 0 0 CV 0 0 0 0 0 0
0.25 0 0 1 CV 0 0 CV 1 0 0
0.5 1 1 1 1 0 1 1 1 1 1
0.5 1 1 1 1 1 1 1 1 1 1
Beloc
1.0 1 1 1 1 1 1 1 1 1 1
0.0 0 0 0 0 0 0 0 0 0 0
0.25 0 0 1 1 0 1 CV 0 1 0
0.5 1 0 1 1 1 1 1 1 1 1
0.5 1 1 1 1 1 1 1 1 1 1
Binotal
1.0 1 1 1 1 1 1 1 1 1 1
0.0 0 0 0 nq 0 0 0 0 0 CV
0.25 0 CV 1 nq 0 0 CV 0 1 0
0.5 1 1 1 nq 1 1 0 1 1 1
0.5 1 1 1 nq 1 1 0 1 1 1
Ethanol
1.0 1 1 1 nq 1 1 1 1 1 1
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
89
0.0 0 0 0 0 0 0 0 0 0 0
0.25 0 0 1 1 0 CV 1 CV 1 1
0.5 1 1 1 CV CV 1 CV CV 1 1
0.5 1 1 1 1 1 1 1 CV 1 1
Fenistil
1.0 1 1 1 1 1 1 1 1 1 1
0.0 0 0 0 0 0 0 0 0 0 0
0.25 0 CV 0 0 0 1 0 0 1 1
0.5 1 1 1 1 0 1 CV 1 1 1
0.5 1 1 1 1 0 1 1 1 1 1
Glucose
1.0 1 1 1 1 0 1 1 1 1 1
0.0 0 0 0 0 0 0 0 0 0 0
0.25 0 CV 1 CV 0 1 0 0 0 1
0.5 1 1 1 1 0 1 CV 0 1 1
0.5 1 1 1 1 0 1 1 1 1 1
MCP
1.0 1 1 1 1 1 1 1 CV 1 1
0.0 0 0 0 nq 0 0 nq 0 0 0
0.25 0 0 1 nq 0 1 nq CV 0 1
0.5 1 0 1 nq 0 1 nq 1 1 CV
0.5 1 1 1 nq 1 1 nq 1 1 1
Sostril
1.0 1 1 1 nq 1 1 nq 1 1 1
0.0 0 0 0 0 0 0 nq 0 0 0
0.25 0 0 CV CV CV CV nq 0 0 0
0.5 1 1 1 1 1 1 nq 1 1 1
0.5 1 1 1 1 1 1 nq 1 1 1
Synto-
cinon
1.0 1 1 1 1 1 1 nq 1 1 1
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
90
0.0 0 CV 0 nq 0 0 nq CV 0 0
0.25 0 0 0 nq 0 0 nq 0 0 0
0.5 1 1 1 nq 0 1 nq CV 1 1
0.5 1 1 1 nq 0 1 nq 1 1 1
A
(saline)
1.0 1 1 1 nq 1 1 nq 1 1 1
0.0 0 0 0 nq 0 0 nq 0 0 0
0.25 0 0 0 nq 0 0 nq 0 0 CV
0.5 1 1 1 nq 0 1 nq CV 1 1
0.5 1 1 1 nq 0 1 nq 1 1 1
B
(saline)
1.0 1 1 1 nq 1 1 nq 1 1 1
Sample size n 46 49 25 48 48 24 42 50 47
Specificity 100 68.4 75 100 77.8 88.9 94.1 80.0 76.5
Sensitivity 93.3 100 100 62.1 100 86.7 96.0 100 100
Table 3: Classifications of samples by all methods and all laboratories in
validation
Grey shading indicates that for these drugs the PPCs did not qualify so that the
PC was used in the PM.
CV = sample showed a variability resulting in exclusion, i.e. CV > 45 % and no
significant outlier present.
nq = not qualified according to quality criteria, i.e. failure of PPCs and PCs
0 = considered/classified negative/ 1 = considered/classified positive
False classifications are in bold type
Predictive capacities
Table 4 summarises the sensitivity and specificity for each method together with
the respective sample sizes. For IPT Method A, eight samples at the NL 1 and
three at the NL 3 were excluded due to their high variability, i.e. CVs > 45%. For
IPT Method B and Method C, the sample sizes were reduced for both methods in
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
91
addition to ten samples with high variability by four drugs tested at the NL 3 that
did fail the quality criteria for both the PPC and the PC.
Test
Inter-laboratory
reproducibility (%)
Sample size:
sensitivity
Sensitivity
(%)
Sample size:
specificity
Specificity
(%)
IPT
Method A
DL-NL 1: 86.7
DL-NL 2: 87.5
NL 1-NL 2: 100
77
97.4
45
82.2
IPT
Method B
DL-NL 1: 66.0
DL-NL 2: 63.3
NL 1-NL 2: 83.3
74
82.4
46
89.1
IPT
Method C
DL-NL 1: 88.1
DL-NL 2: 89.7
NL 1-NL 2: 91.5
84
98.8
55
83.6
Table 4: Inter-laboratory reproducibility and sensitivity/specificity with the
respective sample sizes in validation
Inter-laboratory reproducibility was calculated by the proportion of samples
classified identically for each pair-wise laboratory comparison.
The overall performances of the IPT Methods A and C were very good: High
sensitivities over 90% could be achieved, while specificities around 80% were
established, reflecting the safety approach in the PM emphasizing sensitivity. In
contrast to these methods, the IPT Method B performed differently with a higher
specificity of 89% on cost of a decreased sensitivity (82%). Misclassifications
occurred with one exception only for samples with contaminations close to the
pyrogenicity threshold, i. e. 0.25 and 0.5 IU/ml. The 95%-confidence intervals
for the parameters were calculated, assuming binomial distribution, and are
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
92
presented in Figure 3. The according parameters for the rabbit test, calculated
with the model of Hoffmann et al. 2005 (16), were also included. While the new
test except IPT Method B had a slightly lower specificity than the rabbit test, the
sensitivity was substantially increased by 20% up to 40%.
0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1,0
0,0 0,1 0,2 0,3 0,4 0,5
1-specificity
sensitiv
ity
IPT Method A
IPT Method B
IPT Method C
Rabbit
Fig. 3: Sensitivity and specificity of validation with 95% confidence
intervals including the rabbit pyrogen test
The sensitivity and specificity of the in vitro methods was assessed after the
validation phase and compared to the rabbit pyrogen test assuming binomial
distribution. The y-axis shows the sensitivity in percent, the x-axis the specificity
for the rabbit pyrogen test and all four in vitro assays.
6.5. Discussion
This follow-up study had the aim to refine and improve the already established,
validated systems for pyrogen testing using freshly drawn human blood and
make them available as a safe and standardized reagent in cryopreserved
form. In comparison to the fresh blood assays, the cryopreserved cells showed
a higher response with regard to IL-1β, but the variance also tended to be
higher.
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
93
Therefore, some sets of data of all of the IPT methods had to be excluded due
to variance-controlling quality criteria. In the former validation (48), the whole
blood test, at this time performed with fresh blood in reaction tubes, achieved a
sensitivity of 72.7% (n= 88) and a specificity of 93.2% (n= 59). The inter-lab
reproducibility had been 72.9 and 81.6% between the developing lab and the
two naïve labs. In this study, the sensitivity of the fresh blood assay could be
improved by transferring it to the microtiter plate (98.8%) with a minor reduction
of specificity (83.6%). This approach proved to be more easily transferable to
the naïve labs (inter-lab reproducibility 88-92%). The frozen cells performed
well, although the -80°C method (A) was better transferable than the nitrogen
method (86-100% vs. 63-83% for the inter-laboratory reproducibility). However,
one of the laboratories (NL 3) had particular problems with both cryopreserving
methods as the data of four drugs failed quality criteria and thus had to be
excluded from analysis.
We assume that the IPT assay at the NL 3 posed problems in performance
because of insufficient transfer of the method. This problem could not be
overcome because of the very tight time schedule of the validation.
Nevertheless, IPT-data generated in parallel at Konstanz (Table 5) yielded
good results, although they did not enter the formal evaluation according to the
study protocol which foresaw only the participation of three GLP-concordant
laboratories. The importance of successful assay transfer is stressed
considering the fact that, had the results of the developing lab instead of those
of NL 3 entered the evaluation, the sample size for method A qualifying for
evaluation had increased from 120 to 143.
IPT A IPT B IPT C
sample size n 48 37 47
Specificity 100 100 94.4
Sensitivity 96.7 90.5 86.2
Table 5: Performance of the IPT in the developing laboratory
INTERNATIONAL VALIDATION OF PYROGEN TESTS BASED ON
CRYOPRESERVED HUMAN PRIMARY BLOOD CELLS
94
The developing laboratory in Konstanz performed all three methods of the IPT
in parallel to the naïve labs. This table shows the sample size that qualified for
evaluation and the achieved specificity and sensitivity for these samples.
The specificity of this method would have increased slightly to 89.5% and the
sensitivity would hardly have changed at all, indicating that the method of
evaluation used in this study was adequate. In general, the –80°C version
produced higher ODs and was more sensitive than the nitrogen version of the
cryoblood. It was noticeable, though, that the linear range in the dose response
curve from 0.25 IU/ml to 1 IU/ml is much smaller in the –80°C than in the
nitrogen version. All in all, this validation has shown that the novel, recently
validated pyrogen test based on human blood can be performed with
cryopreserved cell preparations. The IPT could be improved with regard to
performance making it at the same time easier to handle by transferring it to the
microtiter plate. The optimization employing cryopreserved cells allows the
assay to become more standardized and cells as test reagents more widely
available. The fact that safety standards of blood transfusions can be
implemented as shown for the IPT stresses that concerns of possibly infected
donors can be ruled out.
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
95
7 Pyrogen testing of lipidic parenterals with a novel in
vitro test
Application of the InVitro Pyrogen Test (IPT) based on cryopreserved human
whole blood
Stefanie Schindler+, Ute Rosenberg*, Detlef Schlote*, Katja Panse*, Andreas
Kempe*, Stefan Fennrich+ and Thomas Hartung+#
* Schering AG, Biological Quality Control, Müllerstrasse 178, D-13342 Berlin
+ Biochemical Pharmacology, University of Konstanz, Universitätsstr.10, D-
78457 Konstanz, Germany
# European Commission, Joint Research Centre, Institute for Health &
Consumer Protection, ECVAM, I-21020 Ispra (VA)
Abbreviations: CV, coefficient of variation; DMSO, dimethylsulfoxide;
ECVAM, European Center for the Validation of Alternative Methods; ELC,
Endotoxin Limit Concentration; ELISA, Enzyme-linked immunosorbent assay;
EU, Endotoxin Unit; IL, Interleukin; LAL, Limulus Amoebocyte Lysate; LPS,
lipopolysaccharide; MVD, Maximum Valid Dilution; PPC, positive product
control; NPC, negative product control; IPT, In Vitro Pyrogen Test; WHO,
World Health Organisation
Corresponding author
Thomas Hartung, MD, PhD
European Commission, Joint Research Centre
Institute for Health & Consumer Protection
ECVAM/ I-21020 Ispra (VA)
e-mail: [email protected]
Tel: +39-0332-785939/ Fax: +39-0332-786297
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
96
7.1. Abstract
The European Pharmacopoeia has made the testing of small volume
parenterals (< 15 ml) obligatory since 2004. This concerns many
formulations, e.g. vitamins, steroids and hormones, many of which are
applied intramuscularly using a lipidic carrier. Lipopolysaccharides, the best
established endotoxins from Gram-negative bacteria, bind strongly to
lipophilic substances which mask them in Limulus amebocyte lysate assays.
End-product testing, therefore, can only be carried out in rabbit pyrogen tests.
This will of course lead to a pronounced increase in animal experiments if no
alternative procedures become available. We have described a novel in vitro
pyrogen test (IPT) based on human whole blood, which has recently been
validated in a collaborative study including the European Pharmacopoeia.
Here, the utility of the IPT for lipophilic substances and lipid-containing end-
products was assessed.
For a variety of lipids commonly added to formulations of injectable
endproducts, namely peanut oil, sesame oil, miglyol and paraffin, a protocol
which allows interference-free testing was established applying the
pharmacopoeial criterium of 50 to 200% retrieval of an LPS spike.
Furthermore, end-product testing for three formulations was possible. In all
cases a method could be established which allows to determine given or
calculated ELC (endotoxin limit concentrations) according to Pharmacopoeia.
It is concluded that the monocytes react to lipid-bound LPS showing that
immune responses to contaminated endproducts must be anticipated and
that the IPT is suitable for endproduct control of these formulations.
Keywords: Pyrogen testing; alternatives to animals; lipids; IL-1β
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
97
7.2. Introduction
The Limulus amebocyte lysate assay (LAL) has replaced about 80% of the in
vivo tests in rabbits. However, complex preparations challenge this assay
because of interferences. This holds true especially for biologicals such as
blood products through binding of LPS to lipoproteins (166), but also for
lipophilic preparations (167) or liposomes (168). Similar problems are seen
for vaccines (169), which usually contain adjuvants such as aluminium
hydroxide which binds and masks endotoxins, making them undetectable in
the LAL (170).
If such endproducts are controlled (some are not controllable), dilutions in
pyrogen-free distilled water are established to allow endotoxin spike retrieval in
the LAL. Preparing dilutions of lipids in water is extremely difficult, since the
phases will only remain mixed for a few seconds after vortexing and a
homogenous and representative distribution of a contamination in the
hydrophilic diluent cannot be determined. A considerable proportion of LPS
could remain in the lipophilic phase and remain unaccessible for testing and,
additionally, be lost by binding to plastic pipette tips and vials. Very viscuous
samples, such as castor oil, are almost impossible to pipet. Therefore, the true
pyrogen content of the original sample is extremely difficult to judge.
The obvious alternative, the rabbit pyrogen test, is costly, laborious, and
especially in the case of lipid substances, ethically problematic. The procedure
of injecting lipid samples intravenously bears the risk of seriously harming the
animal through the formation of miniscule drops which can obstruct small
vessels e.g. in the kidney. The rabbit assay itself has some shortcomings which
are often overlooked, namely the lack of standardization with regard to animal
strain and age (27), the lack of positive and negative controls, and the small
number of animals used per sample.
Shortly after the human fever reaction was elucidated by identifying
proinflammatory cytokines of leukocytic origin (37), Duff and Atkins (134) as
well as Dinarello et al. (133) suggested using this in
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
98
vitro for future pyrogen testing. Problems in standardizing such cellular assays
initially hampered their routine application. But, recently a collaborative study,
including the European Pharmacopoeia, compared and validated such cell-
based pyrogen tests for inclusion into Pharmacopoeia (48). We contributed a
simplified and therefore highly standardized pyrogen test employing human
whole blood (IPT) in which the release of IL-1.beta. is the endpoint (49) to this
validation exercise. The advantage of testing the pyrogenic reactions of
monocytes in their natural environment comprises the presence of the
endotoxin-presenting plasma protein Lipopolysaccharide-Binding Protein (LBP)
which is capable to transfer LPS to the cellular binding site (171). Further
advantages in using primary cells instead of cell lines lie in their availability, not
requiring laborious maintenance and culture steps, and the fact that, unlike cell
lines, they do not harbour the risk of undergoing dedifferentiation or oncogenes,
therefore ensuring that they express all receptors and signalling proteins
necessary for their response to pyrogens.
Other cell-based approaches (reviewed by Poole and Gaines Das, 2001) (172)
measured IL-6 (63, 142) or TNF-α (65) as well as neopterin (66, 67). The whole
blood assay has since then been further developed and optimized for use with
cryopreserved blood (73) and is available in a standardized format.
Here, we addressed the pressing question whether lipid formulations of drugs
can be adequately controlled by IPT.
7.3. Materials and Methods
Lipids
Sesame oil and castor oil (both Schering AG, Berlin, Germany) as well as
peanut oil, miglyol and paraffin (L+S, Bad Bocklet, Germany) were used.
Another hydrophobic additive was benzyl benzoate (Schering). The
endproducts tested were Testoviron®, Gynodian® and Noristerat® provided by
Schering. Calibration and end-product testing was performed using
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
99
commercially available sesame oil (Sigma-Aldrich, Steinheim, Germany). All oily
substances were kept at room temperature or at 4°C and placed in an incubator
or in a water bath at 37°C for 30 minutes prior to use. Dilution of the castor oil
was done using endotoxin-free DMSO (Cryo-Sure DMSO, Wak-Chemie,
Steinbach, Germany)
Standards
The 2nd International WHO Standard for endotoxin (from E. coli O113: H10:K(-)
(94/580), which is identical to FDA/USP standard EC6/Lot G was used as the
standard endotoxin (56) for calibration. 100 pg of this standard correspond to
the pyrogenic activity of 1 EU (Endotoxin Unit).
The endotoxin used for the testing of the lipids was a calibrated lyophilisate of
E. coli O111: B4 provided in the IPT Kit (Charles River Endosafe, Charleston,
South Carolina, USA).
The Gram-positive standard used for the testing was a calibrated lyophilisate of
B. subtilis (57) provided in the IPT Kit (Charles River Endosafe).
Cryopreserved human whole blood
The procedure of freezing human whole blood has been described previously
(73). Briefly, the freshly drawn heparinized blood of five healthy donors is mixed
with pyrogen-free DMSO at a final concentration of 10% (v/v ratio), and pooled.
The blood is frozen in 1.5 or 4ml aliquots in a computer-controlled freezer
(Nicool Plus PC, Air Liquide, Marne-la-Vallée Cedex 3, France) and stored in
the vapor phase of liquid nitrogen. Before use the cryopreserved blood was
thawed at 37°C for 15 minutes.
Incubation
Basic procedure
A) Tube method
The procedure of the whole blood incubation has been described previously
(Hoffmann et al., 2005a). The incubation was carried out in pyrogen-free
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
100
polypropylene 1.5 ml reaction tubes (Eppendorf, Hamburg, Germany). Briefly,
1000 µl of pyrogen-free physiological saline (Charles River Endosafe or Berlin-
Chemie, Berlin, Germany) were pipetted into each tube, and 100 µl of
standard/samples and 100 µl of cryopreserved blood were added. The tubes
were closed, inverted twice to achieve complete mixing of the contents and
placed overnight in an incubator at 37°C. The next day, the tubes were taken
out and inverted again. The resuspended incubations were then transferred to
the wells of the ELISA plate.
B) Microtiter plate method
During the testing of the lipid substances, the procedure was adapted to a 96-
well sterile, pyrogen-free microtiter plate (96well flatbottom tissue culture plate,
Becton Dickinson Labware, Meylan, Cedex, France). 200 µl of RPMI (Charles
River Endosafe) were pipetted into each well and 20 µl of sample/control and
20 µl of blood were added. Mixing was achieved by aspiring and dispensing the
incubations 4-6 times with a multichannel pipette using sterile, pyrogen-free tips
and changing the tips between the rows in order to avoid cross-contaminations.
The plate was then covered with a lid and placed in an incubator at 37°C and
5% CO2. The next day, the incubations were resuspended using a multichannel
pipette with disposable tips and transferred to the wells of the ELISA plate.
C) ELISA procedure
Cytokine ELISAs were based on commercially available antibody pairs against
IL-1.beta. (R&D, Wiesbaden, Germany). Binding of biotinylated antibody was
quantified using streptavidin-peroxidase (Biosource, Camarillo, California, USA)
and the substrate TMB (3,3’,5,5’-tetramethylbenzidine) (Sigma-Aldrich). The
substrate is transformed from colorless to blue, and the intensity is measured in
terms of optical density (OD), in our case at a wavelength of 450 nm. For final
endproduct testing, the IPT Kit provided by Charles River Endosafe was used
according to the manufacturer’s instructions.
.
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
101
Adaptation of the procedure to the lipid substances
A) Peanut oil, sesame oil, miglyol, paraffin and benzyl benzoate
The endotoxin E. coli O111: B4, serving as Gram-negative control, was diluted
either in saline (Charles River Endosafe) or in the respective oil as indicated.
Dilutions were performed either with saline or oil down to a concentration of 0.5
EU/ml. The saline dilutions served as reference for judging the interferences
caused by the oil. All dilutions were vortexed before use for about 10 seconds at
maximum speed.
B) Testing of the endproducts Testoviron®, Gynodian® and Noristerat®
The Gram-negative control was resuspended in the respective endproduct,
yielding an artificial contamination of 20 EU/ml. After a suitable diluent had been
found for pipetting dose-response curves, a protocol for a PPC (positive product
control) and an NPC (negative product control) was developed for the Gram-
negative and the Gram-positive control, respectively. Furthermore, the
procedure was optimized for the microtiter plate. Since the resuspended
lyophilisate was very viscuous, 1 ml syringes (Kendall, Wollerau, Switzerland)
for used for the dilutions and a multipette with pyrogen-free combitips was used
to pipet the required volumes of the lipid samples onto the microtiter plate. All
dilutions were vortexed before use for about 30 seconds at maximum speed.
7.4. Results
A) Peanut oil, sesame oil, miglyol, paraffin and benzyl benzoate
The protocol established for the testing of hydrophilic parenterals had to be
adapted to the lipidic drugs. In order to avoid problems caused by redistribution
of the LPS, the lyophilized LPS standards in the IPT kit were resuspended and
diluted in the oils. An additional reference curve in saline was performed to
judge the amount of pyrogen retrieved by the monocytes under these
conditions. In all following experiments, a pool of cryopreserved blood from five
individual donors was used. It turned out that the oils differed in their
characteristics with regard to interference in spike retrieval (Table 1).
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
102
With peanut oil and sesame oil, the dilutions of the reconstituted LPS in the oil
resulted in a dose-response curve with a positive signal down to 0.5 EU/ml. A
comparison to the standard curve in saline showed no interference, except for
the 0.5 EU/ml in sesame oil, which gave a higher response than the saline
reference.
Oil control
oil
0.5
EU/ml
oil
1.0
EU/ml
oil
control
saline
0.5
EU/ml
saline
1.0
EU/ml
saline
mean OD 0.1027 0.3413* 0.6507 0.08900 0.1460 0.6840 sesame
oil SD 0.009866 0.04167 0.1134 0.007550 0.03208 0.06077
mean OD 0.0850 0.9547 1.864 0.09567 0.7103 1.610 peanut
oil SD 0.004000 0.02542 0.08154 0.03326 0.1760 0.1664
meanOD 0.07167 0.3063 0.4360 0.05933 0.2773 0.3763 miglyol
SD 0.02060 0.02290 0.09032 0.00208 0.02084 0.05160
mean OD 0.07867 0.2597 0.4520 0.05933 0.2773 0.3763 paraffin
SD 0.02554 0.05952 0.1410 0.00208 0.02084 0.05160
mean OD 0.09167 0.08533 0.1240 0.09567 0.3363 1.392 castor oil
SD 0.007572 0.006506 0.06022 0.009018 0.05659 0.06558
mean OD 0.2493 0.2073 0.2047 0.3083 0.9740 2.000 benzyl
benzoate SD 0.02574 0.02434 0.01617 0.08093 0.1888 0.2435
Table 1: Interference test of six different oily preparations towards a
reference E. coli O111: B4 endotoxin curve
bold letters: negative interference: the mean OD of the sample is below 50%
the mean OD of the saline reference
positive interference: the mean OD of the sample is above 200% the mean
OD of the saline reference
The saline reference curve is the same for miglyol and paraffin, since the oils
were incubated in the same experiment. They tested interference-free at all
dilutions and were easy to pipet. This did not apply to the other substances
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
103
though, and castor oil and benzyl benzoate showed a strong interference when
compared to saline.
Part B) Endproduct testing
Since the endproducts consist mainly of castor oil and benzyl benzoate, a
strong interference occurred for all three (Fig. 1). The reference curve was
performed with saline and with sesame oil, since the latter had proven to be
advantageous with regard to high spike retrieval and very little interference
(Table 1).
0 2.5 50 .0
0 .5
1 .0
1 .5
T e s to vi r o n
G y n o d ia n
N o r is te r a t
s e s a m e o i ls a lin e
O 1 1 1 ( E U / m l)
OD
450
±± ±± S
D
Fig. 1: Dose-response curve of E. coli O111: B4 in the endproducts, in
sesame oil and saline
After modification of the testing protocol for castor oil, this oil could be tested in
the IPT. In this experiment, the castor oil was diluted with 15% endotoxin-free
DMSO (v/v ratio) and LPS dilutions were performed in the oil with the added
DMSO. The reference in saline was done with and without 15% DMSO. An
approach using the Gram-positive stimulus (LTA, lipoteichoic acid from B.
subtilis) was done in parallel.
The modified protocol brought an improvement for the castor oil, but did not
resolve the interference problem with the endproducts (data not shown),
probably due to remaining interference of the benzylbenzoate or the
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
104
pharmaceutically active substances themselves. This approach was therefore
not pursued further.
Using the approach that allowed interference-free testing with some lipids, it
was attempted to dilute the interfering drugs to concentrations at which the
interference would be diminished or entirely gone was developed. For this
purpose, the drug itself was used to reconstitute the Gram-negative standard to
a concentration of 20 EU/ml. This stock solution was then diluted with sesame
oil. Sesame oil was chosen since it is commercially available and showed
excellent pipetting characteristics in the previous experiments. At the same
time, the procedure was transferred from the tubes to the microtiter plate
incubation. The CV (coefficient of variation) was determined in order to see
whether lipids could be pipetted reproducibly at a volume of 20 µl. The CV
equals the standard deviation divided by the mean (expressed as a
percentage). This proved to be successful for all three products (Table 2).
Dose-response curve 1
Mean OD CV (%)
LPS (EU/ml) 0 1 2 4 0 1 2 4
Testoviron 0.032 1.187 1.209 1.483 6.82 8.46 9.77 6.38
Gynodian 0.033 0.804 1.236 1.400 8.65 23.84 6.65 5.39
Noristerat 0.035 0.664 1.072 1.226 35.55 34.74 4.55 13.27
Table 2a: E. coli O111: B4 was reconstituted in the endproduct and
diluted with sesame oil (dose-response curve 1).
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
105
Dose-response curve 2
Mean OD CV (%)
LPS (EU/ml) 0 1 2 4 0 1 2 4
Testoviron 0.042 1.379 1.547 1.674 66.43 5.26 5.56 4.48
Gynodian 0.029 1.290 1.557 1.664 8.16 7.59 4.18 3.26*
Noristerat 0.031 1.438 1.636 1.742 29.26 6.2 4.12 1.65
Table 2b: E. coli O111: B4 was reconstituted and diluted in saline (dose-
response curve 2).
Next, the Gram-positive standard, lipoteichoic acid (LTA), was tested with a
similar protocol. The lyophilisate was reconstituted in saline and in sesame oil,
respectively. Both the reactivity and the variance were satisfying (Fig. 3),
although the retrieval in the oil was lower.
Establishment of a positive product control (PPC) is necessary in order to prove
that a contamination would have been retrieved at a certain product dilution. For
this purpose, the diluted “clean” product is spiked with a sample from the
sesame oil reference dose-response curve and must prove interference-free,
i.e. have an OD between 50 and 200% of the signal of the same spike
concentration in the absence of the product. A PPC of 2 EU/ml at a product
dilution in sesame oil of 1:10 and one of 4 EU/ml at a product dilution of 1:5 was
done for Testoviron® and Noristerat® (Fig. 4 and 5). A corresponding negative
product control (NPC) was done accordingly, spiking with the negative control
from the sesame oil dose-response curve. The mean OD (solid line) was
calculated for the PPC 2 EU/ml as well as the 50 and 200% cut-off (dotted
lines).
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
106
0 LTA 0 LTA0.0
0.1
0.2
0.3
0.4
0.5
saline sesame oil
5.49
19.93one outlierexcluded
5.1one outlierexcluded
20.67
OD
450
Fig. 3: replicates and reactivity of the Gram-positive standard
(Lipoteichoic acid, LTA) in saline and sesame oil.
Both the PPC with 4 and with 2 EU/ml were interference-free when compared to
the dose-response curve. The PPC 2 EU/ml was chosen for end-product
testing, since it lies in the linear range of the reference curve. Additionally, in the
same experiment as in Fig. 4 and 5, a PPC for the Gram-positive control was
established at a product dilution of 1:10 for both drugs. The 1:10 dilution was
chosen since it corresponds to the product dilution of the Gram-negative PPC 2
EU/ml.
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
107
0.0
1.0
2.0
4.0
TV 1 E
U/ml
TV 2 E
U/ml
TV 4 E
U/ml
NPC
PPC 2 E
U/ml
PPC 4 E
U/ml
0.0
0.5
1.0
1.5
O111 Dose-Response insesame oil
Testoviron + O111(20 EU/ml), dilutedin sesame oil
NPC/PPCO111
200%
PPC
50%
OD
450
±± ±± S
D
LTAreferencein sesame oil
LTA in 1:10product dilution
Gram-negative PPC Gram-positive PPC
Fig. 4: Establishment of a PPC (positive product control) for the Gram-
negative and the Gram-positive stimulus: Testoviron®
0.0
1.0
2.0
4.0
1 EU/m
l
2 EU/m
l
4 EU/m
lNPC
PPC 2 E
U/ml
PPC 4 E
U/ml
0
1
2
O111 Dose-Response insesame oil
Testoviron + O111(20 EU/ml), dilutedin sesame oil
NPC/PPCO111
200%
PPC
50%
LTAreferencein sesame oil
LTA in 1:10product dilution
Gram-negative PPC Gram-positive PPC
OD
450
±± ±± S
D
Fig. 5: Establishment of a PPC (positive product control) for the Gram-
negative and the Gram-positive stimulus: Noristerat®
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
108
A commercially available sesame oil preparation (Sigma-Aldrich) was tested to
determine whether it would be a suitable diluent for the preparation of an LPS
dose-response curve. For this purpose, the reaction was compared to that of an
E. coli O113: H10 WHO standard dose-response curve in saline. An O111
dose-response curve in saline was set up in parallel. With this sesame oil
preparation, negative interference occurred, but the reactivity of O111 in
sesame oil and O113 in saline was very similar (Fig. 6).
0 1 2 40.0
0.5
1.0
1.5O111 in salineO111 in sesame oilO113 in saline
O111 (EU/ml)
OD
450
Fig. 6: Comparison of the activity of LPS from E. coli O111: B4 in saline
and in sesame oil with the E. coli O113: H10 WHO standard in saline
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
109
With these results, it was concluded that the O111 saline curve would not be
suitable for judging lipidic endproducts, since the pyrogenic contaminations
would be underestimated. The similarity of the O111 in sesame oil to the WHO
standard allows the conclusion that the legend “EU/ml” is adequate for
endproduct testing. Since it is useful to set up the reference in the same diluent
used for the sample, the O111 dose-response curve was from then on
performed in sesame oil in order to serve as reference for product testing.
Additionally, the following minimum assay suitability requirements for the testing
of the endproducts were determined:
A) The 2.0 EU/ml concentration of the Gram-negative standard in sesame oil
must test positive (more than 1.6 times the mean OD of the negative sesame oil
control). In the In vitro Pyrogen Test, a value higher than 1.6 times the mean of
the corresponding negative control is considered significantly positive. The
factor 1.6 derives from the average cv (below 20%), equal to 0.2 in the
calculation. This 0.2 value is multiplied by 3 in order to create a broad safety
margin, should the cv be higher than average.
B) The Gram-positive (LTA) control must test positive (more than 1.6 times the
mean of the negative sesame oil control).
C) The 2.0 EU/ml PPC of the Gram-negative control must be in the
interference-free (50-200%) range of the 2.0 EU/ml concentration of the dose-
response curve.
D) The Gram-positive PPC must be in the interference-free range of the Gram-
positive control in sesame oil.
E) The OD of the negative sesame oil control as well as the NPC must not
exceed 0.1.
Applying these criteria, all three end products were tested three times
independently and with three different lots of cryopreserved blood in the IPT.
Results are summarized in Table 3.
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
110
Quality criteria (A-E) met Experiment
Number Testoviron Gynodian Noristerat
1 Yes Yes Yes
2 Yes Yes Yes
3 No* Yes Yes
* Criterion B not met (Gram-positive control did not test positive), therefore
criterion D could not be met as well
Table 3: Results of the final testing of all three substances in the IPT
7.5. Discussion
In this study, the strongly varying characteristics of different oily preparations
concerning their interference with the human whole blood test based on pooled
cryopreserved blood were assessed with altogether five oils and three
endproducts. While most oils tested interference-free when compared to an
LPS dose-response curve in saline, others caused strong negative interference.
Interference testing is one of the most critical aspects of pyrogen testing. First of
all, it requires the availability of a pyrogen-free reference material, which is often
difficult to obtain or to judge, especially in the case of negative interference in
the respective pyrogen test. The lipid A portion as the active part of the LPS
molecule is integrated into lipid layers, e.g. into the membranes of liposomes
and reduces IL-1.beta. production (173). Therefore, a strong interference with
the pyrogen test was expected, but the oils as biological products showed
different characteristics that appeared to be independent of integration and
inactivation of LPS. No solution could be found for the substance benzyl
benzoate. In principle, all oils that are interference-free in the test are suitable
as diluents for more complicated preparations that cannot be tested in their
undiluted form.
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
111
Taking into account the fever threshold of humans at 5 EU/kg (48) on which
ELC calculations are based, the maximum contamination that can be tolerated
was calculated. Since the endproducts are applied at a dose of 1 ml
intramuscularly and assuming an average body weight of 70 kg, a theoretical
contamination of 350 EU/ml would still be non-pyrogenic. It should be noted that
this criterion was set for intravenous drugs which might suggest an even higher
possible ELC. In general, these calculations suggest that most end-products
can be controlled at the respective maximum valid dilution. With a protocol
allowing 20 EU/ml to be detected, a broad safety margin could be achieved.
The calibration of the O111 LPS in sesame oil against the international WHO
endotoxin standard from E. coli O113 (56) showed that “endotoxin unit”
(EU/ml) is appropriate in the case of dose-response curves in lipids. At the
same time, this experiment showed that the pyrogenic activity of the O111 in
saline would have been overestimated. This led to the establishment of the
reference dose-response curve in sesame oil. In contrast to the rabbit test, a
positive product control as well as a negative control could be established;
the successful retrieval of the former serving as proof that a contamination of
the “clean” drug would have been detected at the respective dilution. A Gram-
positive control and a corresponding PPC could be established as well.
Furthermore, the distribution of LPS plays a major role. A homogenous
redistribution of LPS in lipidic products could be demonstrated by calculating the
coefficient of variation, thus proving that the lipid samples can be pipetted in a
satisfying way even at very small volumes (20 µl in the microtiter plate). For all
experiments, cryopreserved human whole blood was used (73). An interesting
result was that the reactivity of the three lots of cryopreserved blood differed in
their response to the Gram-positive stimulus. In endproduct testing, one lot of
cryopreserved blood failed to detect the Gram-positive control in the testing of
the Testoviron®; while when was used to test the other two endproducts, the
signals were very low, though positive. The other two lots were highly sensitive
towards this stimulus. The reactivity towards the Gram-negative controls was far
more homogenous in all experiments. This stresses the importance of prior
testing before releasing the cryopreserved blood by the manufacturer and
PYROGEN TESTING OF LIPIDIC PARENTERALS WITH A NOVEL IN
VITRO TEST
112
indicates at the same time that the individual sensitivities towards Gram-positive
contaminations are much higher than those for LPS.
Lipidic substances in general remain a challenge for the IPT. The established
protocol has been optimized for the given endproducts, but can be modified,
e.g. concerning the diluent and/or the necessary dilutions, in order to meet the
requirements of other endproducts. Large volume parenterals for total
parenteral nutrition (TPN), are particularly difficult to test, since, apart from
containing fats, they also require an extremely low detection limit for successful
pyrogen testing. Further interesting preparations to be examined in the IPT are
liposomes. These are used as vehicles for drugs, releasing the drug at or
nearby the desired site of action where they are degraded and could release
possible pyrogenic contaminations.
7.6. Conclusion
This study shows that contaminations of lipidic formulations of drugs cause a
strong and sensitive reaction of human monocytes, indicating that they have the
potential to be a health hazard to patients receiving these drugs.
Taking this into account, this study provides compelling evidence that the IPT
assay can overcome most of the problems associated with the control of lipidic
endproducts. It therefore offers a viable alternative to testing these preparations
in rabbits.
SUMMARIZING DISCUSSION
113
8 Summarizing discussion
The desire to make available an in vitro pyrogen test as an alternative method
to the rabbit pyrogen test is prompted by ethical considerations of animal
protection as well as to matters of patient safety. It is known long that the raise
in temperature in the rabbit represents an unreliable parameter due to stress-
induced hyperthermia, hypothermia (29), sensitivity differences of rabbit strains
(27), age, sex and even between individuals of the same breeds who have
been found to be high, low, and intermediate responders (174). Therefore, the
LAL was introduced in pyrogen testing in the 70s and is now widely used. The
major limitations of this test is, as its alternative name bacterial endotoxin test
(BET) implies, being limited to the pyrogens derived from Gram-negative
bacteria and missing known pyrogens as lipoteichoic acid and peptidoglycans
as well as exotoxins and other non-endotoxin pyrogens. Pyrogenic reactions in
healthy volunteers have been reported during a clinical study when a
preparation of human growth hormone was injected that had tested negative in
the LAL and in the rabbit (133). Adverse reactions after the application of
vaccine against early summer meningoencephalitis have been observed,
although the LAL had tested negative (59). An assay capable of detecting all
relevant pyrogens should mirror closely the physiological events which take
place in the mammalian organism during a pyrogenic reaction.
Upon contact with so-called pathogen-associated molecular patterns (PAMPS)
(127), e.g. bacterial, fungal or viral components, the blood monocytes are
activated. The Cluster of Differentiation (CD) 14 receptor, a glycoprotein whose
significance was recognized in 1990, exists in a membrane-bound (mCD14, a
53 kDa protein) and a soluble form. The former is embedded in the plasma
membrane, the latter, sCD14, is capable of inhibiting LPS activity (175), but
additionally LPS/sCD14 complexes can induce biological responses in certain
CD14-negative cells such as endothelium (176). LPS binding takes place via
the Lipid A component and is of high affinity (177).
SUMMARIZING DISCUSSION
114
The innate immune system of mammals uses the family of toll-like receptors
(TLR) to engage microorganisms by recognizing PAMPS (178). For the
recognition of LPS, the TLR4 receptor is crucial (179). The role of TLR4 was
determined in TLR4 deficient mice which proved to be unresponsive to LPS
(180). Upon binding, the TLR4 receptor dimerizes and the intracellular signal
cascade is initiated.
In contrast, the pyrogen LTA, but also peptidoglycans and lipoproteins, bind to
the TLR2 (181, 182) and cause fever along the same lines as LPS. The
receptor engagement activates intracellular signaling cascades which lead to
the formation of proinflammatory cytokines such as Interleukin-1β (IL-1β), IL-6
and others by the blood monocytes. The IPT measures the endpoint IL-1β.
The IL-1β exists in the mononuclear cells as a preformed molecule, the pro-IL-
1β. Upon stimulation, a protease is activated, the IL-1 Converting Enzyme
(ICE), whose catalytic function is essential for the generation of mature,
extracellular IL-1β (183). It is an interesting phenomenon that in studies with
human subjects, no IL-1β could be found in the blood stream after LPS
injection or in naturally occurring sepsis, whereas IL-6 was always detectable
(184-186). Messengers this potent are probably quickly bound to their
corresponding receptors, e.g. to the soluble Interleukin-1 receptors (187), or
internalized, e.g. by binding to/crossing the blood-brain barrier (188, 189),
making them disappear fast from the blood serum. The individual roles of the
proinflammatory cytokines in the generation of fever are still controversially
discussed, but several lines of evidence suggest that IL-1β is indeed the most
important and potent inducer of fever. Several studies in humans (190-193)
have demonstrated the pyrogenic potency of this cytokine and studies in rabbits
have shown clearly that the doses of IL-1β required are 100fold lower than
those of IL-6 and around 10fold lower than those of TNF-α (46, 47).
Furthermore, the cytoplasmic domains of TLRs and the IL-1β receptor share
the same signaling areas (38). The necessary intracellular enzymatic cleavage
of pro-IL-1β might provide an additional control mechanism of a mediator which
is crucial in the genesis of pyrogenic reactions.
These events in the mammalian organism provided the idea for the principle of
the whole blood test. When in contact with a pyrogen, the monocyte, a
SUMMARIZING DISCUSSION
115
subfraction of the white blood cells, reacts with the production of
proinflammatory cytokines, one of which is IL-1β. By using the whole blood as a
reagent, the monocyte is left in its natural environmant with all components
necessary for a reaction that is as close to the in vivo situation as possible.
IL-1β was chosen as a readout since it is a very stable monomer, which
survives repeated freezing-thawing processes, something that might be
necessary during the establishment of the system and for research purposes,
and since it is highly regulated, which might prevent or reduce false-positive
reactions or basal levels in unstimulated samples. Furthermore, IL-1β has been
recognized as the most potent cytokine when injected in vivo (47, 190-193) and
appears to play a crucial, if not the crucial role in the pathogenesis of fever.
In order to make the new IPT available on a large scale and therefore a true
alternative to animal experiments, all reagents had to be standardized and
certified. A major issue was the certification and standardization of the human
blood. Freshly drawn human blood cannot be stored for more than 4 hours until
the IL-1 formation declines, and is difficult to measure. Individual differences if
not in the sensitivity, but the amount of cytokine produced can make
interpretation of the results difficult. Furthermore, a risk of infection for the user
with e.g. hepatitis cannot be excluded easily. In the first part of this work, a
protocol for pooling and freezing whole blood was developed which maintains
all the characteristics of the fresh blood. The cryopreservation of isolated blood
cells has been successfully performed since the 1960s for transfusion and
research purposes (73, 194-197). By freezing the entire blood without any
isolation procedures, the monocyte is maintained in its natural environment with
all the known and unknown factors that might influence the response, thus
maintaining their reactivity and specificity with regard to pyrogen testing. An
important plasma protein is for example the LPS binding protein (LBP) which in
the blood of healthy donors presents the LPS to the CD14 receptor (162, 163).
While red blood cells require very high freezing rates in order to remain intact,
the white blood cells are best frozen at rates of 1-2°C/min. This is due to the
process of exosmosis of intracellular water which depends on parameters like
cell surface/volume ratio and membrane permeability and which was described
by Mazur in 1963, 1965 and 1977 (198-200). Our protocol therefore had to
SUMMARIZING DISCUSSION
116
tolerate the complete destruction of the erythrocytes with the risk, that the
reaction of the monocytes would be influenced by the cell detritus. This was, at
least in our experiments, never the case. The procedure involved clinical,
endotoxin-free DMSO at a 10% final concentration in the blood (v/v ratio) and a
computer-controlled freezing process using liquid nitrogen. For higher
standardization and in order to level out individual differences in the response,
the blood of 5 donors was pooled and frozen together. The frozen blood proved
to be a highly sensitive and robust reagent with a very high inter-lot
comparability. The possibility of pretesting the donors in question for infectious
agents eliminates the abovementioned health hazard for the personnel.
In order to make the IPT available for routine application, the assay using
cryopreserved blood, allowing storage at -80°C as compared to liquid nitrogen
as well as an improved variation of the fresh blood was validated in an
international collaborative study which followed the procedure in which the
fresh blood had been validated previously in detail. The validation included
three laboratories working under good laboratory practice (GLP) with 10
substances and altogether 50 blinded endotoxin spikes at or around the
pyrogenic threshold of 0.5 EU/ml (16). The IPT achieved sensitivities around
90% and specificities around 80%. The cryopreserved blood described in part
one had a lower sensitivity of 80% with at the same time the best specificity.
Compared to the former study (48), the whole blood assay could be improved
regarding consumer-friendlyness as well as performance. Based on this
outcome, the inclusion of the IPT into the Pharmacoopoeias should be
possible.
In the last part of this work, the validated assay using the cryopreserved blood
was adapted to suit a special application. So far, only hydrophilic substances
mainly for intravenous administration had been used in the test. A change in
regulation in 2004 by European Pharmacopoeia made the testing of so-called
small volume parenterals (SVP) obligatory that had so far not been subjected to
pyrogen testing. Suddenly, several lipophilic parenterals, e.g. with fat-soluble
vitamins or hormones, for intramuscular and subcutaneous application were
SUMMARIZING DISCUSSION
117
concerned. Testing lipophilic substances in the rabbit by intravenous injection
into the ear vein which is the accepted procedure, is extremely dangerous since
the insoluble miniscule drops can clot capillaries in vital organs such as the
kidneys. Apart from the viewpoint of animal protection, the outcome of such an
experiment must be doubtful. The LAL, on the other hand, is impeded by the
fact that the pyrogenic portion, the lipid A, is masked by lipid substances, lipidic
parenterals (122), lipoproteins (121, 168), and liposomes (173) and therefore
no longer accessible to the components of the coagulation cascades, leading to
an underestimation of the pyrogenic contamination. It was therefore a pressing
question whether such substances can be controlled by the IPT. For this
purpose, the standard protocol was modified and different pure oils as well as
three endproducts were measured. It turned out that all products can be
controlled, although the pyrogenic threshold of 0.5 EU/ml for hydrophilic
substances could not be maintained in this product group. Still, since the drugs
are given at a very small volume, a higher contamination can be tolerated
which will predictably not cause any adverse reactions in the recipient.
Taken together, with the new assay based on human whole blood, a
standardized, reliable and highly sensitive method based on the human fever
reactions and measuring all relevant pyrogens is now available for widespread
use. We hope that the replacement of the rabbit pyrogen test by this assay is
therefore only a question of time, simultaneously maintaining and even
exceeding the already high level of safety for patients receiving parenterals.
ZUSAMMENFASSUNG
118
9 Summary
The detection of bacterial contaminations of all origins in parenterals has been
recognized as an important issue. In 1943, the in vivo rabbit pyrogen test was
introduced into the US Pharmacopoeia and has been used since then. In the
70s, another pyrogen test, the limulus amoebocyte lysate test (LAL) was
introduced which uses the lysates of blood cells of the horseshoe crab and
cannot measure a variety of common pyrogens nor distinguish between the
potencies of given pyrogens in the mammal. The immune system is able to
respond to pathogens with production and secretion of cytokines. In 1995, a
novel in vitro alternative based on human whole blood which uses this reaction
for the detection of all possible kinds of microbial contaminations has been
developed. Its reliability and wide spectrum of application possibilities make it a
promising candidate for entirely replacing the rabbit test.
• As a first step, the human whole blood test using freshly drawn human
whole blood was validated in an international collaborative study.
• The most critical reagent of the new test, the human whole blood, was made
available by developing a protocol for pooling and cryopreserving the fresh
blood, at the same time maintaining the characteristics of the fresh blood
regarding the reactivity towards all relevant pyrogens.
• This blood was then validated in an international study in order to be able to
include the assay into Pharmacopoeias.
• With this validated procedure, the applications of the test were extended
from hydrophilic to lipophilic parenterals meeting new requirements of
European Pharmacopoeia.
With all these measures, the new pyrogen assay based on whole blood which
measures all relevant pyrogens inducing a fever reaction is ready for routine
use. This work adds to the replacement of an important and widely used animal
test while at the same time maintaining, even superceding the existing safety
standards and extending the possibilitites of applications beyond those of the
rabbit test and even the LAL.
ZUSAMMENFASSUNG
119
10 Zusammenfassung
Die Detektion von bakteriellen Kontaminationen jeglicher Herkunft in
Parenteralia ist als wichtiges Problem erkannt. 1943 wurde der in vivo
Kaninchenpyrogentest in die US Pharmacopoe eingeführt und wird seither
verwendet. In den 1970er Jahren wurde ein weiterer Pyrogentest, der Limulus
Amoebozyten Lysattest (LAL) eingeführt, der das Lysat der Blutzellen der
Hufeisenkrabbe einsetzt und der weder alle relevanten Pyrogene messen noch
zwischen ihrer jeweiligen biologischen Potenz im Säugetier unterscheiden
kann. Das Immunsystem antwortet auf Pathogene mit der Produktion und
Sekretion von Zytokinen. 1995 wurde eine in vitro Alternative mit humanem
Vollblut entwickelt, die diese Reaktion zur Detektion aller möglichen
mikrobiellen Kontaminationen nutzt. Ihre Zuverlässigkeit und das weite
Spektrum an Anwendungsmöglichkeiten machen sie zu einem
vielversprechenden Kandidaten, den Kaninchentest vollständig zu ersetzen.
• Als ein erster Schritt wurde der Test, basierend auf frisch abgenommenem
humanem Vollblut, in einer internationalen Studie validiert.
• Der wohl kritischste Bestandteil des Tests, das menschliche Vollblut, wurde
verfügbar gemacht, indem ein Protokoll zum Mischen und Einfrieren von
frischem Blut entwickelt wurde. Gleichzeitig blieben die Eigenschaften des
frischen Blutes in Bezug auf die Reaktivität gegenüber allen relevanten
Pyrogenen erhalten.
• Parallel zu dem Frischblut wurde dieses Blut in einer internationalen Studie
validiert, um den Test in die Pharmakopoen einführen zu können.
• Mit dieser validierten Methode wurden die Anwendungen für den Test
erweitert, um den neuen Anforderungen der Europäischen Pharmacopoe
Genüge zu tun: von hydrophilen zu lipophilen Parenteralia.
Damit ist der neue Pyrogentest auf Vollblutbasis, der alle relevanten
fiebererzeugenden Pyrogene mißt, zum routinemäßigen Einsatz bereit. Die
vorliegende Arbeit trägt dazu bei, einen wichtigen und viel genutzten
Tierversuch zu ersetzen, gleichzeitig die bestehenden Sicherheitsstandards
beizubehalten bzw. zu übertreffen und die Anwendungsmöglichkeiten über die
des Kaninchentests und sogar des LAL hinaus auszuweiten.
REFERENCES
120
11 References
1. Semmelweis, I. P. (1861). Die Ätiologie der Begriff und die Prophylaxis des
Kindbettfiebers. C. A. Hartleben´s Verlags-Expedition: Pest
2. Godlee, R. J. (1917). Lord Lister, London
3. von Haller, A. (1757-66). Elements physiologiae corporis humanae Vol III
4. Panum, P. L. (1874). Das putride Gift, die bakterien, die putride infektion
oder intoxikation und die septikämie
Arch Pathol Anat Physiol Klein Med (Virchow´s Arch) 60: 301
5. Billbroth, T. (1862). Beobachtungsstudien über das wundfieber und
accidentelle wund-krankheiten. Arch Klin Chir 2: 578
6. Centanni, E. (1921). Trattato di Immunologia
Societa Editrice Libraria, Rome: 149
7. Hort, E., Penfold, W. (1912). Microorganisms and their relation to fever. J
Hyg 12: 361
8. Seibert, F. B. (1925). The cause of many febrile reactions following
intravenous injection. Am J Physiol 71: 621
9. Rademaker, L. (1933)
The cause and elimination of reactions after intravenous infusions
Surg Gynaecol Obstet 56: 956
10. Welsh, H., Calvery, H. O., McClosky, W. T., and Price, C. W. (1943)
Method of Preparation and Test for Bacterial Pyrogen
J Am Pharm Assoc 3: 65-69
REFERENCES
121
11. McClosky, W. T., Price, C. W., van Winkle, W. J., Welch, H., Calvery, H.
O. (1943). Results of the first USP collaborative study of pyrogens.
J Am Pharm Assoc 32: 69
12. Nowotny, A. (1983). Shedding of bacteria. Biomembranes 11: 1-20
13. Galanos, C., Rietschel, E. T., Luderitz, O., Westphal, O., Kim, Y. B.,
Watson, D. W. (1972). Biological activities of lipid A complexed with bovine-
serum albumin. Eur J Biochem 31: 230-3
14. Rietschel, E. T., Kim, Y. B., Watson, D. W., Galanos, C., Luderitz, O., and
Westphal, O. (1973). Pyrogenicity and immunogenicity of lipid A complexed
with bovine serum albumin or human serum albumin. Infect Immun 8: 173-7
15. U. S. Food and Drug Administration (1980)
Human and veterinary drugs; availability of draft guidelines for use of Limulus
amoebocyte lysate. Fed Reg 45: 3668
16. Hoffmann, S., Luederitz-Puechel, U., Montag-Lessing T., and Hartung, T.
(2005). Optimisation of pyrogen testing in parenterals according to different
pharmacopoeias by probabilistic modelling. J Endotoxin Res 11: 25-31
17. Bodel, P.T., and Atkins, E. (1965). Studies in staphylococcal fever.
V. Staphlococcal filtrate pyrogen. Yale J Biol Med 38(3): 282-98
18. Brunson, K.W., and Watson, D.W. (1974)
Pyrogenic specificity of streptococcal exotoxins, staphylococcal enterotoxin,
and gram-negative endotoxin. Infect Immun 10: 347-51
19. Watson, D. W. (1960). Host-parasite factors in group A streptococcal
infections: pyrogenic and other effects of immunologic distinct exotoxins
related to scarlet fever toxins . J Exp Med 111: 255-284
REFERENCES
122
20. Atkins, E., and Huang W. C. (1958). Studies on the pathogenesis of fever
with influenza virus. I. The appearance of an endogenous pyrogen in the
blood following intravenous injection of virus. J Exp Med 107: 383-401
21. Atkins, E., and Morse, S. I. (1967). Studies in staphylococcal fever. VI.
Responses induced by cell walls and various fractions of staphylococci and
their products. Yale J Biol Med 39: 297-311
22. Rotta, J. (1975). Endotoxin-like properties of the peptidoglycan.
Z Immunitätsforsch Exp Klin Immunol 149: 230-44
23. Schleifer, K. H. (1975). Chemical structure of the peptidoglycan, its
modifiability and relation to the biological activity.
Z Immunitätsforsch Exp Klin Immunol 149: 104-17
24. Kobayashi, G. S., and Friedman, L. (1964). Characterization of the
pyrogenicity of Candida albicans, Saccharomyces cervisiae, and
Cryptococcus neoformans. J Bacteriol 88: 660-6
25. Braude, A. I., McConnell, J., and Douglas, H. (1960). Fever from
pathogenic fungi. J Clin Invest 39: 1266-76
26. Morath, S., Geyer, A., and Hartung, T. (2001). Structure-function
relationship of cytokine induction by lipotechoic acid from Staphylococcus
aureus. J Exp Med 193: 393
27. van Dijck, P. and H. van de Voorde (1977). Factors affecting pyrogen
testing in rabbits. Dev Biol Stand 34: 57-63
28. Deeter, L. B., L. W. Martin, and Lipton, J. M. (1989).
Age- and sex-related differences in febrile response to peripheral pyrogens in
the rabbit. Gerontology 35: 297-304.
REFERENCES
123
29. Grant, R. (1950). Emotional hypothermia in rabbits. Am J Physiol 160:
285
30. Greisman, S. E. and Hornick, R. B. (1969). Comparative pyrogenic
reactivity of rabbit and man to bacterial endotoxin. Proc Soc Exp Biol Med
131: 1154
31. Levin, L. and Bang, F. B. (1964). A description of cellular coagulation in
the limulus. Bull Johns Hopkins Hosp 115: 337-45
32. Levin, J. and Bang, F. B. (1964). The role of endotoxin in the
extraacellular coagulation of limulus blood.
Bull Johns Hopkins Hosp 115: 265-74
33. Yin, E. T., Galanos, C., Kinsky, S., Bradshaw, R. A., Wessler, S., Luderitz,
O., and Sarmiento, M. E. (1972). Picogram-sensitive assay for endotoxin:
Gelation of Limulus polyphemus blood cell lysate induced by purified
lipopolysaccharides and lipid A from gram-negative bacteria
Biochem Biophys Acta 26: 284
34. Devleeschouwer, M. J., Cornil, M. F., and Dony, J. (1985). Studies on the
sensitivity and specificity of the Limulus amebocyte lysate test and rabbit
pyrogen assays. Appl Environ Microbiol 50: 1509-11
35. Wildfeuer, A., Heymer, B., Schleifer, K. H., and Haferkamp, O. (1974)
Investigations on the specificity of the Limulus test for the detection of
endotoxin. Appl Microbiol 28: 867-71
36. Beeson, P. B. (1948). Temperature-elevating effect of a substance
obtained from polymorphonuclear leucocytes (abstract)
J Clin Invest 27: 524
REFERENCES
124
37. Dinarello, C. A., Goldin, N P., and Wolff, S. M. (1974). Demonstration and
characterization of two distinct human leukocytic pyrogens. J Exp Med 139:
1369-81
38. Gay, N. J., and Keith, F. J. (1991). Drosophila Toll and IL-1 receptor.
Nature 351(6325): 355-6
39. Li, S., Wang, Y., Matsumura, K., Ballou, L R., Morham, S. G., and
Blatteis, C. M. (1999). The febrile response to lipopolysaccharide is blocked in
cyclooxygenase-2 (-/-), but not in cyclooxygenase-1(-/-) mice.
Brain Res 825: 86-94
40. Li, S., Ballou, L. R., Morham, S. G., and Blatteis, C. M. (2001)
Cyclooxygenase-2 mediates the febrile response of mice to interleukin-1beta.
Brain Res: 163-73
41. Li, S., Goorha, S., Ballou, L. R., and Blatteis, C. M. (2003).
Intracerebroventricular interleukin-6, macrophage inflammatory protein-1 beta
and IL-18: pyrogenic and PGE(2)-mediated? Brain Res 992: 76-84
42. Chai, Z., Gatti, S., Toniatti, C., Poli, V., and Bartfai, T. (1996)
Interleukin (IL)-6 gene expression in the central nervous system is necessary
for fever response to lipopolysaccharide or IL-1 beta: a study on IL-6-deficient
mice. J Exp Med 183: 311-6
43. Ushikubi, F., Segi, E., Sugimoto, Y., Murata, T., Matsuoka, T., Kobayashi,
T., Hizaki, H., Tuboi, K., Katsuyama, M., Ichikawa, A., Tanaka, T., Yoshida,
N., and Narumiya, S. (1998). Impaired febrile response in mice lacking the
prostaglandin E receptor subtype EP3. Nature 395 (6699): 281-4
44. Dinarello, C. A., Gatt, S., and Bartfai, T. (1999)
Fever: links with an ancient receptor
Curr Biol 9: R147-50
REFERENCES
125
45. Laburn, H. P., Rosendorff, C., Willies, G., and Woolf, C. (1974)
Proceedings: A role for noradrenaline and cyclic AMP in prostaglandin E1
Fever. J Physiol 240: 49P-50P
46. Dinarello, C. A., Cannon, J. G., Wolff, S. M., Bernheim, H. A., Beutler, B.,
Cerami, A., Figari, I. S., Palladino, M. A., and O'Connor, J. V. (1986).
Tumor necrosis factor (cachectin) is an endogenous pyrogen and induces
production of interleukin 1. J Exp Med 163:1433-50
47. Dinarello, C. A., Cannon, J. G., Mancilla, J., Bishai, I., Lees, J., and
Coceani, F. (1991). Interleukin-6 as an endogenous pyrogen: induction of
prostaglandin E2 in brain but not in peripheral blood mononuclear cells.
Brain Res 562:199-206
48. Hoffmann, S., Peterbauer, A., Schindler, S., Fennrich, S., Poole, S.,
Mistry, Y., Montag-Lessing, T., Spreitzer, I., Löschner, B., van Aalderen, M.,
Bos, R., Gommer, M., Nibbeling, R., Werner-Felmayer, G., Loitzl, P., Jungi,
T., Brcic, M., Brugger, P., Frey, E., Bowe, G., Casado, J., Coecke, S., de
Lange, J., Mogster, B., Naess, L. M., Aaberge, I. S., Wendel, A., and Hartung,
T. (2005). International validation of novel pyrogen tests based on human
monocytoid cells. J Immunol Methods: 161-73
49. Hartung, T., and Wendel, A. (1995). Detection of Pyrogens using human
whole blood. ALTEX 12: 70-75
50. Rudloe A, Hernkind A. E. (1983). The effect of heavy bleeding on mortality
of the horseshoe crab, Limulus polyphemus, in the natural environment.
J Invertebr Pathol 42: 167–176
51. Thompson, M.(1998). Assessments of the population biology and critical
habitat for the horseshoe crab, Limulus polyphemus, in the South Atlantic Bight
Master’s thesis Medical University of South Carolina, University of Charleston.
REFERENCES
126
52. Walls E. A., Berkson, J. (2003). Effects of blood extraction on horseshoe
crabs (Limulus polyphemus). Fish Bull 101. 457-459
53. Fennrich, S., Fischer, M., Hartung, T., Lexa, P., Montag-Lessing, T.,
Sonntag, H. G., Weigandt, M., Wendel, A (1999). Detection of endotoxins and
other pyrogens using human whole blood. Dev Biol Stand 101: 131-139
54. Bennett, I. L., Beeson, P. B. (1953). Studies on the pathogenesis of fever.
II. Characterization of fever-producing substances from polymorphonuclear
leukocytes and from the fluid of sterile exsudates. J Exp Med 98: 493-508
55. Dinarello, C. A. (2004). Infection, fever, and exogenous and endogenous
pyrogens: some concepts have changed. J Endotoxin Res 10: 201-22
56. Poole, S., Dawson, P., Gaines Das, R. E. (1997). Second international
standard for endotoxin: calibration in an international collaborative study.
J Endotoxin Res 4: 221-231
57. Morath, S., Geyer, A., Spreitzer, I., Hermann, C., Hartung, T. (2002).
Structural decomposition and heterogeneity of commercial lipoteichoic acid
preparations. Infect Immun 70 (2): 938-44
58. Schindler, S., Bristow, A., Cartmell, T., Hartung, T., Fennrich, S. (2003).
Comparison of the reactivity of human and rabbit blood towards pyrogenic
stimuli. ALTEX 20: 59-63
59. Fischer, M., Keller-Stanislawski, B., Schober-Bendixen, S., Schosser, R.,
Hacke, K., Hartung, T., Montag, T. (2001). Effect of the preservative thiomersal
on the release of interleukin-1 beta from human peripheral blood cells.
ALTEX 18 (1): 47-9.
REFERENCES
127
60. Carlin, G., Viitanen, E. (2003). In vitro pyrogenicity of a multivalent vaccine:
Infanrix. Pharmeuropa 15: 418-23.
61. Carlin, G., Viitanen, E. (2005). In vitro pyrogenicity of the diphtheria, tetanus
and acellular pertussis components of a trivalent vaccine. Vaccine 23: 3709-15.
62. Steere, A. C., Rifaat, M. K., Seligmann, E. B., Hochstein, H. D., Friedland,
G., Dasse, P., Wustrack, K. O., Axnick, K. J., Barker, L. F. (1978). Pyrogen
reactions associated with the infusion of normal serum albumin (human).
Transfusion 18 (1): 102-7
63. Pool, E. J., Johaar, G., James, S., Petersen, I., Bouic, P. (1998). The
detection of pyrogens in blood products using an ex vivo whole blood culture
assay. J Immunoassay 19: 95-111
64. Spreitzer, I., Fischer, M., Hartzsch, K., Luderitz-Puchel, U., Montag, T.
(2002). Comparative study of rabbit pyrogen test and human whole blood assay
on human serum albumin. ALTEX 19 Suppl 1: 73-5
65. Eperon, S., Jungi, T. W. (1996). The use of human monocytoid lines as
indicators of endotoxin. J Immunol Methods 194: 121-129
66. Werner-Felmayer, G., Baier-Bitterlich, G., Fuchs, D., Hausen, A., Murr, C.,
Reibnegger, G., Werner, E. R., Wachter, H. (1995). Detection of Bacterial
Pyrogens on the Basis of their Effects on Gamma Interferon-Mediated
Formation of Neopterin or Nitrite in Cultured Monnocyte Cell Lines. Clin. Diagn.
Lab. Immunol. 2 (3): 307-313
67. Peterbauer, A., Werner-Felmayer, G. (1999). Further development of a cell
culture model for the detection of bacterial pyrogens. ALTEX 16: 3-8
REFERENCES
128
68. Ziegler-Heitbrock, H. W., Thiel, E., Futterer, A., Herzog, V., Wirtz, A.,
Riethmuller, G. (1988). Establishment of a human cell line (Mono Mac 6) with
characteristics of mature monocytes. Int J Cancer 41: 456-461
69. Hartung, T., Fennrich, S., Fische, M., Montag-Lessing, T., Wendel, A.(2000).
Prevalidation of an alternative to the rabbit pyrogen test based on human whole
blood. In: Progress in the reduction, refinement and replacement of animal
experimentation. Balls, M., van Zeller, A. M., Halder, M. E.
2000 Elsevier Science B. V. 2000, pp 991-99
70. Hartung, T., Aaberge, I., Berthold, S., Carlin, G., Charton, E., Coecke, S.,
Fennrich, S., Fischer, M., Gommer, M., Halder, M., Haslov, K., Jahnke, M.,
Montag-Lessing, T., Poole, S., Schechtman, L., Wendel, A., Werner-Felmayer,
G. (2001). Novel pyrogen tests based on the human fever reaction. The report
and recommendations of ECVAM workshop 43. European Centre for the
Validation of Alternative Methods. Altern Lab Anim 29 (2): 99-123
71. Hartung, T. (2002). Comparison and validation of novel pyrogen tests based
on the human fever reaction. Altern Lab Anim 30 Suppl 2: 49-51
72. De Boer, M., Reijneke, R., Van de Griend, R. J., Loos, J. A., Roos, D.
(1981). Large-scale purification and cryopreservation of human monocytes.
J Immunol Methods 43 (2): 225-39
73. Schindler, S., Asmus, S., von Aulock, S., Wendel, A., Hartung, T., Fennrich,
S. (2004). Cryopreservation of human whole blood for pyrogenicity testing.
J Immunol Methods 294 (1-2): 89-100
74. Schindler, S., Spreitzer, I., Löschner, B., Hoffmann, S., Hennes, K., Halder,
M., Brügger, P., Frey, E., Hartung, T., Montag-Lessing, T. (2006). International
validation of pyrogen tests based on cryopreserved human primary blood cells.
J Immunol Methods in print
REFERENCES
129
75. Association for the Advancement of Medical Instrumentation
AAMI Standards and Recommended Practices, vol 3
American National Standards Institute, Arlington 3: 1-332, 2001
76. Shmunes , Darby, T. (1984). Contact dermatitis due to endotoxin in
irradiated latex gloves. Contact Dermatitis 10: 240-4
77. Kure, R., Grendahl, H., Paulssen, J. (1982). Pyrogens from surgeons' sterile
latex gloves. Acta Pathol Microbiol Immunol Scand [B] 90: 85-8
78. Groetsch, W., Leimbach, R., Sonnenschein, B. (1992). On the safety of
medical products: The detection of endotoxin on sterile surgical gloves.
Hyg Med 17: 200-206
79. Mazzotti, F., Beuttler, J., Zeller, R., Fink, U., Schindler, S., Wendel, A.,
Hartung, T., von Aulock, S. (2006). In vitro Pyrogen Test - a new test method for
solid medical devices. J Biomed Mater Res Part A, in print
80. Anderson, J. M., Miller, K. M. (1984). Biomaterials biocompatibility and the
macrophage. Biomaterials 5: 5-10
81. Ziats, N. P., Miller, K. M., Anderson, J. M. (1988). In vitro and in vivo
interactions of cells with biomaterials. Biomaterials 1988; 9: 5-13
82. Rich, A., Harris, A. K. (1981). Anomalous preference of cultured
macrophages for hydrophobic and roughened substrata. J Cell Sci 50: 1-7
83. Soskolne, W. A., Cohen, S., Sennerby, L., Wennerberg, A., Shapira, L.
(2002). The effect of titanium surface roughness on the adhesion of monocytes
and their secretion of TNF-α and PGE2. Clin Oral Implants Res 13 (1): 86-93
REFERENCES
130
84. Refai, A. K., Textor, M., Brunette, D. M., Waterfield, J. D. (2004). Effect of
titanium surface topography on macrophage activation and secretion of
proinflammatory cytokines and chemokines. J Biomed Res 70A: 194-205
85. Kullmann, K. (2002). Adaption des In Vitro Pyrogen tests (IPT) für
prothetische Materialien. Master Thesis, Fachbereich Biologie der Universität
Konstanz
86. Rosenberg, S. A. (2001). Cellular therapy: an introduction.
Cancer J 7 (Suppl 2): 51-2
87. Blajchman, M. A., Beckers, E. A., Dickmeiss, E., Lin, L., Moore, G., Muylle,
L. (2005). Bacterial detection of platelets: Current problems and possible
resolutions. Transfus Med Rev 19 (4): 259-272
88. Perez, P., Salmi, L. R., Follea, G., Schmit, J. L., de Barbeyrac, B., Sudre, P.,
Salamon, R. (2001). Determinants of transfusion-associated bacterial
contamination: results of the French BACTHEM case-control study. Transfusion
41 (7): 862-871
89. Kuehnert, M. J., Roth, V. R., Haley, N. R., Gregory, K. R., Elder, K. V.,
Schreiber, G. B., Arduino, M. J., Holt, S. C., Carson, L. A., Banerjee, S. N.,
Jarvis, W. R. (2001). Transfusion-transmitted bacterial infection in the United
States, 1998 through 2000. Transfusion 41 (12): 1493-1499
90. McDonald, C. P., Hartley, S. , Orchard, K. , Hughes, G., Brett, M. M., Hewitt,
P. E., Barbara, J. A. (1998). Fatal Clostridium perfringens sepsis from a pooled
platelet transfusion. Transfus Med 8(1): 19-22
91. Wollowitz, S. (2001) Fundamentals of the psoralen-based Helinx
technology for inactivation of infectious pathogens and leukocytes in platelets
and plasma. Semin Hematol 38(4 Suppl 11): 4-11
REFERENCES
131
92. Raij, L., Shapiro, F. L., Michael, A. F. (1973). Endotoxemia in febrile
reactions during hemodialysis. Kidney Int 4: 57-60
93. Favero, M S., Petersen, N. J., Boyer, K. M., Carson, L. A., Bond, W. W.
(1974). Microbial contamination of renal dialysis systems and associated health
risks. Trans Am Soc Artif Intern Organs 20A: 175-83
94. Klein, E., Pass, T., Harding, G. B., Wright, R., Million, C. (1990). Microbial
and endotoxin contamination in water and dialysate in the central United States.
Artif Organs 14: 85-94
95. Pegues, D. A., Oettinger, C. W., Bland, L. A., Oliver, J. C., Arduino, M. J.,
Aguero, S. M., McAllister, S. K., Gordon, S. M., Favero, M. S., Jarvis, W. R.
(1992). A prospective study of pyrogenic reactions in hemodialysis patients
using bicarbonate dialysis fluids filtered to remove bacteria and endotoxin. J Am
Soc Nephrol 3 (4): 1002-7
96. Kulander, L., Nisbeth, U., Danielsson, B. G., Eriksson, O. (1993).
Occurrence of endotoxin in dialysis fluid from 39 dialysis units. J Hosp Infect 24:
29-37
97. Bambauer, .R, Schauer, M., Jung, W. K., Daum, V., Vienken, J. (1994).
Contamination of dialysis water and dialysate. A survey of 30 centers. ASAIO J
40(4): 1012-6
98. Phillips, G., Hudson, S., Stewart, W. K. (1994). Persistence of microflora in
biofilm within fluid pathways of contemporary haemodialysis monitors (Gambro
AK-10). J Hosp Infect 27: 117-25
99. Schouten, W. E., Grooteman, M. P., van Houte, A. J., Schoorl, M., van
Limbeek, J., Nube, M. J. (2000). Effects of dialyser and dialysate on the acute
phase reaction in clinical bicarbonate dialysis. Nephrol Dial Transplant 15 (3):
379-84
REFERENCES
132
100. Arvanitidou, M., Spaia, S., Katsinas, C., Pangidis, P., Constantinidis, T.,
Katsouyannopoulos, V., Vayonas, G. (1998). Microbiological quality of water
and dialysate in all haemodialysis centres of Greece.
Nephrol Dial Transplant 13(4): 949-54
101. Laurence, R. A., Lapierre, S. T. (1995). Quality of hemodialysis water: a 7-
year multicenter study. Am J Kidney Dis 25: 738-50
102. Marion-Ferey, K., Leid, J. G., Bouvier, G., Pasmore, M., Husson, G.,
Vilagines, R. (2005). Endotoxin level measurement in hemodialysis biofilm using
"the whole blood assay". Artif Organs 29 (6): 475-81
103. Evans, R. C., Holmes, C. J. (1991). In vitro study of the transfer of
cytokine-inducing substances across selected high-flux hemodialysis
membranes. Blood Purif. 9: 92-101
104. Lonnemann, G., Behme, T. C., Lenzner, B., Floege, J., Schulze, M.,
Colton, C. K., Koch, K. M., Shaldon, S. (1992). Permeability of dialyzer
membranes to TNF alpha-inducing substances derived from water bacteria.
Kidney Int 42 (1): 61-8
105. Urena, P., Herbelin, A., Zingraff, J., Lair, M., Man, N. K., Descamps-
Latscha, B., Drueke, T. (1992). Permeability of cellulosic and non-cellulosic
membranes to endotoxin subunits and cytokine production during in-vitro
haemodialysis. Nephrol Dial Transplant 7: 16-28
106. Tsuchida, K., Takemoto, Y., Yamagami, S., Edney, H., Niwa, M., Tsuchiya,
M., Kishimoto, T., Shaldon, S. (1997). Detection of peptidoglycan and endotoxin
in dialysate, using silkworm larvae plasma and limulus amebocyte lysate
methods. Nephron 75 (4): 438-43
107. Dinarello, C. A., Koch, K. M., Shaldon, S. (1988). Interleukin-1 and its
relevance in patients treated with hemodialysis. Kidney Int Suppl 24: 21-6
REFERENCES
133
108. Miyasaka, N., Sato, K., Kitano, Y., Higaki, M., Nishioka, K., Ohta, K.
(1992). Aberrant cytokine production from tenosynovium in dialysis associated
amyloidosis. Ann Rheum Dis 51 (6): 797-802
109. Baz, M., Durand, C., Ragon. A., Jaber, K., Andrieu, D., Merzouk, T.,
Purgus, R., Olmer, M., Reynier, J. P., Berland, Y. (1991). Using ultrapure water
in hemodialysis delays carpal tunnel syndrome. Int J Artif Organs 14 (11): 681-5
110. Schwalbe, S., Holzhauer, M., Schaeffer, J., Galanski, M., Koch, K. M.,
Floege, J. (1997). Beta 2-microglobulin associated amyloidosis: a vanishing
complication of long-term hemodialysis. Kidney Int 52 (4): 1077-83
111. Daneshian, M., Guenther, A., Wendel, A., Hartung, T., von Aulock, S.
(2006). In vitro pyrogen test for toxic or immunomodulatory drugs
in print
112. Prausnitz, C. (1936). Investigations on respiratory dust disease in
operatives in the cotton industry. Medical Research Council, Special report
series 1936; 212: 1-73. His Majesty´s Stationery Office, London
113. Neal, P. A., Schneiter, R., Caminita, B. H. (1942). Report on acute illness
among rural mattress makers using low grade, stained cotton. JAMA 119: 1074-
1082
114. Cinkotai, F. F., Lockwood, M. G., Rylander, R. (1977). Airborne micro-
organisms and prevalence of byssinotic symptoms in cotton mills.
Am Ind Hyg Assoc J 38: 554-9
115. Donham, K. J., Zavala, D. C., Merchant, J. A. (1984). Acute effects of the
work environment on pulmonary functions of swine confinement workers.
Am J Ind Med 5: 367-75
REFERENCES
134
116. Donham, K.J., Zavala, D. C., Merchant, J. A. (1984). Respiratory
symptoms and lung function among workers in swine confinement buildings: a
cross-sectional epidemiological study. Arch Environ Health 39: 96-101
117. Thelin, A., Tegler, O., Rylander, R. (1984). Lung reactions during poultry
handling related to dust and bacterial endotoxin levels. Eur J Respir Dis. 65 266-
71
118. Zucker, B. A., Draz, A., Müller, W. (2000). Comparison of filtration and
impingement for sampling airborne endotoxin. J Aerosol Sci 31: 751-755
119. Kindinger, I., Daneshian, M., Baur, H., Gabrio, T., Hofmann, A., Fennrich,
S., von Aulock, S., Hartung, T. (2005). A new method to measure air-borne
pyrogens based on human whole blood cytokine response. J Immunol Methods
298 (1-2): 143-53
120. Rietschel, E. T., Kirikae, T., Schade, F. U., Ulmer, A. J., Holst, O., Brade,
H., Schmidt, G., Mamat, U., Grimmecke, H. D., Kusumoto, S. (1993). The
chemical structure of bacterial endotoxin in relation to bioactivity.
Immunobiology 187 (3-5):169-90
121. Emancipator, K., Csako, G., Elin, R. J. (1992). In vitro inactivation of
bacterial endotoxin by human lipoproteins and apolipoproteins. Infect Immun 60:
596-601
122. Paulsson, J., Michaelsen, P. (1984). The limulus amoebocyte lysate test
(LAL) assay for the detection of endotoxin in fat emulsions for total parenteral
nutrition (TPN). Acta Pathol Microbiol Immunol Scand 92: 177-179
123. Schindler, S., Rosenberg, U., Schlote, D., Panse, K., Kempe, A., Fennrich,
S., Hartung, T. (2006) Application of the InVitro Pyrogen Test (IPT) based on
cryopreserved human whole blood for lipidic parenterals. Pharmeuropa 2006, in
REFERENCES
135
124. Zimmermann, M., Busch, K., Kuhn, S., Zeppezauer, M. (1999). Endotoxin
adsorbent based on immobilized human serum albumin. Clin Chem Lab Med
37: 373-379
125. Russel, W. M. S. and Burch R. L. (1959). The principles of humane
experimental technique. Methuen, London
126. European Union (1986). EU-Directive 86/609/EEC. Official Journal of the
European Union L 358
127. Beutler, B. and Rietschel, E. T. (2003). Innate immune sensing and its
roots: the story of endotoxin. Nat Rev Immunol 3: 169-176
128. Mascoli, C. C. and Weary, M. E. (1979). Applications and advantages of
the Limulus amebocyte lysate (LAL) pyrogen test for parenteral injectable
products. Prog Clin Biol Res 29: 387-402
129. Mascoli, C. C. and Weary, M. E. (1979). Limulus amebocyte lysate (LAL)
test for detecting pyrogens in parenteral injectable products and medical
devices: advantages to manufacturers and regulatory officials. J Parenter
Drug Assoc 33: 81-95
130. Twohy, C. W., Duran, A. P., Munson, T. E. (1984). Endotoxin
contamination of parenteral drugs and radiopharmaceuticals as determined
by the limulus amebocyte lysate method. J Parenter Sci Technol 38: 190-201.
131. Roslansky, P. F. and Novitsky, T. J. (1991). Sensitivity of Limulus
amebocyte lysate (LAL) to LAL-reactive glucans. J Clin Microbiol 29: 2477-
2483
132. Dinarello, C. A. (1999). Cytokines as endogenous pyrogens.
J Infect Dis 179 Suppl 2: 294-304
REFERENCES
136
133. Dinarello, C. A., O'Connor, J. V., LoPreste, G., Swift, R. L. (1984)
Human leukocytic pyrogen test for detection of pyrogenic material in growth
hormone produced by recombinant Escherichia coli
J Clin Microbiol 20: 323-329
134. Duff, G. W. and Atkins, E. (1982). The detection of endotoxin by in vitro
production of endogenous pyrogen: comparison with limulus amebocyte
lysate gelation. J Immunol Methods 52: 323-331
135. Tsuchiya, S., Yamabe, M., Yamaguchi, Y., Kobayashi, Y., Konno, T.,
Tada, K. (1980). Establishment and characterization of a human acute
monocytic leukemia cell line (THP-1). Int J Cancer 26: 171-176.
136. Poole, S., Mistry, Y., Ball, C., Gaines Das, R. E., Opie, L. P., Tucker, G.,
Patel, M. (2003). A rapid 'one-plate' in vitro test for pyrogens.
J Immunol Methods 274: 209-220
137. Balls, M., Blaauboer, B., Brusick, D., Frazier, J., Lamb, D., Pemberton,
M., Reinhardt, C., Roberfroid, M., Rosenkranz, H., Schmid, B., Spielmann, H.,
Stammati, A. L., Walum, E. (1990). Report and recommendations of the
CAAT/ERGATT workshop on validation of toxicity test procedures.
Altern Lab Anim 18: 303-337
138. Balls, M., Blaauboer, B., Fentem, J. H., Bruner, L., Combes, R. D.,
Ekwall, B., Fielder, R. J., Guillouzo, A., Lewis, R. W., Lovell, D. P., Repetto,
G., Sladowski, D., Spielmann, H., Zucco, F. (1995). Practical aspects of the
validation of toxicity test procedures. The report and recommendations of
ECVAM workshop 5. Altern Lab Ani. 23: 129-147
139. Worth, A. P. and Balls M. (2002). The principles of validation and the
ECVAM validation process. Altern Lab Anim 30 Suppl 2: 15-21
REFERENCES
137
140. Poole, S. and Mussett, M. V. (1989). The International Standard for
Endotoxin: evaluation in an international collaborative study.
J Biol Stand 17: 161-171
141. Council of Europe (2001). Biological Tests, 2.6.8. Pyrogens. In:
European Pharmacopoeia. Council of Europe, Strasbourg: 131-132
142. Peterbauer, A., Eperon, S., Jungi, T. W., Werner, E. R., Werner-
Felmayer, G. (2000). Interferon-gamma-primed monocytoid cell lines:
optimizing their use for in vitro detection of bacterial pyrogens.
J Immunol Methods 233: 67-76
143. Taktak, Y. S., Selkirk, S., Bristow, A. F., Carpenter, A., Ball, C., Rafferty,
B., Poole, S. (1991). Assay of pyrogens by interleukin-6 release from
monocytic cell lines. J Pharm Pharmacol 43: 578-582
144. Schins, R. P., van Hartingsveldt, B., Borm, P. J. (1996). Ex vivo cytokine
release from whole blood. A routine method for health effect screening
Exp Toxicol Pathol 48: 494-496
145. Clopper, C. J., and Pearson, E. S. (1934). The use of confidence or
fiducial limits illustrated in the case of the binomial. Biometrika 26: 404-413
146. Council of Europe (2001). Biological Tests, 2.6.14. Bacterial endotoxins.
In: European Pharmacopoeia. Council of Europe, Strasbourg: 140-147
147. Nakagawa, Y., Maeda, H., Murai, T. (2002). Evaluation of the in vitro
pyrogen test based on proinflammatory cytokine release from human
monocytes: comparison with a human whole blood culture test system and
with the rabbit pyrogen test. Clin Diagn Lab Immunol 9: 588
REFERENCES
138
148. Eperon, S., De Groote, D., Werner-Felmayer, G., Jungi, T. W. (1997)
Human monocytoid cell lines as indicators of endotoxin: comparison with
rabbit pyrogen and Limulus amoebocyte lysate assay.
J Immunol Methods 207: 135-145
149. Harada, T., Misaki, A., Saito, H. (1968)
Curdlan: a bacterial gel-forming beta-1,3-glucan.
Arch Biochem Biophys 124: 292-298
150. Poole, S., Thorpe, R., Meager, A., Hubbard, A. R., Gearing, A. J. (1988)
Detection of pyrogen by cytokine release. Lancet 8577: 130
151. Goldman, J.M., Th'ng, K.H., Park, D.S., Spiers, A.S., Lowenthal, R.M.
and Ruutu, T. (1978). Collection, cryopreservation and subsequent viability of
haemopoietic stem cells intended for treatment of chronic granulocytic
leukaemia in blast-cell transformation. Br J Haematol 40: 185-95
152. Cavins, J. A., Djerassi, I., Roy, A. J. and Klein, E. (1965). Preservation of
viable human granulocytes at low temperature in dimethylsulfoxide.
Cryobiology 2: 129-33
153. Knorpp, C. T., Merchant, W. R., Gikas, P. W., Spencer, H. H. and
Thompson, N. W. (1967). Hydroxyethyl starch: extracellular cryophylactic
agent for erythrocytes. Science 157: 1312-1313
154. Handin, R. I. and Valeri, C. R. (1972). Improved viability of previously
frozen platelets. Blood 40: 509-513
155. Frim, J. and Mazur, P. (1983). Interactions of cooling rate, warming rate,
glycerol concentration, and dilution procedure on the viability of frozen-
thawed human granulocytes. Cryobiology 20: 657-76
REFERENCES
139
156. Hermann, C., von Aulock, S., Graf, K. and Hartung, T. (2003)
A model of human whole blood lymphokine release for in vitro and ex vivo
use. J Immunol Methods 275: 69-79
157. von Aulock, S., Hermann, C. and Hartung, T. (2003). Determination of
the eicosanoid response to inflammatory stimuli in whole blood and its
pharmacological modulation ex vivo. J Immunol Methods 277: 53-63
158. Hartung, T., Docke, W.D., Gantner, F., Krieger, G., Sauer, A., Stevens,
P., Volk, H.D. and Wendel, A. (1995). Effect of granulocyte colony-stimulating
factor treatment on ex vivo blood cytokine response in human volunteers.
Blood 85: 2482-9
159. Hartung, T., Doecke, W.D., Bundschuh, D., Foote, M.A., Gantner, F.,
Hermann, C., Lenz, A., Milwee, S., Rich, B., Simon, B., Volk, H.D., von
Aulock, S. and Wendel, A. (1999). Effect of filgrastim treatment on
inflammatory cytokines and lymphocyte functions.
Clin Pharmacol Ther 66: 415-24
160. von Aulock, S., Boneberg, E.M. and Hartung, T. (2000).
Intermittent G-CSF (filgrastim) treatment cannot induce lymphocytosis in
volunteers Clin Pharmacol Ther 68: 104
161. Fennrich, S., Wendel, A. and Hartung, T. (1999). New Applications of the
Human Whole Blood Pyrogen Assay (PyroCheck). Altex 16: 146-149
162. Schumann, R.R. (1992)
Function of lipopolysaccharide (LPS)-binding protein (LBP) and CD14, the
receptor for LPS/LBP complexes: a short review
Res Immunol 143: 11-5
163. Fenton, M.J. and Golenbock, D.T. (1998). LPS-binding proteins and
receptors. J Leukoc Biol 64: 25-32
REFERENCES
140
164. Blackman, M. A., and Woodland, D.L. (1995). In vivo effects of
superantigens. Life Sci 57: 1717-35
165. Nakatani, T., Tsuchida, K., Sugimura, K., Yoshimura, R., and Takemoto,
Y. (2002). Response of peripheral blood mononuclear cells in hemodialyzed
patients against endotoxin and muramyldipeptide
Int J Mol Med 10: 469-72
166. Emancipator, K., Csako, G., Elin, R. J. (1992). In vitro inactivation of
bacterial endotoxin by human lipoproteins and apolipoproteins. Infect Immun 60:
596-601
167. Paulsson, J., Michaelsen, P. (1984). The limulus amoebocyte lysate test
(LAL) assay for the detection of endotoxin in fat emulsions for total parenteral
nutrition (TPN). Acta Pathol Microbiol Immunol Scand 92: 177-179
168. Richardson, E. C., Banerji, B., Seid, R. C., Levin, J., and Alving, C. R.
(1983). Interactions of lipid A and liposome-associated lipid A with Limulus
polyphemus amoebocytes. Infect Immun 39: 1385-91
169. Kreeftenberg, J. G., Loggen, H. G., van Ramshorst, J. D., and Beuvery,
E. C. (1977). The limulus amoebocyte lysate test micromethod and
application in the control of sera and vaccines. Dev Biol Stand 34: 15-20
170. Ochiai, M., Yamamoto, A., Kataoka, M., Toyoizumi, H., and Horiuchi, Y.
(2001). Interfering effect of Diphteria-Tetanus-Acellular Pertussis Combined
(DtaP) Vaccines on the Bacterial Endotoxin Test. Biologicals 29: 55-58
171. Hailman, E., Lichenstein, H. S., Wurfel, M. M., et al. (1994).
Lipopolysaccharide (LPS)-binding protein accelerates the binding of LPS to
CD14. J Exp Med 179: 269-77
REFERENCES
141
172. Poole S., Gaines Das RE. Towards a humane pyrogen test. Eur J
Parenteral Sci 2001; 6 (2): 63-64.
173. Dijkstra, J., Mellors, J. W., Ryan, J. L., and Szoka, F. (1987)
Modulation of the biological activity of bacterial endotoxin by incorporation
into liposomes. J Immunol 138: 2663-2670
174. Stitt, J. T. (1985). A study of the variability in the febrile responses of
rabbits to endogenous pyrogen. J Appl Physiol 59: 1254-7
175. Schutt, C., Schilling, T., Grunwald, U., Schonfeld, W., and Kruger, C.
(1992). Endotoxin-neutralizing capacity of soluble CD14.
Res Immunol 143: 71-8
176. Frey, E. A., Miller, D. S., Jahr, T. G., Sundan, A., Bazil, V., Espevik, T.,
Finlay, B. B., and Wright, S. D. (1992). Soluble CD14 participates in the
response of cell to lipopolysaccharide. J Exp Med 176: 1665-1671
177. Kirkland, T. N., Finley, F., Leturcq, D., Moriarty, A., Lee, J.-D., Ulevitch,
R. D., and Tobias, P. S. (1993). Analysis of lipopolysaccharide binding by
CD14. J Biol Chem 268: 24818-24823
178. Beutler, B., K. Hoebe, X. Du, and R. J. Ulevitch. (2003). How we detect
microbes and respond to them: the Toll-like receptors and their transducers.
J Leukoc Biol 74:479
179. Hirschfeld, M., Weis J. H., Vogel, S. N., and Weis J. J. (2000). Cutting
Edge: repurification of lipopolysaccharide eliminates signalling through both
human and murine toll-like receptor 2. J Immunol 165: 618-622
REFERENCES
142
180. Poltorak, A., He, X., Smirnova, I., Liu, M. Y., Huffel, C. V., Du, X.,
Birdwell, D., Alejos, E., Silva, M., Galanos, C., Freudenberg, M., Ricciardi-
Castagnoli, P., Layton, B., and Beutler, B. (1998). Defective LPS-signaling in
C3H/HeJ and C57BL/10ScCr mice: mutations in the tlr4 gene.
Science 282: 2085-2088
181. Schwandner, R., R. Dziarski, H. Wesche, M. Rothe, and C. J.
Kirschning. (1999). Peptidoglycan- and lipoteichoic acid-induced cell
activation is mediated by toll-like receptor 2. J Biol Chem 274: 17406.
182. Takeuchi, O., Hoshino, K., Kawai, T., Sanjo, H., Takanda, H., Takeda,
K., and Akira, S. (1999). Differential roles of TLR2 and TLR4 in recognition of
gram-negative and gram- positive bacterial cell wall components.
Immunity 11: 443-51
183. Thornberry, N. A., Bull, H. G., Calaycay, J. R., Chapman, K. T., Howard,
A. D., Kostura, M. J., Miller, D. K., Molineaux, S. M., Weidner, J. R., et al.
(1992). A novel heterodimeric cysteine protease is required for interleukin-1β
processing in monocytes. Nature 356: 768-774
184. Cannon, J. G., Tompkins, R. G., Gelfand, J. A., Michie, H. R., Stanford,
G. G., van der Meer, H. W. M., Endres, S., Lonnemann, G., Corsetti, J.,
Chrenow, B., Wilmore, D. W., Wolff, S. M., Burke, J. F., and Dinarello, C. A.
(1990). Circulating interleukin-1 and tumor necrosis factor in septic shock and
experimental endotoxin fever. J Infect Dis 161: 79-84
185. Engel, A., Kern, W. V., Murdter, G., and Kern, P. (1994)
Kinetics and correlation with body temperature of circulating interleukin-6,
interleukin-8, tumor necrosis factor alpha and interleukin-1 beta in patients
with fever and neutropenia . Infection 22: 160-164
REFERENCES
143
186. van Zee, K. J., Coyle, S. M., Calvano, S. E., Oldenburg, H. S. A., Stiles,
D. M., Pribble, J., Catalano, M., Moldawer, L. L., and Lowry, S. F. (1995)
Influence of IL-1 receptor blockade on the human response to endotoxemia.
J Immunol 154: 1499-1507
187. Luheshi, G., Miller, A. J., Brouwer, S., Dascombe, M. J., Rothwell, N. J.,
and Hopkins, S. J. (1996). Interleukin-1 receptor antagonist inhibits endotoxin
fever and systemic interleukin-6 induction in the rat.
Am J Physiol 270: E91-5
188. Banks, W. A., Ortiz, L., Plotkin, S. R., and Kastin, A. J. (1991)
Human interleukin (IL) 1α, murine IL-1α and murine IL-1β are transported
from blood to brain in the mouse by a shared saturable mechanism.
J Pharmacol Exp Ther 259: 988-996
189. Banks, W. A., Kastin, A. J., and Gutierrez, E. G. (1994). Penetration of
interleukin-6 across the murine blood-brain barrier. Neurosci Lett 179: 53-56
190. Tewari, A., Buhles, W. C., Starnes, H. F. (1990). Preliminary report:
effects of interleukin-1 on platelet counts. Lancet 336(8717): 712-4
191. Ogilvie, A. C., Hack, C. E., Wagstaff, J. , et al.(1996). IL-1 beta does not
cause neutrophil degranulation but does lead to IL-6, IL-8, and nitrite/nitrate
release when used in patients with cancer. J Immunol 156: 389-94
192. Nemunaitis, J., Ross, M., Meisenberg, B., et al. (1994). Phase I study of
recombinant human interleukin-1 beta (rhIL-1 beta) in patients with bone
marrow failure. Bone Marrow Transplant 14: 583-8
193. Crown, J., Jakubowski, A., Gabrilove, J. (1993). Interleukin-1: biological
effects in human hematopoiesis. Leuk Lymphoma 9: 433-40
REFERENCES
144
194. Du Bois, M. J., Schellekens, P. T., De Wit, J. J., and Eijsvoogel, V. P.
(1976). In vitro reactivity of human lymphocytes after cryopreservation using a
programmed cooling device. Scand J Immunol Suppl 5: 17-22
195. Doebbler, G.F., and Rinfret, A. P. (1962)
The influence of protective compounds and cooling and warming conditions
on hemolysis of erythrocytes by freezing and thawing
Biochim Biophys Acta 58: 449-58
196. Lionetti, F. J., Hunt, S. M., Gore, J. M., et al. (1975)
Cryopreservation of human granulocytes. Cryobiology 12: 181-91
197. Lionetti, F. J., and Hunt, S. M. (1975). Cryopreservation of human red
cells in liquid nitrogen with hydroxyethyl starch. Cryobiology 12: 110-8
198. Mazur. P. (1963). Kinetics of water loss from cells at subzero
temperatures and the likelihood of intracellular freezing.
J Gen Physiol 47: 347-69
199. Mazur, P. (1965). Causes of injury in frozen and thawed cells.
Fed Proc 24:1 75-82
200. Mazur, P. (1977). The role of intracellular freezing in the death of cells
cooled at supraoptimal rates. Cryobiology 14: 251-72