Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Research Collection
Doctoral Thesis
The role of perforin protein in lymphocyte mediated cytotoxicity
Author(s): Kägi, David
Publication Date: 1993
Permanent Link: https://doi.org/10.3929/ethz-a-000891428
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.
ETH Library
Diss.ETH No 10050
The Role of Perforin Protein in Lymphocyte
Mediated Cytotoxicity
A dissertation submitted to the
SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH
for the degree of
Doctor of Natural Sciences
presented by
David Kagi
Dipl.sc.nat.
born June 2,1963
citizen of Zurich, Wallisellen (ZH) and Baretswil (ZH)
accepted on the recommendation of
Prof. T.Koller, examiner
Prof. H.Hengartner, coexaminer
1993
Reprint from: European Journal of Immunology (1989), 19:1253 -1259
Man darfnie aufhoren, sich die Welt vorzustellen,
wie sie am vemunftigsten ware.
Friedrich Durrenmatt
CONTENT
Summary 4
Zusammenfassung 6
Abbreviations 8
General Introduction 9
Perforin Expression during Infection ofMice
with Lymphocytic Choriomeningitis Virus
5.1 Detection of Perforin and Granzyme A in Infiltrating Cells
during Infection with Lymphocytic Choriomeningitis Virus 28
Summary 29
Introduction 30
Material and Methods 33
Results 36
Discussion 49
52 Detection of Perforin Expression during Infection ofMice
with Lymphocytic Choriomeningitis Virus by RNA-PCR. 52
Introduction 52
Material and Methods 53
Results 55
Discussion 59
Generation of Mice with a Disruption of the Perforin
Gene by Homologous Recombination in ES cells 62
Introduction 62
Material and Methods 65
Results 70
Discussion 103
General Discussion 109
References 114
Dank
Curriculum Vitae
4
1. Summary
Cytotoxic T cells and natural killer cells are able to eliminate virus infected
tumorigenic and allogeneic cells. The effector mechanism by which these cells are
able to lyse target cells is still controversial. From all candidate models, most
evidence has been provided for the granula exocytosis model. It proposes that
cytolytic lymphocytes, upon recognition and interaction with the target cell, release
the cytolytic pore-forming protein perforin into the intercellular space, which
subsequently leads to colloid osmotic lysis of the target cell. Since the relevance of
this concept for cytotoxicity in vivo has been questioned, the expression of perforin
was investigated during the infection of mice with lymphocytic choriomeningitis
virus. It is well established that CD8 positive cytotoxic T cells play a pivotal role in
the elimination of this non-cytopathic virus. Perforin expression was analyzed by in
situ hybridisation of serial liver and brain sections from infected mice and by
immunohistochemical staining with T cell marker and virus specific antibodies. A
close histological association of infiltrating lymphocytes expressing perforin mRNA
with virally infected cells was observed. The distribution of perforin expressing
cells on serial sections compared to the distribution of CD8 positive cells was
consistent with the notion, that CD8 positive T cells express perforin. In addition,
the maximal frequency of perforin expressing cells on liver sections preceded by
about 2 days maximal LCMV specific cytotoxicity of the lymphoid liver infiltrating
cells.
Complementary investigation by RNA-PCR confirmed the upregulation of
perforin expression in the spleen by a factor of at least 10 during LCMV infection.
These findings, though somewhat circumstantial, are most consistent with an
involvement of perforin in cell mediated cytotoxicity in vivo.
To provide more direct evidence for the role of perforin in vivo, the generation of
perforin deficient mice by homologous recombination in embryonic stem cells was
planned. Embryonic stem cells with a targeted mutation of the perforin gene were
5
isolated. A neomycine resistance gene with multiple translational stop codons in
all three reading frames was introduced in exon 3 resulting in the formation of
non-functional mRNA from the disrupted perforin gene. By blastocyst injection of
the targeted pluripotent embryonic stem cells and surgical introduction of the
embryo into the uterus of a foster mother chimeric mice were obtained. Eight
chimeras transmitted the disrupted locus to 50% of their offsprings. By breeding
the heterozygot offsprings with each other homozygous perforin deficient mice will
be obtained. These animals will allow a direct assessment of the role of perforin
for the cytotoxicity of lymphocytes and eventually provide a model system to test
the role of cytotoxic T cells and natural killer cells in infections with various
biological agents and in murine models for autoimmune diseases.
6
2. Zusammenfassung
Zytotoxische T-Zellen und sogenannte NK-Zellen haben die Fahigkeit,
virusinfizierte und neoplastische Zellen zu eliminieren. Der molekulare
Mechanismus dieser Zellyse ist nach wie vor kontrovers. Unter den
vorgeschlagenen Modellvorstellungen ist das Granula-Exocytose-Modell das
experimentell am breitesten abgestiitzte. Es beruht auf der Vorstellung, dass
cytolytische Lymphozyten, nachdem sie ihre Zielzelle spezifisch erkannt und
gebunden haben, das Protein Perforin sekretieren, das dann durch Porenbildung
den cytolytischen Effekt bewirkt.
Da dieses Modell speziell fur die in vivo Situation in Frage gestellt worden ist,
wurde die Expression von Perforin wahrend Infektion von Mausen durch das
nicht-cytopathische Lymphocytare Choriomeningitis Virus untersucht. Von diesem
Virus ist bekannt, dass es hauptsachlich durch CD8 positive cytotoxische T-Zellen
bekampft wird. Perforin Expression wurde durch In Situ Hybridisation auf Leber-
und Hirnschnitten nachgewiesen und durch immunhistochemische Farbungen fur
T-Zell Oberflachenmolekule und Vinisantigene auf seriellen Schnitten derselben
Gewebeproben erganzt Es wurde gezeigt, dass Perforin exprimierende Zellen im
infizierten Gewebe in der unmittelbaren Nahe von Virus infizierten Zellen zu
finden sind und dass die Verteilung dieser Zellen im Gewebe identisch mit der
Verteilung der CD8 positiven Zellen war. Das Haufigkeitsmaximum der Perforin
exprimierenden Zellen ging dem Maximum der zytolytischen Aktivitat von
Leberlymphozyten um zwei Tage voran. Die um einen Faktor von mindestens
zehn erhohte Expression von Perforin wahrend der Infektion mit dem
Lymphozytaren Choriomeningitis Virus wurde durch eine Untersuchung von
Milzen infizierter Tiere mit RNA-PCR bestatigt. Alle diese Befunde weisen
demnach auf eine Beteiligung von Perforin an der Zell vermittelten Zytolyse auch
in vivo hin.
Um einen direkten Beweis fur die Relevanz von Perforin in vivo zu fuhren, wurde
die Herstellung von Perforin defizienten Mausen durch homologe Rekombination
7
in embryonalen Stammzellen geplant. Embryonale Stammzellen mit einer
gezielten Mutation in Exon 3, die zum Abbnich der Proteintranslation in alien drei
Leserastera fuhrt, wurden isoliert. Durch Injektion der mutierten Stammzellen in
das Blastocoel von Mauseembryonen und durch chirurgisches Einfuhren dieser
Embryonen in den Uterus von Ammenmuttermausen wurden chimare Mause
erhalten. Acht Chimaren gaben die Mutation an 50% ihrer Nachkommen weiter,
die dann zu homozygoten Perforin defizienten Mause weitergeziichtet werden
konnen. Diese Tiere werden eine direkte Untersuchung der Rolle von Perforin im
Prozess der Zytolyse durch Lymphozyten erlauben. Zusatzlich konnten sie ein
interessantes Testsystem ergeben, mit dem der Beitrag der Cytotoxizitat von
cytolytischen T-Zellen und NK-Zellen sowohl wahrend einer Infektion mit
verschiedenen mikrobiologischen Organismen als auch in Maus-Modellen fur
Autoimmunkrankheiten untersucht werden kann.
8
3. Abbreviations
APC antigen presenting cell
CD complex of differentiation
con A concanavaline A
CTL cytotoxic T cell
DNA desoxy-ribonucleic acid
EScell embryonic stem cell
FCS fetal calf serum
i.c. intracerebral
i.v. intravenous
Ig immunoglobulin
IL interleukin
L.U. lytic unit
LCMV lymphocytic choriomeningitis virus
LCMV-WE strain WE ofLCMV
LGL large granular lymphocyte
MHC major histocompatibility complex
mRNA messenger ribonucleic acid
NK natural killer
NOD nonobese diabetes
PCR polymerase chain reaction
pfu plaque forming unit
RNA ribonucleic acid
TNF tumour necrosis factor
9
4. General Introduction
The task of the immune system is to maintain the homeostasis of the body by
defending its integrity against numerous microbial infections, parasites and foreign
toxic substances. For this purpose, the immune system is capable to discriminate
between self and nonself antigens. Nonself antigens generally generate a
protecting immune response, while against self determinants a response is
prevented by a state of nonresponsiveness or tolerance, which is established at an
early time point in ontogeny when the immune system is not completely functional
and mature yet. Already back at the beginning of the twentieth century, Ehrlich
recognized the absence of reactivity towards the organisms own structures, the so
called "horror autotoxicus", as a key property of the immune system. But from
todays point of view, the notion, that the immune system never shows any
autoreactivity, appears not absolutely correct; under certain physiological
conditions, the immune system overrides the principle of non-reactivity against self
by the elimination of virus infected or tumourigenic cells. In such situations, it is
advantageous to the inflicted organism to sacrifice a small number of cells in order
to prevent further damage by spreading of the virus or the tumour. The cells of the
immune system able to lyse virus infected and tumour cells are the cytotoxic T cells
and NK cells. Both of them are part of cell mediated immunity, in which
antibodies play only a subordinate role.
Eversince the capacity of lymphocytes to lyse nucleated cells was discovered
(Govaertz, 1960), speculations about the molecular mechanism of cell lysis by
lymphocytes were made. But the issue has not found its definitive solution yet and
is still controversial. Recent advances suggest a surprising similarity of the effector
functions in the humoral and cellular branch of the immune system. It is now very
probable, that cytotoxic T cells and NK cells perform cell lysis by a mechanism that
is very similar to the antibody induced cell lysis by the complement system.
But as it applies also for the complement system, the drastic impact of cell lysis not
always remains beneficial to the host: On one side it can protect the body from
10
fatal virus infections and tumour growth but on the other side, its excessive and
unfavorable manifestations can lead to autoimmune disease and life threatening
transplant rejection. Understanding the cells and molecules involved in the process
of cytotoxicity could help to interfere one day with the course of such diseases in a
more specific way.
Discovery of the Cytotoxic Capacity of Lymphocytes
The saying of the war being the father of all things applies also to the discovery of
cytotoxic lymphocytes. World war n, with its many victims, generated an intense
interest in transplant immunology. It was in England of the 40s that Medawar,
Snell, Gorer, Billingham and Brent showed that lymphocytes are responsible for
the vigorous rejection reaction after transplantation of allogeneic tissue. This was
just after the central role of lymphocytes in allergy had been recognized
(Landsteiner and Chase, 1942). That a direct contact between grafted and effector
cells is essential for rejection of the grafted tissue was concluded from experiments
in which tumour fragments enclosed in diffusion chambers were not rejected after
transplantation (Algjre, 1957). The transfer of allograft rejection by primed
lymphoid cells but not by antisera established the cellular basis of allograft
rejection (Mitchison, 1954). That indeed a lytic effect is responsible for allograft
rejection was demonstrated by the cytopathic effect of thoracic duct lymphocytes ,
procured after kidney allograft rejection, on donor renal cells cultured in vitro.
(Govaertz, 1960). Miller established the importance of the thymus in transplant
rejection by demonstrating the absence of skin transplant rejection in neonatally
thymectomized mice (Miller, 1961). Thymus dependent T cells were postulated
and identified as the main cytoloytic lymphocyte population (Cerrotini et al.,
1970). Later, they were further characterized as a T cell subpopulation carrying the
CD8 surface marker (Cantor and Boyse, 1975).
11
Cytotoxic T Cells
Cytotoxic T cells defend the host mainly against intracellular agents like viruses,
tumour cells and intracellular bacteria. The capacity to lyse virus infected cells
before new infectious agents have formed in an infected cell renders cytotoxic T
cells mainly effective against viruses. Beyond these natural functions, cytotoxic T
cells are the main effector cells in rejection of allogeneic tissue grafts as well.
T cells can be divided into the two functional subclasses by the presence of the
CD4 and CD8 surface markers. Cytotoxic T cells belong to the CD8 positive subset
and require, like all T cells, the thymus for development. Activated, functional
cytotoxic T cells develop from resting CTL precursor cells by triggering via their
receptor and by additional Interleukin 2 mediated stimulation.
In contrast to B cells, T-cells do not recognize native protein antigen, but instead
short octa- or nonameric peptides bound to MHC class I molecules (Townsend et
al., 1986; Falk et al., 1991). Peptides derived from intracellular proteins by a
limited proteolysis called processing are mainly presented on MHC class I
molecules on the cell surface (HLA-A, B, C loci in humans, H-2 K, H-2 D and H-
2-L loci in mouse) while extracellular antigenic proteins are endocytosed via a
phagolysosome, processed and presented on MHC class II molecules (Townsend et
al., 1986; Morrison et al., 1986; Germain, 1986). The T cell does not recognizes the
free peptide but rather the association of peptide with a matched MHC molecule,
a property of T cells called restriction.(Zinkernagel and Doherty, 1974).
The MHC class I molecule consists of the association of an a chain with an fy
microglobulin chain. On the surface of this heterodimer, two a helices confine the
peptide binding groove (Bjorkman et al., 1987b; Bjorkman et al., 1987a; Garrett et
al., 1989; Jardetzky et al., 1991; Madden et al., 1991). This peptide-MHC class I
complex is specifically recognized by the polymorphic T cell receptor (TCR) of the
cytotoxic T cell (Yoshikai et al., 1984; Allison and Lanier, 1985; Hedrick et al.,
1984; Sim et al, 1984). The TCR is a a/0 heterodimeric membrane protein
expressed on all mature T cells. The polymorphism of both chains is generated by
12
recombination of V, D (only /} chain) and J elements similar to the recombination
mechanism functional in B cells (Bjorkman et al., 1987b).
In addition to the specific interaction conferred by the trimolecular MHC-peptide-
TCR complex, cytotoxic T cells bind to their target cells via a number of unspecific
accessory molecular interactions. These receptor ligand pairs involve CD8-MHC
class I molecule (via a non polymorphic determinant), LFA-1 -ICAM-1, CD2 -
LFA-3, CD28 - B7, CDS - CD27 and CD45RO - CD22 (reviewed by Springer,
1990). The function of these accessory interactions is not completely clear yet but
probably helps to increase the relatively low affinity of the TCR-peptide-MHC
binding (Weber et al., 1992).
Also, the precise molecular events that link receptor engagement to response of
the cytotoxic T cell remain unresolved. Nonetheless, it is thought that certain
transduction mechanisms exist in T cells: Phosphorylation of protein kinase C and
of the f chain of the CD3 complex by the tyrosine kinases p56ldt and p59fyn of the
src family, dephosphorylation by CD45RO and subsequent G protein dependent
breakdown of phosphatidylinositol leading to an increase of intracellular Ca2+ and
activation of protein kinase C have been observed during T cell activation.
13
Antigen-presenting cell
Tcell
Fig.l. Molecules involved in T cell interaction with antigen presenting cells (APC).
Interaction of CD8 positive and CD4 positive T cell with an APC combined in this
drawing for illustrative purposes only. Human CD11a correspond to LFA-1, CD54
to ICAM-1 and CD58 to LFA-3. Lck and fyn are the tyrosine kinases p56lck and
pS^. CD45RO is a tyrosine phosphatase.
NK Cells
NK cells appear with large granular lymphocyte (LGL) morphology and are
thought to play a role in tumour resistance, host immunity to viral and perhaps
other microbial agents, especially as a first line of defense against virus. In nude
mice an increased NK activity, at least partly, compensates for the absence of
cytotoxic T cells. NK cells do not show rearrangement of TCR genes and do not
express the CD3 surface molecule. However, they usually express the asialo GM1,
the NKl and the CD56 (Leu 19) surface marker molecules. A consistent feature of
NK cells is their association with a subpopulation of cells, the LGLs. LGLs can be
14
isolated from blood and spleens of unprimed animals by centrifugation on Percoll
gradients and elimination of cells that form high affinity rosettes with sheep red
blood cells.
Some target cell lines are highly susceptible for target cell lysis (human K562 and
murine YAC-1 cells) and serve as prototype NK activity indicator cell lines. It is
unclear why these cell lines are so sensitive to NK cells, but recent publications
support the interesting idea, that NK cells preferentially lyse cells lacking a
protective peptide associated to class I (reviewed by Versteeg, 1992). Following
these reports, NK sensitivity would be a consequence of either downregulation of
MHC class I molecule in tumour cells, which is an immunological trick to
circumvent recognition by tumour antigen specific T cells, or alternatively results
because a conserved peptide of an intracellular parasite competes with a related
endogenous protective peptide..
The Granule Exocytosis Model of Target Cell Destruction by Lymphocytes
Already back in 1968, it was suggested that lymphocytes lyse target cells by a
mechanism involving an active release of a cytolytic factor by the cytotoxic cell
(Granger and Kolb, 1968). Since then, this suggestion has been confirmed by
numerous observations, which are in detail listed below:
The presence of cytoplasmic granules has been noted in most types of cytolytic
lymphocytes.
Localisation of granules become localized between nucleus and target cell by a
rearrangement of the cytoskeleton after target cell binding (Bykovskaja et al.,
1978a; Bykovskaja et al., 1978b).
The observation of acid phosphatase deposition, originally found in the
lysosomal granula of cytotoxic T cells, at the effector-target cell junctions
during cytolytic-target cell interaction.
The isolation of a protein named Perforin from the granula of cytolytic
lymphocytes with potent cytolytic properties when added to cells in vitro
(Masson and Tschopp, 1985; Podack et al., 1985; Liu et al., 1986).
15
The direct demonstration of complement like pore structures on target cell
membranes of an internal diameter of 15 nm ( complement pores: 10 nm)
(Dourmashkin et al., 1980; Podack, 1983; Bykovskaja et aL, 1978a)
All these observations were subsumed in the granule exocytosis model of target
cell lysis (Fig.2) (Henkart, 1985). It consists of the following discrete steps: (i)
Receptor and adhesion molecule mediated binding of the target cell to the
lymphocyte, (ii) An increase of intracellular Ca2+in the cytotoxic cell and a
subsequent polarization of the Golgi apparatus, the granules and the cytoskeleton
towards the target cell, (iii) Vectorial granule exocytosis and the release of pore
forming protein (Perforin) monomers in the intracellular space between cytolytic
and target cell, (iv) Calcium dependent assembly of membrane lesions with
polymerization of the monomer into a functional channel in the target cell
membrane.
The granule exocytosis model includes the notion that the formation of a
transmembrane polyperforin pore of 16 - 20 nm diameter allows the equilibration
of ion concentration gradients over the membrane. The passive flow of water into
the injured cell from the outside in response to the colloid osmotic pressure
difference generated by the proteins causes the cell to swell and eventually to
burst.
16
Fig.2. Granule exocytosis model of lymphocyte mediated cytotoxicity as it was
suggested by (Henkart, 1985). This model shows the adhesion and recognition
phase mediated by the T cell receptor and accessory molecules, followed by the
cytoplasmic rearrangement bringing the granula in close proximity to the target
cell and the granule exocytosis step. Today, the granule exocytosis model includes
the notion, that target cell lysis is caused by the polymerization of secreted perforin
monomers into a membrane pore causing colloid osmotic lysis of the target cell.
After target cell lysis, the effector can recycle and, probably after replenishment of
the granules by resynthesis of granula proteins, kill another target cell.
17
Alternative Mechanisms of Target Cell Lysis
The granule exocytosis model has been challenged mainly because target cell lysis
was not completely abolished in Ca2+ depleted medium (Trenn et al., 1987). The
complete absence of intracellular and extracellular Ca2+ is thought to inhibit
exocytosis and to prevent the polymerization of perforin monomers.
Another challenge came from the difficulties to detect cytolytic granules in
peritoneal exudate lymphocytes (PEL), potent alloreactive cytolytic cells, by light
microscopy and by the hemolytic bioassay.(8erke and Rosen, 1987). But
subsequent more sensitive analysis with Northern blotting demonstrated
considerable amounts of Perforin mRNA in these cells (Nagler Anderson et al.,
1989).
The main alternative to the granule exocytosis model is the notion that target cells
destroy themselves by a suicidal process named apoptosis (Ucker, 1987; Russell,
1983). Apoptosis is a concept originating from embryology to explain the
controlled cell death without scaring during development. It is thought to be
associated with extensive degradation of genomic nuclear DNA into multiples of
about 200 bp. The internal disintegration model of cytolysis by lymphocytes is
supported by the appearance of early nuclear condensation in target cells during
the cytolytic process. However, the finding that the characteristic "laddering" of
genomic DNA occurs only in some, but not all target cell lines during lysis, and
that fibroblasts derived from mice transgenic for the apoptosis inhibiting bcl2
protein are still susceptible to cytolytic T cells seriously question the relevance of
the apoptosis model. It will be important to verify the exact cause of DNA
fragmentation, especially the question if DNA fragmentation represents the cause
or the merely a consequence of cell death.
Additional candidate mediators of cytotoxicity are the several serine proteases
called Granzymes that have been isolated from the granules of cytotoxic T cells
(Gershenfeld and Weissman, 1986; Masson and Tschopp, 1987). The blockage of
cytolytic lymphocytes by the addition of serine esterase inhibitors has been taken
as evidence for their involvement in target cell lysis (Redelman and Hudig, 1980;
18
Hudig et al., 1981). Since they have been found subsequently also in non cytolytic
lymphocyte subsets, their function in the granules of cytotoxic T cells needs further
investigation, (reviewed by Jenne and Tschopp, 1988).
Several factors with homology to tumour necrosis factors (TNF) have been
isolated from cytolytic granules (Granger and Kolb, 1968; Ruddle and Waksman,
1968; Traub, 1936). Some of these factors can induce slow (over several hours)
target cell lysis and DNA fragmentation. Due to the slow activity, which is hard to
reconcile with target cell lysis occurring in minutes, their relevance remains
unclear.
Molecular Characteristics of Perforin Protein
So far, the granule exocytosis model remains the best characterized to account for
cytotoxicity by lymphocytes. It relies on the recently identified pore forming
protein perforin as the main effector molecule.
Originally, perforin was isolated from granula of cytotoxic lymphocytes with the
help of a bioassay for hemoloysis of sheep red blood cells (SRBC). SRBC are
easily lysed by granula and purified perforin and after lysis display the
characteristic polyperforin lesions on the cell membrane, reminding of the ringlike
pores observed on the surface of cells lysed by the complement system (Podack et
al., 1985; Lichtenheld et al., 1988). Subsequent to the isolation of perforin and the
demonstration of its potent cytolytic activity, the complete cDNA from mouse and
human was cloned and sequenced with the help of a degenerate oligonucleotide
probe derived from the sequence of amino acids 21-29 (Shinkai et al., 1988;
Lichtenheld et al., 1988)
Both mouse and human perforin are encoded by a single genomic locus located on
mouse chromosome 10 (Trapani et al., 1989; Zink et al., 1992). The localization of
human perforin to chromosome 17 (Shinkai et al., 1989) is contradicted by a recent
publication reporting evidence for a localization to human chromosome 10 (Zink
et al., 1992). Both genes are composed of only three exons (Fig.3). Exon one codes
for the majority of the 5'-untranslated sequence; exon two for the remainder of 5'
19
untranslated sequence, the translation start signal, the signal peptide, and the N-
terminal part of the molecule up to, but not including, the putative membrane
binding site. Exon three encodes for the remainder of the molecule including the 3'
untranslated sequence.
Genomic organisationof Perforin
Coding sequence 1-534
C9 homology 44-410
TM region 168-210
EGF domain 356-386
FigJ. Genomic structure of murine perforin. The perforin locus consists of
untranslated exon 1 and the two coding exons 2 and 3, contained in 10 kb of
genomic sequence. The distribution of sequence homologous to C9 complement
component is indicated. The organisation of the human perforin locus is very
similar.
The main inductors of perforin expression in human peripheral blood cells known
so far is the stimulation of T cells via the TCR-CD3 complex or the exposure to
high amount of IL-2. IL-6 synergizes with IL-2 in the induction of perforin
expression and cytolytic potential. In the presence of IL-6, suboptimal doses of IL-
2 can induce perforin synthesis (Smyth et al., 1990b; Smyth et al., 1990a).
The promoter of the human and murine perforin genes are highly homologous.
However, the previously reported potential regulatory elements of the human
promoter were not found in the mouse perforin promoter and consequently, novel
transcription factors are at least partly regulating perforin transcription
(Lichtenheld and Podack, 1989).
-m-
Start Sma1] I
BstE2 StopI L
20
The perforin transcript is 2.6 - 2.9 kb long as determined by Northern blot analysis.
Only one major transcription initiation site has been found in the human perforin
gene. In the murine gene, in contrast, transcription starts from at least two
different initiation sites (Youn et al., 1991). This results in two differently spliced
mRNA species, with a 5'-untranslated region of either 222 or 115 bp. This
alternative splicing is generated by the use of different splice acceptor sites.
Perforin is expressed in vivo in NK cells, CD8 positive cytolytic T cells and 76
TCR positive T cells (Nakata et al., 1992). It is generally not expressed in CD4
positive T cells, although it was found in synovial CD4 positive T cells from
patients suffering from rheumatoid arthritis (Griffiths et al., 1992).
After cleavage of the 20 (mouse) or 21 (human) amino acid long signal peptide
with an amino acid sequence typical for secretory proteins, the mature protein
consists of 534 amino acids with a calculated molecular weight of about 60 kDa for
the unmodified peptide core. Probably due to glycosylation, it migrates with an
apparent molecular weight of 65-70 kDa under non-reducing and with 70 kDa
under reducing conditions on acrylamide gels. Two potential N-linked
glycosylation sites are conserved in both species; murine perforin contains an
additional third potential glycosilation site. Human and mouse perforin share 68%
amino acid identity. All 20 cysteine residues are completely conserved between
these different perforin species. With the exception of the cysteins homologous to
the LDL class B or the epidermal growth factor precursor domains, the disulfide
bridges are not known.
21
100 200 300 400 sog 600—1
PERFORIN
0 100I I L.
aoo, 309 400 500 60p 709 800
1 ' 909
C6 I tsp I I TSP I fTl-s-s-i
C7 -fTSPTfT
C8a*
ri^r
is^r
TSP 1» A
iSsTST
C8P { tsp 1 )|"TEiP~
C9 -TtspT) C5ST" X
-ILBis-s
t-i~s«s-»
-(LB>CT
s-ss-s->
B fl TSP I \-r-4SCR1 YSCR2
-ft*V¥T
B< TSP I
sVs «—S-S—i
-tlfll—¥T
8)( TSP I }-•S-S—'
-»LB>T
<D-
fimi F1M2
Fig.4. Structural organisation of human perforin in comparison to late complement
components. (Figure taken from Tschopp and Nabholz, 1990). The various
structural motives present are: TSP 1, type 1 thrombospondin module; B, LDL
receptor class B module (EGF module); SCR, short consensus repeat; FIM, factor
1 module; LB, candidate for lipid binding protein. Only cysteines outside motives
or differing from invariant cysteines within motives are indicated. By analogy to
the complement proteins, Cys 236 and 258 of perforin are likely to be disulfide-
bonded. Putative asparagme-linked giycosylation sites are shown as black
hexagons. Regions in perforin which show no detectable homology to complement
proteins are boxed in with thick solid lines.
22
Comparison of the amino acid sequences of C6, C7, C8cr, C8/J and C9 reveals no
detectable homology between the hundred N-terminal amino acids on the C-
terminus of perforin and complement components. However, Perforin and
complement share a central region covering approximately 280 amino acids or two
thirds of the perforin molecule. In this region, about 20% of the amino acids are
shared by all of these molecules. This suggests that this segment is responsible for
the structural and functional similarities between C9 and Perforin, in particular,
for the capacity to polymerize into similar tubules and to insert into membranes.
The region contains two structural elements that have been identified previously:
(a) Amino acids 191 - 211 (LB) are the most conserved residues in the entire
sequence, and the corresponding residues in C9 have been proposed to form an
amphipathic or-helix that interacts with the lipid bilayer. (b) The central segment
common to complement proteins and perforin contains a cysteine-rich region of
LDL receptor class B (or EGF precursor) type. The function of this module,
present in many other molecules such as coagulation proteins and cell surface
receptors, is not known.
In the presence of Ca2+ and at 37°C perforin interacts with the phospholipid
headgroups of the target cell membrane, inserts into the membrane and
polymerizes in the membrane irreversibly (Yue et al., 1990; Young et al., 1987). At
4°C, insertion and polymerization do not take place and perforin can be removed
from the membrane by chelation of Ca2+ or by high salt concentartions. It is
thought that a conformational change takes place during membrane insertion and
polymerization. This leads to the exposition of a hydrophobic surface, enabling
perforin to insert into the membrane and to traverse it.
In the target cell membrane perforin is found as a tubular polymer with an
apparent molecular weight of approximately 1000 kDa. Between 12 and 18
monomers can form a cylinder with a defined height of 16 nm and a variable
diameter between 5 and 20 nm (Fig.5). In the presence of 1 mM Ca2+, purified
perforin exhibits potent lytic activity against nucleated cells and against red blood
cells.
23
ca^Jt-JE!:PI poly PI
CSU-8+C9 C5b-9 MAC
Fig.5. Formation of the pofy perforin pore in the membrane of a target cell The
scheme depicts the binding of perforin to the membrane, its insertion after
undergoing a conformational change and the polymerization in the membrane
compared to the assembly of the membrane attack complex (MAC) by the
terminal complement components. The figure is taken from (Podack et al., 1991).
Introduction of Targeted Germline Mutations into Mice by Homologous
Recombination in Embryonic Stem Cells
Molecular cloning techniques have in the recent twenty years allowed the cloning
and sequencing of a cornucopia of mammalian genes. However for many of these
genes, it has proven difficult to establish a clear role and function in vivo. Naturally
occurring mutations were of big help to clarify the role of some genes in vivo. But
for most genes no mutant mouse strains are available and the possibilities to
search for such mutations are quickly limited by restrictions in time and breeding
space. Especially recessive lethal mutations are very rare in the collection of
isolated spontaneous mutant mice. In the light of this situation the interest for a
technique that allows the targeted introduction of desired mutations into the
germline of mice is understandable. The making of mice deficient for a gene of
interest allows a direct assessment of its function in vivo.
Establishment of this techniques relied on a number of progresses made in the
1980s, namely the establishment of pluripotent embryonal stem cells (ES cells)
from mouse blastocysts (Evans and Kaufman, 1981), the generation of germline
24
chimeras from ES cells (Bradley et al., 1984), successful homologous
recombination in eucaryotic cells (Smithies et al., 1985) and later ES cells
(Thomas and Capecchi, 1987; Doetschmann et al., 1987), the preservation of
pluripotency through the electroporation procedure and finally the establishment
of germline chimeras with targeted ES cells (Hooper et al., 1987; Thompson et al.,
1989; Koller and Smithies, 1989).
The procedure, as it is today well established (reviewed by Capecchi, 1989a;
Hogan and Lyons, 1988; Robertson, 1986; Capecchi, 1989b; Rossant and Joyner,
1989; Fung-Leung and Mak, 1992), is outlined in Fig.6.
Pluripotent embryonic stem cells are transfected with the targeting vector
containing the desired mutation. This mutation is flanked by homologous
sequence, which mediates the recombination at the desired genomic locus. In only
a little fraction of all ES cells, which incorporate a DNA construct into their
genome, the integration takes place at the targeted locus, in most other cells the
vector integrates at a random site into the genome. The rare targeted cells are
identified by a screening procedure, usually by polymerase chain reaction (PCR),
cloned and maintained as a homogeneous population. These clonal cells are
injected into the blastocoel cavity of a preimplantation mouse embryo and the
blastocyst is surgically reintroduced into the uterus of a foster mother where
development is progressing to term. The resulting animal is chimeric in that it is
composed of cells derived from the blastocyst, here denoted by the white fur
colour, and the ES cells denoted by the black fur colour (Fig.6). Breeding of the
chimera to an albino mouse yields a certain proportion of black mice indicating
that the ES cells contributed to the formation of the germline in the chimeras.
Genomic screening of the progeny is used to determine which mice received the
mutated allele and can bred in order to obtain animals homozygous for the
mutation.
25
Transfection with
a targeting vector
m
(microinjection orksvafflsys»aeysysj
Embryo-derived eiectroporation) ES cell culture with
stem (ES) cells rare targeted cell
Screening/or enrichment for |targeted cell
Injection of ES cells
into blastocyst
Implantation into
foster mother
Pure population of
targeted ES ceils
Chimeric mouse
Breed with
+/+ animals
Germ-line transmission of
ES cell genome containingtargeted modification
Fig.6. Outline of the procedure to obtain mice with a targeted mutation in a gene of
interest by homologous recombination in ES cells. ES cells with a targeted mutation
(denoted by the black colour) are selected, amplified and injected into blastocysts
of a white furred mouse strain. Chimeras derived from these manipulated
blastocysts by transfer into the uterus of a pseudopregnant foster mother can pass
their genotype indicated by the black fur colour to offsprings. To generate
deficient mice, heterozygous offsprings are bred to homozygousity. Illustration
taken from (Capecchi; 1989a).
26
There are two different approaches currently applied to induce the desired
genomic alteration. Insertion constructs create a partial duplication of the gene
and require only a single crossover, while replacement vectors lead to the
exchange of the endogenous DNA with the construct and involve a double
crossover. Ah interesting application of insertion type vectors is the introduction of
subtle point mutations into a locus of interest by the hit and run method (Hasty et
al., 1991; Nickoloff and Reynolds, 1990).
For standard gene inactivation experiments, where a coding exon of the gene is
interrupted by a neomycine resistance expression cassette, usually the replacement
type vector is chosen. In this case the targeting construct consists of a stretch of
homologous DNA varying in length interrupted by a neomycine resistance gene
that allows the selection of transfected ES cells by growing them in G418
containing medium. The neomycine resistance gene that is introduced into a
coding exon of a targeted gene not only serves to provide a resistance marker for
cells with an integrated construct but also interrupts the coding sequence of the
targeted gene. It displays multiple stop codons in all three reading frames leading
to a premature stop of translation and a noncomplete, nonfunctional protein
product. In addition, transcriptional inactivation of the targeted gene by the
introduced expression cassette has been observed as well (Zijlstra et al., 1990).
An enrichment for homologous recombinant cells has been achieved by the
addition of one or two Herpes Simplex Virus thymidine kinase genes to the ends of
the construct. (Mansour et al., 1988). Homologous recombination leads to the loss
of the kinase genes and to a resistance of the transfected cells towards the selecting
compound ganciclovir. In contrast, randomly targeted cells obtain one or two
copies of the thymidine kinase gene and thereby get sensitive to ganciclovir.
The influence of transcription of the targeted locus in the ES cells on the efficiency
of homologous recombination is not clear. Early experiments did not indicate any
enhancement of homologous recombination by transcription (Johnson et al., 1989)
while others later reported such an effect (Farrar et al., 1980). The considerable
number of successful targeting experiments with loci not transcribed in ES cells
27
suggests, that transcription has only a limited, if any, influence of transcription on
the recombination efficiency. If transcription of the targeted locus in ES cells is
known, this can be exploited to enhance homologous recombination by applying a
promoterless construct (Sedivy and Sharp, 1989; Schwartzberg et al., 1990). In this
case, the resistance gene is preferentially expressed at the targeted locus'while it
remains silent at most random integration sites. This results in an apparent
increase of the recombination frequency.
Gene targeting has had already a strong impact on many fields of biology, most
notably on developmental biology and immunology. It should not be concealed,
that the analysis of mutant mice is sometimes not trivial, because the complexity of
the mammalian organism. Networks of gene interactions, pluripotency of action
and redundant functional mechanisms can obscure the effect of a mutation and
prevent the insight into the immediate function of a gene. Nevertheless, the ability
to access and mutate virtually every nucleotide in the mouse genome has provided
and will provide a valuable tool to carry out a comprehensive genetic analysis of a
mammalian organism.
In the first part of this thesis, perforin expression during infection of mice with
lymphocytic choriomeningitis virus was investigated. Kinetics and histological
localization of expression provided evidence for an involvement of perforin in
antiviral cytotoxicity in vivo. In part two, mice with a targeted disruption of the
perforin gene were generated by homologous recombination in embryonic stem
cells. Perforin deficient mice will allow a direct functional assessment of perforin
protein in vivo. Especially, they will help to address the controversial question,
whether several independent mechanism of cytotoxicity exist or whether perforin is
the sole effector molecule of cytolytic lymphocytes.
28
5. Perforin Expression during Infection of Mice with
Lymphocytic Choriomeningitis Virus
5.1 Detection of Perforin and Granzyme A mRNAs in Infiltrating
Cells during Infection of Mice with Lymphocytic
Choriomeningitis Virus
Christoph Muller*, David Kagi+, Toni Aebischer+, Bernhard Odennatt"1",Werner
Held*, Eckhard R. Podack*, Rolf M. Zinkernagel+, Hans Hengartner+
'Department of Pathology, University of Bern, 3010 Bern
+Institute for Experimental Pathology, University Hospital, 8091 Zurich
SDepartment of Microbiology and Immunology, University of Miami, Miami
33101
Reprint from European Journal of Immunology (1989), 19:1253-1259
29
Summary
The analysis of gene expression in cytotoxic T-cells by in situ hybridization of serial
liver and brain sections from mice infected with lymphocytic choriomeningitis virus
and immunostaining with T-cell-marker- and virus-specific antibodies revealed a
close histological association of infiltrating lymphocytes expressing the perforin
and granzyme A gene with virally infected cells.Maximal frequency of perforin
and granzyme A mRNA containing cells on liver sections preceded by about two
days maximal LCMV-specific cytotoxicity of the lymphoid liver infiltrating cells.
These results are most consistent with an involvement of perforin and granzyme A
in cell-mediated cytotoxicity in vivo.
Abbreviations: LCMV: Lymphocytic choriomeningitis virus LU: Lytic unit
NK: Natural killer cell CTL: Cytotoxic T-lymphocyte
30
Introduction
Cytotoxic lymphocytes seem to play an important role in immune defense against
cell associated antigens. Specific cytotoxic CD8+ T-cells are the main effectors
against virus infected cells (Doherty and Zinkernagel, 1974; Blanden, 1974). The
same class of effector cells and natural killer cells are involved in anti-tumor
immunity (Timonen et al., 1988). Antibody dependent cell mediated cytotoxicity is
conferred by a variety of Fc-receptor positive lymphocytes including NK-cells.
Their lytic mechanism has been studied extensively during the past 10 years
(Moller, 1988).
Analysis of the contents of the large cytoplasmic granules, isolated from cloned
cytotoxic T-cells and NK cells grown in vitro, allowed the characterization and
intracellular localisation of various candidates for effector molecules in cell
mediated cytotoxicity, including perforin (Dennert and Podack, 1983; Young et al.,
1986b; Young et al., 1986d; Shinkai et al., 1988; Lowrey et al., 1989; Podack et al.,
1988) which is also named cytolysin (Henkart et al., 1984) or pore forming protein
(Young et al., 1986c) and several serine proteases (Pasternack and Eisen, 1985;
Simon et al., 1986; Masson and Tschopp, 1987; Moller, 1988; Jenne et al., 1988).
Immunoelectron-microscopical studies have shown, that perforin and granzyme A
are stored in the same electron-dense cytoplasmic granules in cultured cytotoxic T
cell clones (Groscurth et al., 1987; Jenne et al., 1988). Purified perforin was shown
to be hemolytic and cytolytic for various tumor cell lines in vitro at low
concentrations (Young et al., 1986b; Henkart et al., 1984). The role of perforin in
cell mediated cytolysis in vivo is controversial (Clark et al., 1988; Ostergaard and
Clark, 1987; Berke and Rosen, 1987; Young et al., 1987; Clark, 1988), because so
far all attempts have failed to demonstrate perforin directly in primary cytotoxic
cells with polyclonal antisera. Perforin was also not found in alloreactive, highly
cytotoxic peritoneal exudate lymphocytes by immunohistochemistry and Western
blot analysis (Berke and Rosen, 1987). Furthermore, granules of cytotoxic T-cells
from primary mixed lymphocyte cultures, from in vivo poly IC induced NK or
peritoneal exudate cells failed to exhibit any lytic activity when tested on sheep
31
erythrocytes (Dennert et al., 1987). However, Northern blot analysis using cloned
murine perforin cDNA (Lowrey et al., 1989) as a probe elucidated a strict
correlation between the presence of perforin mRNA and cytotoxicity in primary
and cloned cytotoxic T-lymphocytes. The levels of the cytolytic activity of the
various killer cell types did however not correlate strictly with the levels of perforin
mRNA (Podack et al., 1988).
The involvement of serine proteases in cell-mediated cytotoxicity was implied by
the blocking of cytotoxicity with protease inhibitors (Chang and Eisen, 1980;
Redelman and Hudig, 1980; Acha-Orbea et al., 1983). In the meantime, at least 6
different serine proteases, which may constitute up to 1% of the total cellular
protein, have been isolated from cytotoxic T-cells (Pasternack and Eisen, 1985;
Masson and Tschopp, 1987; Simon et al., 1986; Young et al., 1986a; Pasternack et
al., 1986; Podack and Konigsberg, 1984; Gershenfeld and Weissman, 1986; Cole et
al., 1971). Recent comparison suggest that several of those may be identical. As for
perforin, the role of serine proteases in cytotoxic activity in vivo is still debated
(Dennert et al., 1987; Brunet et al., 1987; Ostergaard et al., 1987).
Infection with lymphocytic choriomeningitis virus (LCMV) in the mouse provides
an excellent experimental model to investigate the cellular and molecular events
during cell mediated cytolysis in vivo because this model infection is well studied
(Lehmann-Grube et al., 1985; Lehmann-Grube, 1971; Zinkernagel et al., 1986;
Hotchin, 1971) and because T-cell mediated damage of infected host cells is
histologically well localized (Lehmann-Grube and Lohler, 1988). Intracerebral
inoculation of LCMV in mice causes fatal choriomeningitis, characterized by a
well defined mononuclear infiltrate along the virus infected cells of the meninges
and of the chorioid plexus (Lehmann-Grube and Lohler, 1988; Cole et al., 1971).
Intravenous injection of the hepatotropic strain LCMV-WE leads to a massive
mononuclear infiltration of the liver. An increase of serum levels of liver enzymes
like alanine aminotransferase or glutamate dehydrogenase signals an ongoing
destruction of hepatocytes (Zinkernagel et al., 1986).
32
We used antisense RNA probes, which are complementary to naturally occurring
mRNA, to assess the expression of the perforin and granzyme A genes in the
infiltrating cytotoxic cells during a LCMV-infection of the liver and the brain by in
situ RNA-hybridization. This method has already been used successfully to
demonstrate the expression of the two serine protease genes granzyme A and B
(HF (Gershenfeld and Weissman, 1986) and Cll (Lobe et al., 1986)) in the
cytotoxic effector cells during an allograft rejection in vivo (Mueller et al., 1988).
We provide direct evidence for activation of the perforin and the granzyme A
genes in LCMV specific infiltrates during virus infection.
33
Material and Methods
Mice
Inbred C57BL/6 were obtained from the Institut fur Zuchthygiene, Tierspital
Zurich. Mice of either sex were used at an age of 8-16 weeks.
Virus
The LCMV isolate was originally obtained from Prof.F.Lehmann-Grube,
Hamburg, Federal Republic of Germany, and was subsequently propagated on
L929 cells. LCMV Armstrong was obtained from Dr.M.Buchmeier, Scripps Clinic
and Research Foundation, and was propagated on BHK cells. LCMV was titrated
in vivo. Mice were infected i.v. with 3 x 105 pfu of LCMV-WE or i.c. with 300 pfu
ofLCMV-ARM.
Cytotoxic T-cell assay
Single spleen cell suspensions from C57BL/6 mice were incubated with 104 LCMV
infected or uninfected MC57G (H-2b) or YAC-1 (H-2a, a NK-cell susceptible
thymoma line) target cells at various ratios in round bottomed 96 well plates
(Petra Plastik; Inotech, Wohlen, Switzerland) for 5 hrs at 37°C in air supplemented
with 5% C02. Specific 51Cr-release from target cells was calculated by correcting
for spontaneous 51Cr-release. Maximal lysis was determined in wells containing
targets treated with IN HCl without effector cells. Lytic units,a measure of
relative cytolytic activity, were determined to be the number of lymphocytes
necessary to lyse one third of the standard number of target cells (104 cells/wells)
during the standard test duration. LU per spleen or liver were compared by
dividing the total number of lymphocytes per organ by the number of lymphocytes
for 1 LU (Cerottini and Brunner, 1974).
Immunohistochemistry
Sections were stained with the monoclonal antibodies 53.6.7 (anti-Lyt-2), GK 1.5
(anti-L3T4) (both hybridomas obtained from the American Type Culture
34
Collection, Rockville, MD) and VL-4 (anti-LCMV, gift from J. Banziger,
unpublished) as a first stage and alkaline phospatase-conjugated goat-anti-rat IgG
(Sigma), used at a 1:50 dilution in TBS containing 10% normal mouse serum, as a
second stage reagent. Naphthol-As-Bi phosphate (Sigma) together with neufuchsin
(Fluka) was used as a substrate for alkaline phosphatase. Levamisole was added to
the substrate solution to inhibit endogenous phosphatases.
Preparation oflabeled RNA-probes
A 521 bp cDNA fragment of the 5' end of the murine perforin gene (position 255 -
766) (Lowrey et al., 1989) was subcloned as an Eco RI-HindDI-fragment into the
polylinker of the transcription vector pGEM-2 (Promega Biotech. Madison WI). A
750 bp cDNA fragment of the granzyme A gene (HF-gene) (Gershenfeld and
Weissman, 1986; Mueller et al., 1988) was subcloned as an EcoRI - Sail fragment
into pGEM-2 using standard techniques. After linearization of the plasmids with
appropriate restriction enzymes, antisense RNA probes of the granzyme A and
perforin gene, and sense probe of the perform gene were prepared using the T7
RNA polymerase, and SP6 polymerase, respectively. A typical reaction (35 fil)
contained 8.4 /d 35S-UTP 800 Ci/mMol (final concentration: 12 fiM) (Amersham),
7 fil 5x SP6 buffer (40 mM Tris-HCl pH 7.9, 6mM MgC^, 2 mM spermidine); 3.5
111 100 mM dithiothreitol; 3.5 Ml ribonucleotides (CTP, ATP; and GTP, 10 mM
each); 3.5 fil BSA 5mg/ml; 1 fil Rnasin 40 U//d(Boehringer); 1 fil SP6-RNA
polymerase 20 13/id (or T7-RNA polymerase (both from Boehringer, Mannheim,
FRG); 1 ill linearized DNA template (llig/fil); 6.1 ill Hp. SP6- and T7-reactions
were incubated for 90 min at 40°C, and 37°C, respectively. DNA template was
digested with Dnase I (Worthington) for 15 min at 37°C. Labeled RNA probes
were precipitated twice and resuspended at a concentration of 2X106 cpm/jul in TE,
boiled for 2 min and stored at -70°C for up to two weeks. The labeled probe was
mixed for hybridization at a concentration of 2 x 105 cpm/fil with deionized
formamide (final concentration 50%), dextran sulfate (10%), dithiothreitol
35
(lOOmM), NaCl (300 mM), Tris-HCl, pH 7.5 (20 mM) EDTA pH 8.0 (5 mM),
Denhardt's solution (lx).
In situ hybridization
In situ hybridizations of cryostat sections of the brain and liver from LCMV-
infected C57B1/6 mice were done as previously described in detail (Mueller et al.,
1988). Hybridized tissue sections were dipped into NTB-2 nuclear track emulsion
(Eastman Kodak, Rochester, NY), diluted 1:2 with 600 mM ammonium acetate.
Sections hybridized with a 35S-labeled RNA antisense probe of the granzyme A
gene were exposed for 15 days at 4°C in a light-tight box. Sections hybridized with
35S-labeled RNA antisense probe of the perforin gene and control slides
hybridized with a 35S-labeled RNA-sense probe of the perforin gene were all
exposed for 30 days in the dark. Slides were developed with Kodak developer PL-
12 for 2.5 min and fixed with Kodak fixer for 5 min at room temperature.
Counterstaining was done with nuclear fast red (0.05% in 5% aluminium sulfate).
Evaluation ofthe sections
A microscope with a grid in one of the two oculars was used for evaluation of the
liver sections. Cells located on, or tangential to one line of the grid were
considered for evaluation. For the assessment of the frequency of cells containing
transcripts of the perforin or the granzyme A gene at least 2000 cells were
considered per mouse. Cells were considered positive for gene expression when
they showed at least twice as many silver grains as the cells with the highest
background on control slides hybridized with a sense probe of the perforin gene.
Generally, less than ten grains per cell were found on these control sections. For
the assessment of the frequency of CD4+-,CD8+-cells and virally infected
hepatocytes, at least 1000 nucleated cells per animal were considered for
evaluation.
36
Results
LCMV-induced T-cell mediated hepatitis.
C57BL/6 mice were inoculated with 3 x 105 pfu of the hepatotropic strain LCMV-
WE. Animals were sacrificed at 4, 6, 8, and 10 days after virus inoculation. For the
immunohistochemical analysis and in situ-hybridization the livers of 4 animals
were snap-frozen in liquid nitrogen at each time point (except on day 10 when only
2 animals were used). For the determination of the cytolytic activity of the
splenocytes and the liver infiltrating cells of LCMV-WE infected mice, groups of 8
animals were sacrificed on day 4, 6, 8 and 10, respectively.
The inflammatory infiltrate of the liver increased steadily until day 8 after i.v.
infection with LCMV-WE; it had decreased again by day 10, i.e. the end of the
observation period (Fig.7). The histologically observed kinetics of infiltration was
also reflected by the number of lymphoid cells isolated from the liver of these
animals, which peaked on day 8 after inoculation (data not shown). The LCMV-
specific cytolytic activity of unseparated splenocytes from infected animals tested
on LCMV-infected fibroblasts was maximal on day 8 and dropped sharply
thereafter (TABLE 1 and FIG.8). The NK-cell activity of isolated liver-infiltrating
cells and unseparated splenocytes determined on YAC-1 target cells was low
during the entire observation period and did not show the distinct kinetics
observed for the specific cytotoxicity against LCMV-infected target cells (TABLE
1) (Mclntyre and Welsh, 1986). The determination of the anti-LCMV-specific
cytotoxicity (expressed as lytic units per liver) of the lymphoid cells isolated from
the infected liver revealed a similar time course as the measured cytotoxicity of
unseparated splenocytes from the same animals (FIG.8). Calculated lytic units
were maximal on day 8 and decreased thereafter. The number of liver infiltrating
lymphoid cells on day 0 and day 4 after LCMV-infection was too low for a
determination of lytic units per liver.
37
The morphology of the cryostat sections after immunostaining or in situ
hybridization did not allow an unambiguous distinction between the infiltrating
cells and hepatocytes. Thus, we determined the percentage of cells, positive for T-
cell markers, LCMV, or expression of the granzyme A or perforin gene from the
total number of nucleated cells on a liver section. The phenotypic analysis of T-
lymphocytes in the liver revealed a massive infiltration of mainly CD8+ cells after
more than 4 days after inoculation. Based on the immunostaining of serial sections
with an anti-CD4 and an anti-CD8-antibody respectively we found a CD4+ : CD
8+ ratio of approximately 1.7:1 on day 4; 1:7 on day 6; 1:5 on day 8, and 1:3 on
day 10 (Fig.9).
38
TAB.l. Percent specific 51Cr-release from target cellsa>b)
Splenocytes
MC57G MC57G YAC-1
uninfected LCMV-infected
E/T ratio 50 17 6 2 50 17 6 2 50 17 6 2
Day 4 1 0 0 0 10 1 5 0 11 8 3 4
Day 6 1 2 1 4 37 17 9 6 8 6 2 3
Day 8 3 0 1 4 75 41 20 13 16 5 3 2
Day 10 0 0 0 1 49 37 18 8 13 7 5 7
Liver-derived Lymphocytes0-)
E/T ratio 30 10 3 1 30 10 3 1 30 10 3 1
Day 6 3 0 2 3 85 56 21 19 18 7 7 2
Day 8 6 0 0 3 83 29 21 17 6 4 0 5
Day 10 5 0 0 0 80 46 20 13 18 2 9 2
a) C57 Bl/6 mice were infected intravenously with 3x10s pfu of LCMV-WE
b) 104 51Cr-labeled target cells were incubated at the indicated effector : target
ratio for 5 hours at 37°C. Data represent means of duplicates. Spontaneous
release values were below 33% of maximal release for uninfected MC57G and
LCMV-infected MC57G, and less than 17% for YAC-1 cells.
c) Splenocytes of three animals were pooled and tested in duplicates at each
timepoint.
d) Liver-infiltrating lymphocytes were isolated from 8 animals at each timepoint
as described in Experimental Procedures.
39
:\**'V:»' - •
*• •
• * •
4 ''%*•.£'
• tM/'.'V'-..* i*V«* "A
FIG.7. Histology of the liver at different timepoints after LCMV-infection.
Cryostat sections of the liver from C57 Bl/6 mice on day 0 (uninfected) (A) and
day 8 (B) after intravenous infection with 3 x 105 pfu of LCMV-WE were stained
with H&E using standard techniques
40
120- -200C/i
3S
Oo
o
JS 40J
to
o
Time after infection (days)
FIG.8. Virus-infected cells and specific cytotoxicity in the liver after LCMV-infection.
LCMV-infected liver cells ( -A ) were detected by immunostaining of cryostat
sections at different timepoints after intravenous infection with 3 x 105 pfu of
LCMV-WE with the LCMV-specific antibody VL-4. Lytic units per liver ( ), as a
relative measure of cytotoxicity were determined as described in Experimental
Procedures.
41
J2 2401
5 200
"2 160
I 120i
Ml
ell 80<j
o
40C/J
c
Ci.0
8 10
Time after infection (days)
FIG.9. T lymphocyte subset analysis of cells infiltrating the liver after LCMVinfection.
The numbers of CD4+( ^) - and CD8+( ) cells per thousand nucleated liver
cells were determined by immunostaming with the monoclonal antibodies
H129.19, and 53.6.7, respectively, on cryostat sections of the liver at different
timepoints after intravenous infection of C57B1/6 mice with 3 x 105 pfu of LCMV-
WE.
42
Maximum expression of the perforin gene precedes maximum LCMV-specific
cytotoxicity in infected livers.
For the detection and localization of cells expressing the perforin or the granzyme
A gene on tissue sections we prepared 35S-labeled RNA-antisense probes of a 521
bp EcoRI - Hind in -fragment of the perforin gene and a 775 bp EcoRI - Sail
fragment of the granzyme A (HF) gene, respectively. Cells expressing the
granzyme A gene were found on all liver sections from day 4 on. The frequency of
infiltrating cells containing transcripts of the granzyme A gene at a detectable level
increased dramatically between day 4 and 6. Thereafter the number decreased
again, and on day 10 cells with transcripts of the granzyme A gene were as
frequent as they were on day 4 (Fig. 10). The expression level of the granzyme A
gene (determined as the number of silver grains over a positive cell) was also
maximal on day 6, thereafter it decreased to a level on day 10 that was lower than
on day 4 (quantitative analysis not shown). Hybridization of the liver sections with
the perforin antisense probe revealed a comparable time course for the expression
of this gene in the infiltrate of LCMV-infected livers. Cells possessing transcripts
of the perforin gene were absent on liver sections from uninfected animals and
only very few positive cells were found on day 4. The frequency of perforin positive
cells peaked on day 6 and, similar to the results obtained with a probe of the
granzyme A gene, declined thereafter (Fig. 11). The largest amount of perforin
mRNA in the infiltrating cells was also found on day 6 after LCMV-infection, i.e.
two days before the number of lytic units against LCMV-infected targets was
maximal in livers. The signal of the in situ hybridization with the perforin probe,
was at least ten fold weaker than the signal obtained with the granzyme A probe
throughout the entire observation period. Some variation in the frequency of
granzyme A and perforin expressing cells was seen among the individual animals
at each time point. However, in every animal examined 6 days after infection, the
frequency of granzyme A and perforin expressing cells was at least twice as high as
in any animal examined at day 4 or 8 after infection.
43
«512Ch
c/i
C/l
c
8 10
Time after infection (days)
FIG. 10. Granzyme A and perforin expressing cells infiltrating the liver in LCMV-
infected mice. Number of cells (per thousand nucleated cells) containing transcripts
of the perforin ( ) and the granzyme A gene ( • ) detected by in situ
hybridization of liver sections with 35S-labeled RNA antisense probes at different
timepoints after intravenous infection of C57B1/6 mice with 3 x 105 pfu of LCMV.
As a negative control, slides were hybridized with a 35S-labeled RNA sense probe
of the perforin gene ( +). At least 2000 cells were evaluated at each timepoint.
B
*«*•» *. « *
'.
•*' • •* '
*
*
'
'•*• V. *.. • ,;* * v* >v* '**
FIG. 11. Perforin expression by infiltrating cells in the liver of LCMV infected mice.
Cryostat sections of the liver at day 0 (uninfected) (A), day 6 (B) and day 10 (C)
after intravenous infection of C57B1/6 mice with 3 x 105 pfu LCMV-WE were
hybridized with a 35S-labeled RNA antisense probe of the perforin gene as
described in Experimental Procedures. As a negative control, a liver cryostat
45
section of the same animal as in (B) was hybridized with a 35S-labeled sense probe
of the perforin gene.
Association of viralfy infected cells with cells expressing the genes for perforin and
granzymeA in infected liver and brain tissue.
For the analysis and correlation of surface markers, LCMV infection and gene
expression of cells in LCMV infected livers and brains serial sections of tissue were
prepared at different time points after intravenous and intracerebral infection with
LCMV. The sections were subsequently used, in the following order, for in situ
hybridization with an antisense RNA probe of the granzyme A gene,
immunoperoxidase-staining for CD8, in situ hybridization with an antisense RNA-
probe of the perforin gene, immunoperoxidase-staining for LCMV,
immimoperoxidase-staining for CD4 and in situ hybridization with a sense RNA
probe of the perforin gene as a negative control.
C57BL/6 mice, inoculated i.c. with 300 pfu of LCMV-Armstrong showed clinical
signs of severe paresis and convulsion on day 6 and died between day 7 and 8 after
infection. Histologically, the choriomeningitis was characterized by a well defined
infiltration along the meninges, the ependymal cells and the choroid plexus. On
days 5 and 6 in some areas up to 80%-95% of ependymal and meningeal cells were
infected with LCMV. At the later time points, some animals showed already
histological signs of cellular destruction of these LCMV-infected cells. In the brain
of infected animals on day 5 the infiltrating T-cells were mainly of the CD8+
phenotype (CD4+:CD8+ = 1 : 1.8), while no infiltrating CD4+ or CD8+ were
detected in the brain of uninfected controls. The inflammatory infiltration was
much more pronounced on day 6 and the CD4+ : CD8+ ratio further dropped to 1
: 3.2. Cells expressing the granzyme A gene at a low level were found in the
infiltrate in small numbers on day 5, whereas infiltrating cells expressing the
perforin gene at detectable levels were observed one day later, i.e. approximately
one day before the infected animals died on day 7.
46
Since we were unable to obtain reliable results when we tried to combine
immunostaining with monoclonal antibodies with in situ hybridization, we used
serial sections of the brain to correlate the site of LCMV-infected cells with the
infiltrating CD4+ and CD8+ cells, and also with the site of granzyme A and
perforin expressing cells. Fig.12 shows the result of the immunostainings and in
situ-hybridization of serial sections from a representative area of the
choriomeningitis of an animal 6 days after intracerebral inoculation with LCMV-
Armstrong.
CD4+ and CD8+ cells were always closely associated with LCMV-infected cells.
The frequency of granzyme A and perforin expressing cells was considerably
higher than the number of CD4+ cells in all animals on day 6 after infection and
the distribution of the positive cells on sections that were hybridized with RNA-
probes of the granzyme A or the perforin gene was very similar to the distribution
of CD8+ cells.
Serial sections were also prepared from liver tissue at different time points after
intravenous infection with 3 x 105 pfu LCMV-WE. In contrast to the
choriomeningitis, in LCMV induced hepatitis the virus infected cells were not
histologically confined to the mononuclear, perivascular infiltrate, but rather were
found also in parenchymal tissue. Perforin, Granzyme A, CD8 and CD4 positive
cells were scattered in the same pattern over the entire area of the liver section
(Data not shown). The frequency of virus infected cells was maximal on day 6,
when also the frequency of perforin mRNA containing cells was maximal in the
liver infiltrate (Fig.8).
47
•••• ''I*' : .>•> V •
'
.
,*4. •
,«r: . •
<**.
• • >•
_*»* " t" ** .rt«*»
•^v -•
••• SLIP ^ ..-\. *'•
^ •• ••^•/; ,\>.***
#,.TT «*3(
'"
«..
' '•„
. * >
* - '%»
i£^ • =,* <- .>> *v*& v^;, ^- -*>y:
V •.»•
-v ^f'** T|»'•
• j. ••
-;V J^-. :•/•»•./.•:'-<^3tf ..:
< »
^ •
..." *"!
•v
• >
r
^ \$- *#:
flWf
.X
^1
*.ft.* ^*>*~-i;
v## ^
A
C
E
A'
4- #>
B
D
F
C
#^
48
FIG.12. (previous page) Immunostaining and in situ-hybridization showing the same
histological situation on serial brain sections from LCMV-infected mice. C57B1/6
mice were inoculated intracerebrally with 300 pfu of LCMV-Armstrong. Serial
sections of the brain were prepared 6 days after infection and were used for:
(A) hybridization with 35S-labeled RNA antisense probe of the granzyme A
gene
(B) immunoperoxidase-staining with the anti-CD8 mAb 53.6.7
(C) hybridization with 35S-labeled antisense probe of the perforin gene
(D) immunoperoxidase-staining with the anti-LCMV mAb VL-4
(E) immunoperoxidase-staining with the anti-CD4 mAb H129.19
(F) hybridization with 35S-labeled RNA sense probe of the perforin gene as
a negative control. (See Experimental Procedure for details.)
(A') Detail from section (A).
(C) Detail from section (C), histological situation corresponds to (A').
49
Discussion
The present investigation of perforin gene expression is good evidence for an
involvement of perforin in cell mediated cytotoxicity in vivo. The close association
between perforin (and granzyme A) positive cells and LCMV-infected cells in
infected livers and brains is a strong indication of a direct involvement of perforin
containing cells in the elimination of virus infected cells.
The frequency and the signal intensity of perforin versus granzyme A mRNA
positive cells differed considerably. This was found on all sections and cannot be
explained solely by the different size of the cDNA fragments used for the
preparation of the probes. Most likely, this difference reflects differing amounts of
free mRNA available for hybridization and that there are more T-cells expressing
serine proteases than perforin. We cannot exclude that some cells with very few
mRNA copies of the perforin gene were below the detection level of our method.
Therefore the frequency of perforin mRNA positive cells may be an
underestimate.
The expression of the perforin gene was found to be slightly higher in infiltrating
cells of the brain than in the liver. This, however, reflects either a difference in the
actual copy number of perforin mRNA per cell or a more rapid degradation
caused by endogenous RNA degrading enzymes in liver tissue.
The number of infiltrating lymphocytes and the lytic LCMV-specific T-cell activity
expressed as LU per liver clearly peaked on day 8 (Fig.8), although the relative
cytolytic activity measured on a cell to cell basis was comparable on days 6, 8 and
10 (TABLE 1). The delay between the peak of T-cell cytotoxicity on day 8 and
perforin expression peaking on day 6 is most readily explained by a delay of
production and storage in granules versus release. Thus after the initial activation
of the CTLs and induced transcription of the perforin gene sufficient perforin was
synthesized in LCMV-specific CTLs for optimal cytolytic activity probably without
any new transcription of the perforin gene during the following 24-48 hours.
Perforin is apparently a storage protein as was shown by immunohistochemistry
and indirectly by the fact that cytotoxic activity does not require any protein de
50
novo synthesis (Thorn and Henney, 1988). The contribution of NK-cells to the
measurement of 51Cr-release and the implied elimination of virus infected cells in
the liver appeared to be minimal as indicated by the low lysis of NK-sensitive
YAC-1 target cells (Mclntyre and Welsh, 1986).
A key question is obviously, whether only CD8+ T-cells produce and release
perforin. We tried to address this question by a combination of immunostaining
against T-cell surface markers and subsequent in situ hybridization. Unfortunately,
cytoplasmic mRNA was not retained well enough in tissue sections used first to
detect T-cell subclasses by immunohistochemistry. Based on the staining of serial
sections with antibodies against T-cell markers and in situ hybridizations (Fig. 12),
it appears most likely that the perforin and the granzyme A gene are expressed in
the CD8 positive subset during an LCMV infection of the liver and brain (Fig.12).
This conclusion is supported by the finding, that the CD4:CD8 ratio of the
infiltrating T-cells dropped during the LCMV-infection of the liver and the highest
frequency of perforin and granzyme A mRNA positive cells was found in both
organs at the time of maximal frequency of LCMV infected cells. We cannot
formally exclude however the possibility, that a minor fraction of CD4 positive
cells or even non-T-cells can express the perforin and/or granzyme A gene during
a viral infection in vivo.
The presented evidence for perforin and granzyme A positive cells in a T-cell
dependent infiltrate causing LCM differs from results obtained with an acute
bacterial meningitis caused by Listeria monocytogenes, when evaluated on day 2
after infection and shortly before death. The cells in the bacterial infiltrate
expressed the gene for TNFct, but only very few granzyme A and no perforin
mRNA containing cells could be detected. (K.Frei, in preparation). This difference
may therefore be most readily explained by the absence of T-cell involvement at
this early stage of Listeria meningitis.
In conclusion, in situ hybridizations with a radiolabeled RNA probe of the
granzyme A and the perforin gene clearly demonstrated the induction of cells with
transcripts of either gene in T-cell dependent inflammatory infiltrates of LCMV
51
infected mice. These positive cells were associated with virus infected cells in
infected liver and brain, suggesting that they are involved in antiviral cytotoxicity.
This study therefore provides
evidence for the induction of the perforin and granzyme A gene during the course
of an antiviral immune response in vivo, thus implying a role for these two proteins
in cell- mediated cytotoxicity in vivo.
Acknowledgment
We wish to thank Dr.I.L.Weissman for providing the HF cDNA clone and
T.P6rinat and G.Walle for expert technical assistance.
52
5.2 Detection of Perforin During Infection of Mice with
Lymphocytic Choriomeningitis Virus by RNA-PCR
Introduction
Transcription of the perforin gene during infection of mice with lymphocytic
choriomeningitis virus (LCMV) is readily detected by in situ hybridisation as
described in chapter 5. A strong upregulation of perforin transcription was found
during the course of the immune response with maximal levels of perforin mRNA
in infected organs on day 6 after infection. But detection of perforin mRNAwith in
situ hybridisation has several drawbacks: First, the method is very slow, especially
because the low level of perforin mRNA present in transcribing cells in vivo calls
for autoradiographic exposition times of up to 20 days. Secondly, quantification of
gene expression by in situ hybridisation is easily biased by the choice of the plane
of section and the region of observation. Because of this, a laborious analysis of a
statistical number of different sections and fields of observation is required.
In order to circumvent some of the problems of in situ hybridisation, the increase
of perforin mRNA expression during infection with LCMV was confirmed and
complemented by the method of RNA-PCR. The high sensitivity of the polymerase
chain reaction (PCR) renders this method suitable to analyze the expression of
perforin, a gene expressed at low levels in only a small portion of lymphocytes in
certain tissues in vivo. Compared to the histological approach of in situ
hybridisation, the method of RNA-PCR gene expression provides an integrated
analysis of expression over the complete volume of an organ. In addition, RNA-
PCR allows the characterization of mRNA transcribed from the wildtype or
alternatively from a mutated perforin allele by applying a set of primer pairs
amplifying different regions of the perforin cDNA molecule.
53
Material and Methods
Infection ofmice
Inbred C57BL/6 mice were infected with 200 pfu of LCMV-WE intravenously.
Mice were killed by cervical dislocation and spleens removed for RNA
preparation.
Isolation oftotal cellular RNA
RNA was extracted from spleens of mice (Chirgwin et al., 1979). Briefly, cell
suspensions were made by forcing the spleens through a steel mesh. 5x107 cells
were lysed in 6 M Guanidiniumisothicyanate, 0.1 M Tris.HCl pH 7.0, 0.1 M 8-
Mercaptoethanol, 0.5% sarcosyl. RNA was pelleted in 5.7 M CsClj, 0.1 M EDTA,
0.01 M Tris.HCl pH 7.0 by ultracentrifugation for 21 hours at 11500 g. Proteins
were removed by Phenol/Chloroform extraction and the RNA precipitated with
Ethanol. RNA was dissolved and stored in TE (10 mM Tris.HCl pH 7.5, 0.1 mM
EDTA).
First strand cDNA synthesis
100 pg total RNA was digested with 30 U of DNAse I in the presence of 30 U
RNAse inhibitor from human placenta in 10 mM Tris.HCl pH 7.5, 5 mM MgCI2
for 60 minutes at 37°C. After phenol/chloroform extraction, precipitation with
ethanol and resuspension in TE 2 /tg RNA were retrotranscribed (Gubler and
Hoffman, 1-983) with 15 U reverse transcriptase form Moloney Murine Leukemia
Virus in the presence of 30 U RNAse inhibitor, 0.2 /iM oligo (dT)^ and 0.5 mM
of dNTPs in 50 mM Tris.HCl pH 7.6, 8 mM MgCl2, 10 mM Dithiothreitol for 2
hours at 37°C. 30 U RNAse Tl was added and the RNA digested for 60 minutes at
37°C followed by heat inactivation for 5 minutes at 70°C.
PCR amplification ofcDNA
54
cDNA synthesized from 0.2 /fg of total RNA was amplified (Saiki et al., 1988) by
0.05 U of Super Taq Polymerase (Stehelin & Cie, Basel) in the presence of 0.4
lM of each primer, 200 iM dNTPs in 10 mM Tris.HCl pH 9.0, 500 mM KO, 0.1
% gelatine, 1.5 mM MgC^, 1 % Triton X-100 for 35 temperature cycles. Each
cycle was 1 minute at 94°C, 2 minutes at 60°C and 3 minutes at 72°C. Amplification
was completed with an 10 minutes incubation at 72°C.
PCR products were analysed on a 1.2% agarose gel by standard techniques.
Primers usedforPCR amplification
For designing and optimising the PCR primer pairs the Oligo Program, Version
4.0 (Wojcieck Rychlik, National Biosciences, USA) was used.
Perforin specific primer sequences were derived from the published cDNA
sequences for perforin (Lowrey et al., 1989; Shinkai et al., 1988) (Genebank
accession numbers J04148 and X12760) and for exon 1 from unpublished data
(E.Podack, personal communication).
Upper Primer a: 5' - TGG GCT TCA GTG GCG TCT TG- 3'
Lower Primer b: 5' - TGT GTG GCG TCT CTC ATT AG - 3'
Upper Primer c: 5' - AGG CAG CTG CTA ATA TCAAT - 3'
Lower Primer d: 5' - TGT GCT GTT TCT TCT TCT CC - 3'
Upper Primer e: 5' - AGC CCC TGC ACA CAT TAC TG - 3'
Lower Primer f: 5' - CCG GGG ATT GTT ATT GTT CC - 3'
Upper Primer g: 5' - AGA TGG ACT TTG AGA ATG TG - 3'
Lower Primer h: 5' - TTTTTGAGACCCTGTAGACCCA - 3'
The expected fragment lengths were for primer set a/b 472 bp and 366 bp
(differential splicing), for primer sets c/d, e/f and g/h 481 bp, 350 bp and 407 bp
respectively.
55
6-Actin specific primer sequences were derived from the published cDNA
sequence of cytoplasmatic B-actin (Alonso et al., 1986) (Genebank accession
number M12481).
Upper Primer i: 5' - CCA ACC GTG AAAAGA TGA CC - 3'
Lower Primer k: 5' - TCG TTG CCA ATA GTG ATG AC - 3'
A 418 bp DNA fragment was amplified by primer pair i/k from 6-actin cDNA.
Results
C57BL/6 mice were infected with 200 pfu of lymphocytic choriomeningitis virus of
strain WE i.v.. 8 days after infection total cellular RNA was extracted from the
spleens of infected and uninfected mice. Polyadenylated mRNA was
retrotranscribed to cDNA and was analyzed with three different primer pairs by
PCR as shown in Fig. 13. Despite extensive treatment of total cellular RNA with
DNAse during preparation, a minor contamination of the RNA preparation with
genomic DNA can never be excluded completely. In order to prevent false positive
bands due to residual genomic DNA in the RNA preparation, the primer pairs a/b
and c/d were designed to span a complete intron from one exon boundary to the
next exon boundary. In this configuration, amplification of contaminating genomic
DNA yields a fragment of considerably higher molecular weight than the fragment
amplified from cDNA. No contaminations with genomic DNA were detected in all
PCR reactions (Fig. 14).
A fourth set of primers (i/k) was used to amplify ubiquitous /J-actin sequence. The
comparable intensities of the bands in lane 1 and lane 2 confirm, that equal
amount of cDNA from the different samples were compared.
Amplification of a limiting amount of control plasmid sequence was used to assure
the sensitivity of the polymerase chain reaction (Fig. 14, probe 4). The sequence
contained in the control plasmid can be amplified with primer sets c/d and e/f, but
not with primer sets a/b and i/k. The prominent band amplified from 50 pg of
plasmid indicates the high sensitivity of the PCR reaction with primers c/d and e/f.
56
Despite this high sensitivity, perforin mRNA was not detectable in spleens of
uninfected mice (Fig.14, lane 2, primers a/b, c/d and e/f), while it was detected in
splenic RNA from infected mice (Fig. 14, lane 1, primer sets a/b, c/d and e/f). This
finding was verified with all three primer pairs specific for different regions of the
perforin gene.
Although the exponential amplification process during the polymerase chain
reaction impedes accurate quantification, the differences in band intensities
between infected and uninfected mice can be estimated to correspond to an at
least 10-fold increase of perforin expression during LCMV infection.
Besides the main PCR fragment, the primer set a/b amplified a second fragment
of slightly higher molecular weight. This second fragment is readily explained by
the reported differential splicing of perforin mRNA between exon 1 and exon 2,
resulting in a 106 bp difference between the two alternatively spliced mRNA
populations (Youn et al., 1991). This is in accordance with the result of RNA-PCR
presented here. However, the proportion of intensities between the two PCR
bands on lane 1 can not be taken as a representation of the stochiometric relation
between the two mRNA populations. Most likely the longer mRNA species is
present in a higher proportion than it is reflected by the band intensities, since the
amplification of a longer fragment competes unfavorably with the amplification of
a shorter fragment in the same PCR reaction.
57
a b
Starti
c d e fL 1
Stop
•i 2 31
Coding sequence
Fig.13. Locations ofprimers used for the analysis ofperforin transcripts. The PCR
primers, their relative locations and directions of amplification in the perforin gene
are indicated by arrows. Note that primer sets a/b and c/d are located at
neighboring ends of exons, spanning intron 1 or intron 2 respectively thus
amplifying different fragments from contaminating genomic DNA than from
cDNA.
58
Primers: a/b c/d e/f i/k
Fig. 14. Expression of perforin in noninfected and LCMV infected mice. RNA was
extracted from spleens of LCMV infected and uninfected mice, retrotranscribed,
amplified by PCR and separated on an agarose gel. 1: Mouse 8 days after
infection n with 200 pfu of LCMV-WE. 2: Uninfected mouse. 3: No cDNA added,
blank. 4: Positive control plasmid for primer sets c/d and e/f. Primer sets a/b, c/d
and e/f amplify perforin cDNA, primer pair i/k amplifies /J-actin cDNA. Relative
positions of primers a-f are shown in Fig. 13. The observed fragment lengths
corresponded to the expected lengths of 472 bp and 366 bp (differential splicing)
for primer set a/b and of 481 bp, 350 bp and 418 bp for primer sets c/d, e/f and
i/k respectively.
59
Discussion
It has been demonstrated that the overall expression of perforin mRNA in spleens
of mice is induced by infection with LCMV. Two independent methods were used
to confirm this finding. The observed upregulation of perforin expression was
consolidated by RNA-PCR in accordance with the results obtained with in situ
hybridisation. RNA-PCR circumvented some of the problems associated with the
histological restraints of expression analysis by in situ hybridisation. It has been
clearly confirmed that the overall amount of perforin mRNA in spleens of mice on
day 8 after LCMV infection increased by an estimated factor of at least 10
compared to uninfected mice. The result of the in situ hybridisation, the described
pattern of perforin expression in primary spleen cells and the low NK activity
observed during LCMV infection lead us to the conclusion that perforin expression
in the spleen occurs mainly during activation of resting virus specific T cells.
Subsequently, the mature cytolytic effector cells probably infiltrate the infected
tissues to eliminate virus infected cells by cytolysis. In addition, the observed
expression of perforin in the virus infected tissues of liver and brain argues for a
potential of mature activated cytotoxic T cells to replenish their cytolytic granula
by resynthesizing perforin in order to lyse several target cells sequentially.
It has been objected against the granule exocytosis model of perforin action that
perforin expression would be an in vitro artifact occurring in cells exposed over
longer periods of time to high amounts of IL-2. Such high amounts of IL-2 are
present in tissue culture media complemented with supernatant from concanavalin
A stimulated rat spleen cells (Berke and Rosen, 1988; Clark, 1988), as they are
widely used to cultivate T cells in vitro. The demonstration of significant perforin
expression during the course of an antiviral immune response in vivo directly
disproves these arguments.
A modification of this objection consists in the notion, that perforin expression
would be triggered by exposure of cytotoxic T cell precursors, mature cytotoxic T
cells and NK-cells to high local concentrations of IL-2 effective in a paracrine way
(Clark, 1988). Following this hypothesis, perforin expression would be the
60
hallmark of some kind of unspecific cytotoxic activity exerted by bystander
lymphocytes, similar to the unspecific activation of B-cells by certain bacterial cell
wall products. The detection of perforin expression during LCMV infection with
low levels of unspecific cytotoxic activity on uninfected fibroblasts and NK
sensitive YAC-1 cells compared to the prominent LCMV specific activity are
difficult to reconcile with this model. If indeed there had been an upregulation of
this stipulated unspecific, perforin dependent cytotoxicity it was not measurable by
our assays and in this case was much weaker than the specific part of the response.
Although this does not formally disprove the hypothesis, it renders it very
improbable.
It would be interesting to know, which lymphocyte subsets express perforin during
LCMV infection in vivo. In chapter 5, evidence for expression mainly in CD8
positive T cells was obtained by in situ hybridisation and immunohistology on
serial sections. To directly identify the perforin expressing lymphocyte subsets an
alternative method is suggested here. It should be feasible to sort spleen cells by a
fluorescence activated cell sorter (FACS) into different lymphocyte subsets
characterized by the presence of markers like CD8, CD4 and asialo GM1 and test
for perforin expression in the sorted cell populations by RNA-PCR. The technical
problems associated with this procedure would probably be more easily
surmountable than with the double staining procedure suggested above. Due to its
high sensitivity, this approach appears also more promising than earlier
experiments testing the cytolytic activity of granule preparation from different
lymphocyte subsets on sheep red blood cells (Berke and Rosen, 1987; Garcia-Sanz
et al., 1987).
The demonstration of perforin expression during infection with a virus mainly
controlled by cytotoxic T cells strongly suggests an involvement of perforin in the
mechanism of cell lysis by cytotoxic T cells and natural killer cells in vivo. However
the possibility, that perforin expression in these physiological situations is rather an
epiphenomenon than the effective principle at work, can not be ruled out
completely. In order to gain more information about the physiological role of
61
perforin, the ability to directly interfere with the activity of perforin in vivo would
allow a more direct assessment of the role of perforin in cytolysis by lymphocytes.
62
6. Generation of Mice with a Disruption of the Perforin Gene by
Homologous Recombination in Embryonic Stem Ceils
Introduction
Cytotoxic T-cells and Natural Killer cells are both capable of rapidly lysing target
cells and are involved in lysis of virus-infected cells, allograft rejection and tumor
surveillance. The killing process requires a one to one cell to cell contact (Marker
and Volkert, 1973; Martz, 1976; Berke, 1983) conferred by the specific T-cell
receptor in the case of cytotoxic T-cells and a yet hypothetical receptor for NK
cells. In addition, several accessory adhesion molecules participate in the
interaction between cytotoxic and target cell to increase the affinity of the
interaction and/or to provide additional stimulatory signal to the lymphocyte.
Eversince cytolytic activity of lymphocytes has first been described (Cerrotini et al.,
1970; Herberman et al., 1975), the mechanism of lysis has been the subject of great
interest. Early attempts to generate inhibiting antibodies against surface
determinants expressed only on conjugates between cytotoxic T-cells and target
cells and not on cytotoxic cells per se did not reveal the presumed cytolytic effector
molecules but rather accessory binding proteins (Pierres et al., 1982). The cloning
of cytolytic lymphocytes and the identification of the content of their cytolytic
granules have brought some although still limited understanding of the process.
Several candidate effector molecules, partly isolated from these granules, and
killing mechanisms have been suggested, among them perforin (also referred to as
cytolysin, pore forming protein, PFP and PI), lymphotoxin and other TNF
homologs (Granger and Kolb, 1968; Ruddle and Waksman, 1968; Liu et al., 1987),
granzymes (Pasternack and Eisen, 1985; Masson and Tschopp, 1987; Masson et al.,
1986; Simon et al., 1986), secreted ATP (Di Virgilio et al., 1990) and the apoptosis
model of cytotoxicity.
The best characterized of these candidate mechanism is the granula exocytosis
model (Henkart, 1985; Podack, 1985; Lachmann, 1983), which proposes, that
63
cytotoxic T-cells upon recognition of the target cell release the contents of their
cytolytic granules. These granules contain the effector molecule perforin. The pore
formed in the target cell membrane by the polymerizing perforin molecules
renders the membrane permeable and leads to osmotic lysis of the target cell.
Indeed electronmicroscopical studies have demonstrated ringlike pore structures
on the membranes of lysed cells deposited by cytotoxic T-cells and NK-cells :
These structures resemble those found during complement mediated lysis
(Dourmashkin et al., 1980; Dennert and Podack, 1983). Subsequently, the cytolytic
activity was shown to reside in the granules of cytolytic lymphocytes and from these
granules the 70 kD protein perforin was isolated. The membranolytic perforin
molecule is capable of polymerizing into pore like structures as they are observed
on the surface of cells lysed by cytolytic lymphocytes or granule preparations. The
functional analogy between complement C9 and perforin appears in the primary
structures of these poreforming proteins as well, since the central third of the
amino acid sequence for perforin (Shinkai et al., 1988; Young et al., 1986a)
displays homology to C9. The carboxy- and aminoterminus,
where a
membranolytic peptide was localized (Ojcius et al., 1991), are specific to perforin.
The perforin genomic locus is contained in three exons spread over a short stretch
of 10 kb of the mouse genome. Since translation starts in exon 2 and ends in exon
3, exon 1 remains untranslated. The region homologous to complement component
C9 and thought to be functionally essential is distributed over the 3'-half of exon 2
and the 5'-half of exon 3.
Perforin expression was demonstrated in tissue cultured CD8 positive cytotoxic T-
cells (Shinkai et al., 1988; Podack et al., 1988; Garcia-Sanz et al., 1987) and in NK
cells (Kawasaki et al., 1992), but generally not in CD4 positive T-cells (Garcia-
Sanz et al., 1987) although recently evidence for expression of perforin in CD4
positive T-cells isolated from patients with rheumatoid arthritis (Griffiths et al.,
1992) and other autoimmune diseases has been reported. Intracellularly, perforin
was localized to the granula of cytotoxic lymphocytes by
immunoelectronmicroscopy (Groscurth et al., 1987; Peters et al., 1989). Perforin
64
expression in vivo is associated with the kinetics and histological localization of
lymphocytic choriomeningitis virus specific, murine cytotoxic T-cells (Miiller et al.,
1989), with rejection of allogeneic transplants in mouse and man (Mueller et al.,
1988; Young et al., 1990), with spontaneous diabetes in NOD mice (Young et al.,
1989) and with cytolytic lymphocytes isolated from the synovial fluids of patients
with rheumatoid arthritis (Griffiths et al., 1992). In addition, the expression of
perforin was recently demonstrated in the reproductive tract of female mice, not
only in infiltrating cells of the regressing corpus luteum but also in granular metrial
gland cells of the placenta (Zheng et al., 1991). Because of the relation of granular
metrial cells to NK cells, it was suggested that they prevent embryo derived cells to
pass to the mother or alternatively, that they binder virus infected maternal cells to
pass to the embryo.
Evidence for an essential role of perforin at least for in vitro generated cytotoxicity
was provided by the finding, that perforin expression and CD3 specific mediated
cytotoxicity of primary mouse spleen cells could be significantly inhibited with
perforin antisense oligonucleotides. (Acha-Orbea et al., 1990).
Despite this considerable evidence, the role of perforin in cell mediated immunity
has remained controversial. In order to shed some light on the role of perforin in
vivo, we applied the method of homologous recombination in ES cells (Evans and
Kaufman, 1981; Bradley et al., 1984; Thomas and Capecchi, 1987; Hooper et al.,
1987) to generate perforin deficient mice. Such mice would provide an in vivo
system to clarify the question whether perforin is indispensable or only
redundantly together with one or several additional mechanisms involved in
cytolysis by lymphocytes.
65
Material and Methods
DNA sequencing
Perforin intron 2 was sequenced from the double stranded plasmid pTZ19R.GC
with a perforin exoni primer.( 5' - CCT ATC AGG ACC AGT ACA ACT - 3') by
Taq cycle sequencing (Taq Dye Deoxy Terminator Cycle Sequencing Kit,
Applied Biosystems Inc., Foster City, CA). The reaction products were separated
on one lane of a 6% acrylamide gel and analysed by an Applied Biosystems 373A
DNA sequencer.
Synthesis ofDNA oligonucleotides
For designing and optimising the PCR primer pairs the Oligo Program, Version
4.0 (Wojcieck Rychlik, National Biosciences, USA) was used.
Perforin specific primer sequences were derived from the published cDNA
sequences of perforin (Lowrey et al., 1989; Shinkai et al., 1988) (Genebank
accession numbers J04148 and X12760) and for exon 1 from unpublished data
(E.Podack, personal communication).
Primer a (Fig.20)
b (Fig.20)
c(Fig.l8)
d (Fig.18)
e(Fig.l8)
f (Fig.18)
g:
h:
5'- TTT TTG AGA CCC TCTAGA CCC A -3'
5'- GCA TCG CCT TCT ATC GCC TTC T -3'
5'- GTA GAG TTG AGG CTG ATG CC -3'
5'- ATT CGC CAATGA CAA GAC GCT G -3'
5'- CTG ACC GCT TCC TCG TGC TTT A -3'
5'- GAG AGA TCA GGA TTT GGT GT -3'
5'- GCC ACT CGG TCA GAA TGC AAG C -3'
5'- AGG GTC ACA GCA TTA GGA AG -3'
5'- CCG GTC CTG AAC TCC TGG CCA C -3'
5'- CCC CTG CAC ACA TTA CTG GAA G -3'
Primers were synthesized by the phosphoramidite method on a solid support
(Matteucci and Caruthers, 1981). The synthesis was performed automatically on an
66
Applied Biosystems Model 391 DNA synthesizer. After removal of the protecting
groups, the neutralised primer solutions were precipitated with ethanol and
purified by gelfiltration on a Sephadex G50 column (Pharmacia, Uppsala,
Sweden). The oligonucleotides were quantitated by optical density at 260 nm
wavelength.
Plasmids
The plasmid pMCI.neoPA containing the neo gene with a point-mutation
presumably decreasing its expression (Yenofsky et al., 1990) was a gift from Silvio
Hemmi, Institut fur Molekularbiologie I, Universitat Zurich.
The plasmid pICI19R.MCI-TK containing the herpes simplex virus thymidine
kinase gene under the control of a promoter optimised for expression in embryonic
stem cells was obtained from Kurt Biirki, Sandoz AG, Basel.
Hybridisation probesforgenomic Southern blotting
The locations of the probes is indicated in Fig.18 and Fig.20.
Probe A was a 620 bp PCR fragment amplified from perforin exon 2 with primer
set g/h (see above). Probe B was a 600 bp restriction fragment prepared from
plasmid pMCLneoPA containing 3'-coding sequence for the neomycin-
phosphotransferase. Probe C was a 298 bp PCR fragment amplified from perforin
exon 3 with primer set i/k.
Mice
Mice were obtained from BRL (Fullinsdorf, Switzerland).
Preparation ofembryonalfibroblast layers
To prepare G418 resistant feeder cells, CDl-MTKneo2 foeti transgenic for the
neomycin-resistance gene were used (Stewart et al., 1987). About 30 13-14 days old
foeti with heads and livers removed were washed with CMF-PBS, homogenized by
pressing through a 18-gauge injection needle and trypsinized for one hour at 37°C
67
in 10 ml Trypsin/EDTA solution (Gibco/BRL, Paisley, Scotland). The cells in the
supernatant were pelleted, taken up in feeder cell medium (DMEM
complemented with 0.1 mM fl-mercaptoethanol and 10% FCS and transferred to a
25 ml cell culture flask. Trypsinization of the carcasses was repeated four times for
30 min as above. Each time, the fibroblasts pelleted from the supernatant were
transferred to a new flask. The embryonal fibroblasts were grown to confluent
layers, passaged once and frozen. Thawed fibroblasts were expanded for not more
than seven passages.
For the production of feeder layers, the fibroblasts were either treated for 150 min.
with a 10 /ig/ml mitomycin C solution or alternatively irradiated with a dose of
3000 rads from a 60Co source and seeded at 5 x 104 cells/cm2 onto gelatinized
cell culture dishes.
In vitro culture ofES cells
The male D3 ES cell line was originally derived from an embryo of the 129/Sv
strain (Gossler et al., 1986) and was provided by Dr. Marco Schilham (Toronto,
Canada). D3H and D3M are sublines from the same line and obtained from Dr.
Michel Aguet (Zurich, Switzerland) and Dr. Manfred Kopf (Freiburg i.Br.,
Germany) respectively. The male ES cell line BL/6-IQ was established from an
embryo of the C57BL/6 mouse strain (Ledermann and Burki, 1991). Its
pluripotency was confirmed by the generation of germline chimeras from
blastocysts injected with BL/6-HI cells.
ES cells were passaged every second day by trypsinization with a Trypsin/EDTA
solution (Gibco/BRL, Paisley, Scotland). Medium (DMEM complemented with 15
% FCS, 0.1 mM B-mercaptoethanol and 1000 U/ml human Leukaemia Inhibiting
Factor (LIF, Sandoz AG, Basel, Switzerland)) was changed daily to prevent
differentiation. ES cells were grown routinely on growth arrested G418 resistant
embryonic feeder cells in an incubator gassed with 10% CQ2 at 37°C.
68
Transection and selection ofES cells
The ES cells were split the day before electroporation. For electroporations, 5 x
106 ES cells were taken up in 800 /d ES cell medium without purification from
feeder cells, electroporated in the presence of 15 /fg of linearized plasmid DNA
with a Bio-Rad Gene Pulser equipped with a capacitance extender set at 500 /fF
and 240 V. The cells were then immediately plated on two fibroblast coated 10 cm
dishes and were allowed to recover for two days. G418 (Geneticin, Gibco/BRL,
Paisley, Scotland) selection was performed for 10 days at 400 /fg/ml. Some plates
were additionally selected from day 7 of the G418 selection with 2 /fM gancyclovir
(Cymevene, Syntex Pharm AG, Allschwil, Switzerland).
Screening ofES coloniesfor homologous recombination
When colonies were grown to about 50 cells on day 10 or 11 of G418 selection, the
medium was replaced with PBS and single colonies were removed by scraping with
a 10 /fl microtip. After trypsinisation in a drop of Trypsin/EDTA solution half of
each colony was transferred onto a fibroblast layer in a 96 well plate and the other
half was transferred to an Eppendorf tube. While one half of the colony grew to a
confluent clone in the 96 well plate, the other half was screened for homologous
recombination by PCR amplification. In detail, the cells were washed with PBS,
hypotonically lysed with 30 /fl of water, boiled for 10 min. at 95°C and digested for
90 min. at 55°C with 1/fg of Proteinase K (Sigma Chemicals, St. Louis, USA). The
proteinase was subsequently inactivated by boiling for 10 min at 95°C. 20 /fl of this
crude extract was then amplified in a 50 /fl PCR reaction (Saiki et al., 1988) with
0.05 U of Super Taq polymerase (Stehelin & Cie, Basel) in the presence of 0.4 /fM
primers each, 200 /fM dNTPs in 10 mM Tris.HCl pH 9.0, 500 mM KCl, 0.1 %
gelatin, 1.5 mM MgCl2, 1 % Triton X-100 for 35 temperature cycles. Each cycle
was 1 minute at 94°C, 2 minutes at 60°C and 3 minutes at 72°C. Amplification was
completed with 10 minutes incubation at 72°C. For cells electroporated with
construct pICI.PHR the primer sets c/d or e/f were used, for constructc pICI.HKl
69
and pICI.HK2 the primer set a/b detected homologous recombination and for
constructs pICI.HK3 and pICI.HK4 primer set b/d was used,
the PCR products were separated on a 1.2 % agarose gel, which was blotted with a
vacuum blotter (Pharmacia, Uppsala, Sweden) to a Genescreen positively charged
nylon membrane (NEN, Boston, MA)..The membrane was prehybridized for at
least 5 hours in a rotating glass cylinder at 65°C with 10 ml of 5x SSC, lOx
Denhardt's solution, 10% dextranesulfat, 0.1% SDS an 100 Mg/ml denatured
salmon sperm DNA. A DNA fragment identical to the expected PCR product was
amplified before from a construction intermediate and used as a probe. The
respective PCR fragment was radioactively labelled with 32P by random priming
(Oligo labelling kit, Pharmacia, Uppsala, Sweden) and separated from
unincorporated nucleotides by gel filtration over a Sephadex G50 column. 107 cpm
of the labelled probe were added to the prehybridisation mix and allowed to
hybridise for at least 10 hrs at 65°C. The filter was then washed stringently with
0.2x SSC, 0,1% SDS at 65°C, exposed to a phosphorimager screen for 4 hrs and
analysed with the Digiscan system (Siemens AG, Munich, Germany). PCR positive
clones were grown up, frozen and analysed by genomic Southern blotting. Ten
micrograms of genomic DNA were digested with 50U of EcoRI, separated on a
0.8% agarose gel, blotted and hybridised as described above. The filters were first
hybridized with a probe flanking to the construct, stripped and rehybridized with a
probe internal to the targeting construct, stripped again and finally hybridized
with a third probe derived from the neoresistance gene. Only when PCR and
Southern hybridisation with all three probes showed the predicted result, a clone
was regarded as homologously recombined.
Generation ofgermline chimerasfrom targeted ES cell clones
All blastocyst injections were carried out by Drs. B. Ledermann and K. Biirki at
Sandoz AG, Basel.
Donor females of strain BALB/c were induced to ovulate by an intraperitoneal
injection of 5 I.U. pregnant mare serum (PMS, Folligon, Intervet), followed 46 hr
70
later by an intraperitoneal injection of 5 I.U. human choriogonadotropin
(Chorulon, Intervet). Morulae were flushed from the uterotubal junction 3 days
after mating and cultured overnight in standard egg culture medium (Biggers et al.,
1971). The ES cells were purified from embryonal fibroblasts by incubation for 1
hr in a non gelatinized tissue culture dish. Blastocysts were transferred together
with suspended ES cells into drops of injection medium (DMEM, 10 mM Hepes
pH 7.4,10 mg/ml BSA, 300 U/ml DNAse I (Sigma)). About 10 to 20 ES cells were
injected into each blastocyst (Hogan et al., 1986). After a recovery period of about
2 hours, injected blastocysts were transferred into both uterine horns of
pseudopregnant CD-I foster mothers. Chimeric offsprings were identified by eye
and coat pigmentation. Chimeric males were set up to breed with C57BL/6
females at the age of 6 weeks and the black coat color of the offsprings indicated
germ line chimerism.
Results
Sequencing oftheperforin genomic locus
In order to design the targeting constructs, the cloning strategies and the primers
to screen for homologous recombination it was necessary to compile the available
information about the sequence of the genomic perforin locus and to complement
it with sequence from intron 2. Intron 2 has not been sequenced so far, however
sequence information going beyond the end of the targeting vector pICLPHR
(Fig.15) was required to design an improved primer pair for screening
electroporated ES cells. The sequences have been compiled from [Eckhard
Podack, ©personal communication, 2051, 5266, 6463]. There are considerable
discrepancies in the published perforin sequences, especially in the first two exons,
but the sequence presented here (Tab.2) is in agreement with all the PCR
reactions and Southern blotting analyses done during this work and can be
regarded as a bona fide consensus sequence.
/+
f+
1+
f+ ,
+
1 GGGTGAGATGTGACCATGTGGCCTGGGGTCTGTGGCTACTTATTCCCATC 50
51 ATGTACrCACTTAGGGGTGTTGAGAACTACCCCCACCCTGCATGGTTTTA 100
101 CCCACCTCATAACCAGCACAATTTTTTTTTTTAAGCCCCCAGTAGGTATC 150
151 CATGAGTCACrCTGCCCATGGAAACCTCCCACAGTAACCTCAGGCAGAAC 200
201 AGAGTGGCCAACTGTCTrTCCACAGTACTAGCCTGCTCAAACCTGAGGTT 250
,+
,+
,+
,+
,+•
251 CAGCTGGGATGTAGGTATGGACAGGAAGTGGGTACCAGCTTTGAGTATCT 300
301 CTCCCCACTCCCAGGGAGGAGGGAACAGTCACATGTATTGAGAATGAGTA 350
351 AAGATCGTGTCTAGTCCrAAGCACTGCACCATGTCTTCATGTCCCCTCTC 400
401 ATGGGACACTGTGATTAGCCACACTTTTGAATGAGGAGC3X3AGATGTGAA 450
451 GCGGCTGAAACCCCn^^CC^GCC^CTCAAGCCCAGGCCACTCPrCAGCA 500
501 ACTCTTCCTGACAGCGCATCCCATCCCAAACACTCAACTCAGAAGCAGGG 550
551 AGCAGTCAGTTGGCCTGCTGGTCCACACCAGAGTCCCGCACCCCAACTGC 600
601 TGCTCTGCTTGGCAGATGAGCCCCTGGCCirK^^ 650
651 GCTGTCACAAGCCGGATGAGGAGGTGACAACAGGGTGGGTGCTGGTGGGA 700
701 ACCTGTGACCACACTCTGGGGGCAGGGCAGGAAGTAGTAATGATATGACG 750
,+
,+
,+
,+
,+
Exon 1
751 TTGGCCaGGGTGGGCCTGCCTGGGGAIAGATCGCAGCATTTTAAAAGCCT 800
801 CC^TTGACAACGCAGGTGTCCCCCTGGGCTTC^GTGGCGTCTTGGTGGGA 850
851 CTTCAGGTAAGGAGGAGGAAGGGTAGGAAGCAGGGAGGTTGAGATACCCT 900
901 GGGGAAAGAGAGTTGGAAGGTATAGCCAGGAGGAACCAGAAATGCTAGTA 950
951 GCCAGCCTTCAAGCTCATCCCTCATTCCTTCATTGGGTTGGTGACAACTC 1000
,+
,+
,+
,----+
,+
1001 GGGGGTTCATGTGTCAGTGACAACTGTCCACAGTTTCCAAGACTATTTCC 1050
1051 AGAAGTTTTGGCAGTTAGAGTTCGTAAAGAATGGAAATGAGCCAAGTATG 1100
1101 GAGCACATATATGACTATACCATCAGGGAACGGGGTAAACCCCATCTCGA 1150
1151 AAGGTGXAAGGAAGGGGAGGTGAGGGGAGAGGACAGGAGAGGAAGGAAAG 1200
1201 GGAGGGGAGGGAAAAGGAGGAAGAATCCAAGATCTGTAAATAAACTCCCT 1250
1251 GACAATATAATGAAGGCCAGAAGGGGACTGATGTCCATTGATCAAAGTCA 1300
1301 CCAATGTGTTGGTCTCTTAAGCAGGTTGACCAGAAAGCrTGACAGGATTC 1350
EcoRI
1351 TCCCCAGCTTGTTTATCTCCCAGCCTCACACATCACCTGAATTCTGCCTC 1400
1401 CAGCTCACTGCriTAGGTAGAACATGGAAGGCrGCAGCTAGATATGAGGA 1450
1451 AGACCTCraZKACCCCACTCCCTCACTTAAGAC^ 1500
,+
,+
,+
,+
,+.
1501 AAAAACAAACAAAOUACAAACAAACACCTCTGAGGGTCTTAGAAGGAGC 1550
1551 CAGTTCTCCATCCGTGGGTTGTGTCCACTGGGGGAAAAAAACCCACCTAT 1600
1601 TTCCAATGGATGGTCTTAGAAAACGCAAGATAGCAAAATTACAGTCATGA 1650
1651 AGTCATGAAAATAAATTTATGGTTGGGGGTTCCa^GGACCATCAGAAATG 1700
1701 TGTGTTTTCCTATGGCTGCAATCCAC^GGTGGAGAAACGCTGAGACAGAG 1750
,H
,H
,H
,H
,+
1751 AGGAGAa^GAGGGAGATTCAAAAACTGGGACCCTGTGGCAAGGAAGATCA 1800
1801 ATGAATTGCTGCCCACIAGCAGGCGAaU3GTAGAaaCCTGTCCTGCCATA 1850
1851 GTGTGAAGTATTCGGCAGGGAATGTCAAGACAGAAAACCTGAGCAGTGAG 1900
1901 C^AAAGGGCTTGCAAGGGCGCACTTAAAGGTCCCCCACACTAGAGCCAAC 1950
1951 AGCACAGATACAGGACAGCCTTTGTCCATTGTAGAGTTGAGGCTGATGCC 2000
,+
,+
,+
,+
..+
Sac I
2001 CTTGCAGGAGGCP^GAGCAGCCACGGGGGATATGCAGGGAGCTCACTGT 2050
2051 CTCCTAGCATCTGAAAGGGAGAGGCAGAAGTCACTGCTGGAATTAGAGAT 2100
2101 GGAGATTAAGTAAGAATCAGCTCGGAGCTATGCCTGGTGCCACAGGATTG 2150-
2151 AGATCGTAGCTCCTCCAC^TGTTCAAGGCrTGGCAACAGAATCACTTCAA 2200
2201 AGCTATCCTGAGCAATGATACCTTGTCTCAAGATAGATAGACAGACAGAC 2250
2251 AGACAGACAGATAGATAGATAGACAGACAGACAGACAGATAGATAGATGA 2300
2301 TAGATAGACAGATAGATGATAGATAGATAGGTAGAXAGACAGATAGATGA 2350
2351 TAGATAGACAGACAGACAGATAGATAGATAGATAGATAGATAGACAGACA 2400
2401 GATAGATGATAGATAGACAGATAGATGATAGATAGGTAGATAGATAGATA 2450
2451 GATAGATAGATAGATAGATAGATAGATGATTGATAGATAGATGATAGGTA 2500
/+
, +,
+ ,+
,+
2501 GATGAGGAGTAGGGTGTGGCTCAGTGGTGGAACACTTGCCTACAGTGCGT 2550
2551 AAGACTCTGGGGGATGAAGGATATATCCCTTCATCCCAGAACCGGGGAGG 2600
2601 GGAATCAGGTAGGAAAAGTGTCCCACTCTGGGTGGAAAATGACACATCTC 2650
2651 CAGAGCCGAAACCTATTACAGATGCTGTGTCCAGGGTCACAGCCrrCAAT 2700
2701 CTCGATTCCTGCTAGGTCTGAGATAATAAAATCTCAGCACAGGGTTCCTC 2750
,+
,+
,+
1+
,+
2751 TTGGCTGTGCGACTCTGAGCTATATGTTTAAATCCCrCTGGATTTCTCTG 2800
Exon 2
2801 TACAGTGGACGAAGCCAGTCCAGCTCTGGCATGAATAACTAAAGAGAAGT 2850
2851 TCACTCTCTCCTGATGTTCCCCAGTCGTGAGAGGTCAGCATCCTTCATCC 2900
Start
2901 CTGTTCCCACAGCTTTCCAGAGTTTATGACTACTGTGCCTGCAGCATCAT 2950
2951 GGCCACGTGCCTGTTCCTCCTGGGCC'riTrCCTGCTGCTGCCACGACCTG 3000
,+
,+
,+
,+
,+
3001 TCCCTGCTCCCTGCTACACTGCCACTCGGTCAGAATGCAAGCAGAAGCAC 3050
3051 AAGTTCGTGCCAGGTGTATGGATGGCTGGGGAAGGCATGGATGTGACTAC 3100
3101 CCn^CGCCGCTCCGGCTCCTTCCCAGTGAACAC^CaGAGGTTCCTGAGGC 3150
3151 CTGACCGCACCTGCACCCK^X^AAAAACTCCCTAATGAGAGACGCCACA 3200
3201 CAGCGCCTACCTGTGGCAATCACCCACTGGCGGCCTCACAGCTCACACTG 3250
,+
,+
,+
,+
,+
3251 CCAGCGTAATGTGGCCGCAGCCAAGGTCCACTCCACGGAGGGTGTGGCCC 3300
3301 GGGAGGCAGCTGCTAATATCAATAACGACTGGCGTGTGGGGCTGGATGTG 3350
3351 AACCCTAGGCCAGAGGCAAACATGCGCGCCTCCGTGGCTGGCTCCCACTC 3400
3401 OVAGGTAGCCAATTTTGCAGCTGAGAAGACCTATCAGGACCAGTACAACT 3450
Exon 2
3451 TTAATAGCGACACAGTAGAGTGTCGCATGTACAGGTAAGAGGAGAGAGGC 3500
3501 ATGGGAGAAAGGAACTGGAAGAATAGTGGGACAATCTAAGGGGTTTAGGA 3550
Sac I
3551 GCTCTGGAATAGTGGCGTGAAGTTTACCTACTGTTTCTCAGGAAGACACC 3600
3601 AAATCCTGATCTCTCTAGAGTCAGTAGTTCTCAGCCTTCCTAATGCTGTG 3650
3651 ACCCTCTAATACAGTTCCTAATGTGGTGACTCCAACCATAAAATGATTGT 3700
3701 GTTGCTACTTTAGAATCTTAATTGTGCTACTGATATGAACCATAATGTAA 3750
74
/+
,+
,+
,+
,+
3751 ATATCTGATATCCAGGCCATCTGACATGCGATCCCTGTGGAGGTCTTGAC 3800
3801 CCACAGGTTAAGAATCTCTGCTCTAGCCGGGCGTNNNNNNNNN3NNNNNNN 3850
< INTRON 2 (2-3 kb) >
Exon 3
5751 NNNNNCTCCTCTAATGCTCTGTCCTCTCCACAGTTTTCGCCTGGTACAAA 5800
5801 AACCTCCACTCCACCTTGACTTCAAAAAGGCGCTCAGAGCCCTCCCCCGC 5850
5851 AACrrrAAC^GCTCCACAGAGCATGCTTACCACAGGCTCATCTCCTCCTA 5900
5901 TGGCACGCACTTTATCACGGCTGTGGACCTCGGTGGCCGCATCTCGGTCC 5950
5951 TTACaGCCCTGCGTACCTGTCAGCTGACCCTGAATGGGCTCACAGCTGAT 6000
,+
,+
,+
,+
,+
6001 GAGGTAGGAGACTGCCTGAACGTGGAGGCCCAGGTCAGCATCGGTGCCCA 6050
6051 AGCCAGCGTCTCCAGTGAATACAAAGCTTGTGAGGAGAAGAAGAAACAGC 6100
6101 ACAAAATGGCCaCCTCTTTCCACCAGACCTACCGTGAGCGTCACGTCGAA 6150
6151 GTACTTGGTGGCCCTCTGGACTCCACGCATGATCTGCTCTTCGGGAACCA 6200
6201 AGCTACACCAGAGCAGrrCTCAACCTGGACAGCCTCACTGCCCAGCAACC 6250
,+
,+
,+
,+
,+
6251 CTGGTCTGGTGGACTACAGCCTGGAGCCCCTGCACACATTACTGGAAGAA 6300
6301 CAGAACCCGAAGCGGGAGGCTCTGAGACAGGCTATCAGCCATTATATAAT 6350
6351 GAGCAGAGCCCGGTGGCAGAACTGTAGCAGGCCCTGCAGGTCAGGCCAGC 6400
6401 ATAAGAGTAGCCATGATTCATGCCAGTGTGAGTGCCAGGATTCAAAGGTC 6450
6451 ACCAACCAGGACTGCTGCCCACGACAGAGGGGCTTGGCCCATTTGGTGGT 6500
6501 AAGCAATTTCCGGGCAGAACATCTGTGGGGAGACTACACCACAGCTACTG 6550
6551 AAGCCTACCTAAAGGTCTTCTTTGGTGGCCAGGAGTTCAGGACCGGTGTC 6600
6601 GTGTGGAACAATAACAATCCCCGGTGGACTGACAAGATGGACTTTGAGAA 6650
6651 TGTGCTCCTGTCCACAGGGGGACCCCTCAGGGTGCAGGTCTGGGATGCCG 6700
6701 ACTACGGCTGGGATGATGACCTTCTTGGTTCTTGTGACAGGTCTCCCCAC 6750
6751 TCTGGTTTCCATGAGGTGACATGTGAGCrAAACCACGGCAGGGTGAAATT 6800
6801 CTCCTACCATGCCAAGTGTCTGCCCCATCTCACTGGAGGGACCTGCCTGG 6850
6851 AGTATGCCCCCCAGGGGCTTCTGGAGATCCTCCAGGAAACCGCATGTGGG 6900
Stop6901 GCTGTGTGGTAACATAATAACAACAATAACATGCCTGAGAGCTGGGTGTA 6950
6951 GTAGCACACGCCTTTAATCCCAGCATTTGGGAGGCAGAGACAGGTGGATA 7000
Exon 3
7001 TCTATGAGTTC3AGGCCAGCCTGGGTCTACAGGGTCTC 7038
75
Tab.2. (Previous pages) Incomplete genomic nucleotide sequence of the murine
perforin locus. The sequence has been compiled from [Eckhard Podack, personal
communication, 2051, 5266, 6463].. The partial sequence data from intron 2 has
not been published before. Exon 1, exon 2 and exon 3 are underlined. Start and
stop indicate the first and last translated codons.
Construction ofthe targeting vectors
Homologous recombination of the replacement type occurs when vectors with
homologous sequence of at least 1 kb are introduced into embryonic stem cells.
Exploiting the fact that interruption of the homologous sequence with small
stretches of non-homologous sequence does not preclude homologous
recombination a neomycin resistance gene can be introduced into a coding exon.
The resistance gene not only serves as a positive selection marker but also contains
multiple translation stop codons in all three reading frames leading to inactivation
of the targeted allele.
Five different targeting constructs suitable for homologous recombination at the
perforin locus were produced. For all constructs, the vector pMCLneoPA (Lowrey
et al., 1989; Trapani et al., 1989; Youn et al., 1991) provided the neoR cassette. A
point mutation of the neomycin resistance gene occurring in some of the
circulating plasmids pMCLneoPA has been reported to decrease the G418
resistance conferred by the neomycine resistance cassette (Thomas and Capecchi,
1987). Retrospectively, the mutation in the coding sequence of the
phosphotransferase was detected by a restriction digest in the neo cassette of all
constructs.
Homologous recombination with the first construct pICLPHR would result in
disruption of the perforin gene in exon 2. To construct pICLPHR (Fig. 15 and
Fig. 16), a 1.2 Sall/Xhol fragment from pMCLneoPA was cloned into the vector
pTZ19R.GC, that contains a genomic DNA fragment of perforin (kindly provided
by E.Podack, Miami ). This construction intermediate pTZ19R.mPlneoPA was
linearized with Smal in exon 2 and the Smal ends were converted to Xhol
76
compatible ends by ligation with an Xhol linker and digestion with Xhol. The
mutated perforin SacI/SacI fragment was subsequently subcloned into
pICI19R.MCI-TK (Yenofsky et al., 1990) containing the HSV-TK gene under the
control of the HSV-TK promoter engineered for optimal expression in ES cells.
The electroporation construct pICLPHR was linearized with Seal before use in
electroporation experiments.
Since no homologous recombinant ES cells were recovered from electroporation
with linearized plasmid pICLPHR, the four related constructs pICLHKl,
pICI.HK2, pICI.HK3 and pICLHK4 (Fig.17) were established subsequently. These
constructs lead to disruption of the perforin gene in exon 3.1n order to clone the
targeting constructs pICLHKl -HK4 (Fig.15) the vector pTZ19R.GC containing a
genomic fragment from the perforin gene was opened at the BstE2 restriction site
in exon3. By a ligation of blunt ends an Sall/Xhol fragment containing the neo
cassette from pMCLneoPA was inserted into exon3. Parallel and antiparallel
orientation of the neo cassette gave rise to the two construction intermediates
pTZ19R.VKp and pTZ19R.VKap. The EcoRV/EcoRV fragments from the
intermediates were then subcloned each into pICI19R.MCI-TK. Different
orientations of the EcoR5 fragments to the HSV-TK gene of the vector yielded the
four electroporation constructs pICLHKl, pICI.HK2, pICI.HK3 and pICI.HK4.
77
.0.9 0.2 1.7 1.2 0.2 1.0 0.25 1.8 kb
CO
<8
3 9a
3 S.S
0.XI-
HSV-TK
oCO
CO
Perforin
genomic ONA
1512S^
8a>
1 Perforin
£ genomic ONA
<
1
COoCO
Fig. 15. Structure of the exon 2 derived construct pICLPHR The construct used to
electroporate ES cells is contained in the plasmid pICLPHR. The perforin
homologous sequence is a SacI/SacI fragment from intron 1, exon 2 and intron 2
of the genomic perforin locus. The perforin introns are indicated by a thicker line.
The interruption with the neo cassette at the Bgll site of exon 3 introduces several
stop codons in all three reading frames. The structural neo gene from Tn5
(Mansour et al., 1988) is driven by a promoter optimized for expression in ES cells
(beck et al., 1982). In detail, it consists of a tandem repeat of the enhancer region
from the polyoma mutant PYF441, bases 5210 - 5274 (Thomas and Capecchi,
1987), the HSV-TK promoter from bases 92 - 218 (Fujimura et al., 1992) and a
synthetic translation initiation sequence. The HSV-TK structural gene included in
the construct (McKnight, 1992) is placed behind the same promoter. The construct
was linearized in the bacterial plasmid sequence with Seal before electroporation.
78
a) 1.5 kb
pTZ19R.GCo
t 2 neo
pTZ19R.mP1neoPA
b)
Sac1
-
Hind 1X
t
ICO
r^
krH2 | neo I 2 -
i 1>£/
H'WTK N' .i i 1 n
PICI19R.MC1-TK
c nov-ris, - 2 1 neo
pICLPHR
*—i
1 2 -\
1 l^
Seal
Fig. 16. Cloning strategy for targeting construct pICLPHR. pICLPHR was cloned in
two steps, (a) First, the neo cassette from pMCLneoPA was inserted as a
Sall/Xhol fragment into pTZ19R.GC. (b) The Sad fragment carrying perforin
sequence and the neo cassette was isolated, the ends were converted to Hindm
compatible ends and ligated into the Hindlll site of pICI19R.MCl-TK to yield the
targeting construct pICLPHR.
79
a)
EwM
at£»
EcoflV
EeoRI
12
Ecoflv
I '
Sail
PTZ19R.GC
BstEil
EmRV
EmRV
QTCACO 1MI
» -k
Xhol Sail
C3
pMCLNeoPA
Xhoi/Sall
TCGACIIS
iCT
Klenow
Klenow
CIP
TCGAC
AGCTC!IS
EetftV
EMM
2
EeoRV1
OOTCAC OTCACCJlSHI
1I jSCAQTd CAOTQG|____
Ligation
Eton
1 •
Snat
1 Hna
HkMM EM*
1 1EmRV
21
1 llfNtoojf 1 3 fip
1M
~~r iSot f>m
1
| TEwM Ewm
1 |
12
1 1 fHH „r
PMfn
T'
Fti MM
3<a1 J
pT219R.VKp pTZ19R.VKap
SMI
„„;;., ;sS»«*«:u«»«|!»|i!«!»HW.TICI.i§>
PICI19.MCI-TK
Sail
;-,T;7ft,f jTCOAC
acr o"
Klanow
CIP
SBwIifiliaiiiiiiii!PSv-TX
XTCOA TCOAC
5AGCT AOCTO-
plCI.HK1
Ligation
PICI.HK2
Ecofil EcoH MM* EooH
tin
Been
Sail
Eoofli
Sail
cm
EcofM
EcoM CM
Sail
M HSV-TK
Sail
EooAt Goofll
Saat
PICI.HK3 PICI.HK4
81
Fig. 17. (previous two pages) Cloning strategy for targeting constructs
pICLHKl, pICLHKl, pICI.HK3 andpICI.HK4
a) The intermediates pTZ19R.VKp and pTZ19R.VKap were constructed by
insertion of the Xhol/Sall neo cassette from pMCLneoPA into
pTZ19R.GC. Insertion of the cassette in a parallel fashion in respect to the
transcriptional directions of the genomic perforin sequence yielded
pTZ19R.VKp. pTZ19R.VKap carries the neo cassette in the antiparallel
direction.
b) Ligation of an EcoRV/EcoRV fragment from pTZ19R.VKp or VKap
respectively into the blunt ended Sail site of pICI19.MCI-TK in both
possible directions gave rise to the four electroporaton constructs
pICLHKl, pICI.HK2, pICI.HK3 and pICI.HK4. Note that all four
constructs differ in the relative orientation of perforin genomic sequence,
neo cassette and HSV-TK gene.
The constructs were linearized with Clal before electroporation.
Transfection of ES cells with targetting constructs and screening for homologous
recombination at the perforin locus
The targeting constructs described above were introduce into embryonic stem cells
by electroporation. After growing the eiectroporated cells under selecting
conditions, the clonal colonies were screened for homologous recombination by
PCR and positive colonies were expanded for further characterization.
A constant number of ES cells were eiectroporated with 15 /Jg of linearized, exon 2
derived targeting construct pICLPHR and grown in the presence of G418. Some
experiments were performed with G418 and gancyclovir double selection hoping to
enrich for homologous recombinants (Thomas and Capecchi, 1987). To provide a
negative control, cells that were eiectroporated in the absence of plasmid DNA
were exposed concomitantly to the same selecting conditions.In all control
82
experiments no G418 resistant colonies were observed confirming the stringency of
the selection conditions.
Usually, the number of G418 resistant colonies recovered from 5 x 106
electroporated ES cells varied to a considerable degree with the ES cell line, the
batch of ES cells, the number of in vitro passages they have gone through and their
state of proliferation.
In order to establish the conditions for the PCR screening procedure, ES cells
were electroporated with the construction intermediate pTZ.mPlneoPA (Fig.16).
Random integration of this control construct can be used to simulate the screening
procedure for homologous recombination, since it results in the same PCR
fragment when tested with the screening primer set c/d (Fig. 18). The suitability of
the primer screening method was confirmed by the detection of the 12 kb
fragment amplified from as little as 10 lysed intermediate electroporated ES cells
in the presence of genomic DNA from 10000 cells. A fraction of the
electroporation experiments were screened with the improved primer set e/f,
which yielded the same sensitivity but worked more reliably because of the smaller
fragment amplified by this primer pair.
Results of the electroporation of D3H, D3 and BL/6-IH ES cells with pICLPHR
are shown in Tab.3. The electroporation of D3H cells yielded 190 neo resistant ES
cell colonies per 5 x 106 cells used for electroporation. The corresponding
frequencies were for D3 ES cells 134 and for BL/6-III cells 124 colonies per 5x10s
electroporated ES cells in average. The BL/6-HI cells compared to cells of D3
origin consistently yielded 25-50% less G418 resistant colonies. This fact most
probably reflects their slower proliferation rate and lower plating efficiency in in
vitro culture.
After electroporation with construct pICLPHR over 1176 colonies (not all data
shown) were screened for homologous recombination by PCR and subsequent
Southern blotting. The theoretical crossovers for targeting of the perforin locus,
the diagnostic Southern blotting fragments and the position of PCR primers
employed are shown in Fig. 18. No homologous recombinants could be recovered
83
from these experiments. Since there was a possibility, that the frequency of
homologous recombination with this construct was too low to find a targeted clone
in a screenable number of colonies, a positive and negative selection procedure
was applied (Fig. 17). This double screening procedure with G418 and gancyclovir
has been reported by some groups to increase the targeting frequency of constructs
flanked by one, two identical or two different (TK1 and TK2) herpes simplex virus
thymidine kinases (HSV-TK) by a factor of up to 2000 (Mansour et al., 1988) .
When double selection was applied to electroporation with the construct
pICLPHR, the frequency of resistant colonies dropped about fivefold compared to
selection with only G418 indicating that the additional gancyclovir selection
conditions were stringent. But still no homologous recombinant ES cell clone was
isolated from 342 screened G418/gancyclovir double resistant colonies.
Consequently, either the frequency of homologous recombination is not increased
by double selection, an observation reported also by other groups (Mansour et al.,
1988; Donehower et al., 1992; Chisaka et al., 1992; Thomas and Capecchi, 1990;
DeChiara et al., 1990; McMahon and Bradley, 1990) or the screening procedure by
PCR was not able to detect the mutation in Exon 2 of the perforin gene.
Alternatively, despite the fivefold enrichment by double selection, the frequency of
homologous recombinants might have been too low to allow the detection of
homologous recombination in the limited number of screened colonies.
Possibly, the failure to find homologous recombination with construct pICLPHR
was not due to an intrinsic property of the perforin locus but rather was a problem
associated with this specific construct. Probably it either resulted because the
length of homology was not sufficient to mediate homologous recombination or
alternatively because the secondary structure of the perforin locus or proteins
bound to the chromatin precluded homologous recombination or its detection by
PCR. In order to circumvent these potential problems, four additional targeting
constructs pICLHKl, pICI.HK2, pICI.HK3 and pICI.HK4 were established.
These vectors contain perforin homologous sequence derived from exon 3 and
from the flanking introns. In addition, the length of homologous sequence was
84
more than doubled from 1.5 kb of the exon 2 derived construct pICLPHR to 3.3
kb. Fig.20 shows the theoretical crossovers for homologous recombination with the
construct pICI.HK2, the lengths of the expected restriction fragments for the
wildtype and the targeted allele and the positions of the employed PCR primer
pairs. The restriction fragment lengths and PCR fragments of the other three
constructs are given in Tab.4. The results of the electroporation of ES cells with
constructs pICI.HKl-HK4 are listed in Tab.5. The frequency of G418 resistant
colonies recovered varied again considerably with the ES cell line and the batch of
cells used. No viable colonies were recovered from control electroporations
without construct DNA added, showing the stringency of the G418 selection
procedure. In a first experiment the same batch of D3H cells were electroporated
with the four targeting constructs and the colonies resulting from 5 x 106
electroporated ES cells were screened for homologous recombination. The
frequencies of neo resistant colonies per 5 x 106 ES cells varied between 30 for
HK2 and 240 for HK4. Homologous recombinants were isolated from
electroporation experiments with pICI.HK2 and pICI.HK4 at a frequency of 1 in
30 respectively of 1 in 200. These two constructs differing only in the orientation of
the neo cassette in respect to the homologous sequence yielded surprisingly
differing frequencies of homologous recombination and G418 resistant colonies
per 5 x 106 electroporated ES cells. Since both constructs gave about the same
number of homologous recombinant clones per 5 x 106 electroporated ES cells,
pICI.HK4 comparatively produced much higher numbers of colonies with the
construct integrated at a random site of the cellular genome. Thereafter, the
frequencies of homologous recombination could be reproduced with ES cell lines
D3, BL/6-III and D3M without any statistically significant difference in
frequencies.
85
E poly A
1 Kb
•1.5 kb-
H-1200bp-H MM** |HomoJogousCM* -
•16kb-
recombination
Slop
-2.1 kb 15.2kb-
Fig.18. Disruption of the perforin gene in mouse embryonic stem (ES) cells with the
targeting constructpICLPHR
a) The targeting construct pICLPHR consisted of a 1.5 kb Sad fragment, which
was derived from perforin exon 2 and from parts of the flanking introns,
interrupted by a neo cassette of 1.2 kb length. The neo gene is driven by a HSV-
TK promoter. EcoRI (E), Sad (S) and Smal (M) sites are indicated.
b) Map of the murine perforin gene. The hypothetical crossovers with the targeting
construct are shown with crossed lines. The endogenous locus yields a 16 kb EcoRI
band when analyzed by Southern blotting with a Exon 3 probe (B) and a neo probe
(Q.
c) The disrupted perforin locus can be detected by a PCR reaction with primer c
and d or alternatively with primers e and f. Note the additional EcoRI site in the
disrupted allele that the neo cassette introduces into the locus. Consequently,
EcoRI digestion of the targeted locus will yield Southern blotting fragments of 2.1
kb and 15.2 kb depending on the indicated DNA probes.
86
a Gene Targeting
neo'
HSV-tk
1~T
f 11 II 1 1 1 1
Gene X
\noo'
1 1 1 V//////A 1 1
x-neo' HSV-tk"(G418^GANC)
b Random Integration
HE
neof HSV-tk
mm-Tv-mm
x'noo' HSV-tk*<G418r,GANC')
Fig. 19. The positive-negative procedure used to enrich for ES cells containing a
targeted disruption of gene X (Mansour et al., 1988). a, A gene X replacement
vector that contains an insertion of the neor gene in an exon of gene X and a
linked HSV-TK gene, is shown pairing with a chromosomal copy of gene X.
Homologous recombination between the targeting vector and genomic X DNA
results in the disruption of one copy of gene X and the loss of HSV-TK sequences.
Such cells will be X-, neor and HSV-TK- and will be resistant to both G418 and
Gancyclovir. b, Because non-homologous insertion of exogenous DNA into the
genome occurs through the ends of the linearized DNA, the HSV-TK remains
linked to the neorgene. Such cells will be X+, neor and HSV-TK+ and therefore
resistant to G418 but sensitive to Gancyclovir. Open boxes denote introns or
flanking DNA sequences, closed boxes denote exons and cross-hatch boxes denote
the neor or HSV-TK genes.
87
E poly A
Start
X
—fr - 2
3Jkb 1
16kb-
fr
Homologous recombination
Start
J R
I l-665bp—|
i. b
StopR
.5.3 kb. +
8 B
C
tkb
•115 Kb
Fig.20. Disruption of the perforin gene in mouse embryonic stem (ES) cells with the
targeting constructpICLHKZ
a) The targeting construct pICLHK2 consisted of a 33 kb EcoRV fragment, which
was derived from perforin exon 3 and from a part of the preceding intron,
interrupted by a neo cassette of 1.2 kb length. The neo gene is driven by a HSV-
TK promoter. EcoRI (E), EcoRV (R) and BstI (B) sites are indicated.
b) Map of the murine perforin gene. The hypothetical crossovers with the targeting
construct are shown with crossed lines. The endogenous locus yields a 16 kb EcoRI
band when analyzed by Southern blotting with an exon 2 probe (A), an exon 3
probe (B) and a neo probe (C).
c) The disrupted perforin locus can be detected by a PCR reaction with primer a
and b. Note the additional EcoRI site in the disrupted allele introduced by the neo
cassette. Consequently, EcoRI digestion of the targeted locus will yield Southern
blotting fragments of 5.3 kb and 11.9 kb depending on the indicated DNA probes.
88
Tab.3 Electroporation ofES cells with exon 2 derived construct pICLPHR and effect of Gancyclovir
selection on thefrequency ofhomologous recombination.
Construct ES cell line
Colonies
per5xl0*cells Colonies screened Targeted ES clones
pICLPHR D3H 190 160 0
pICLPHR D3 134 270 0
pICLPHR BL/6-in* 165 496 0
pICLPHR + Ganc BL/6-m* 31 188 0
pICLPHR BL/6-m# 83 250 0
pICLPHR + Ganc. BL/6-m# 25 154 0
5xl06 ES cells of different lines were electroporated with linearized plaamid construct DNA. After 10
days of G418 selection die numberofresistant colonieswas deteniiined and a certam number ofresistant
colonies assessed for homologous recombination by PCR and genomic Southern Blotting. In certain
experiments, double selection with G418 and Gancyclovir was applied. Symbols (*, #) indicate dut die
same batch of ES cells was used for a series of dectroporations and that die frequencies can be compared
directly.
Tab.4. Expected restriction fragmentpatterns and PCRproducts after homologous recombination of
the perforin locus with thefour exon 3 derived constructs. For die localization of die probes
seeFig20.
EcoRl fragments
Construct Probe A Probe B Probe C Primer pair PCR fragment
pICLHKl 16.0. 6.0 16, 11.2. 6.0 6.0 b/d 735 bp
pICLHK2 16.0. 5.3 16. 11.9, 5.3 11.9 a/b 665 bp
pICLHK3 16.0, 5.3 16. 11.9, 5.3 11.9 b/d 665 bp
pICI.HK4 16.0. 6.0 16. 11.2, 6.0 6.0 a/b 735 bp
89
Tab.5. Statistics ofES cell electroporation withperforin exon 3 derived targeting constructs
Construct
Colonies
ES cell line per 5xl0*cells Colonies screened Targeted ES clones
pICLHKl D3H * 60 60 0
pICLHK2 D3H * 30 30 1
pICLHK3 D3H * 180 180 0
pICLHK4 D3H * 240 200 1
pICLHK4 D3 105 200 1
pICLHK2 BL/6-m # 120 150 1
pICLHK4 BU6-HI # 420 150 0
pICLHK2 BL/6-m 51 142 1
pICLHK2 D3M 43 130 2
pICLHK2 B1V6-HI 21 62 4
pICLHK2 BU6-IH 14 83 2
5xl06 ES cells of die indicated ES cell lines were electroporated with linearized plasmid DNAi After 10
days of G418 selection die number of resistant colonies was determined and a certain number of resistant
colonies assessed for homologous recombination by PCR and genomic Southern blotting. Symbols (*, #)
indicate, that the same batch of ES cells was used for a series of electroporations and that the frequencies
can be compared directly.
90
Example for the screening ofa successful electroporation experiment
Six lots of BL/6-QIES cells were electroporated with 30 Mg of linearized pICI.HK2
plasmid DNA. After selection for G418 resistance, the 83 resistant colonies were
picked as single clones. The colonies were divided into two equal halves, one half
passaged to a 96 well plate and the other half analyzed as crude cell lysates of
about 50 ES cells for PCR. The PCR reaction products were separated on an 1.2
% agarose gel and since no bands were visible by ethidiumbromide staining, the
gel was blotted to a nylon membrane and hybridized with the labelled respective
PCR fragment. The result of the autoradiographic exposure of the membranes is
shown in Fig.21. Besides two clearly positive clones on lane 59 and 75 (translating
to clones 134-74 and 134-96 ) eight further clones on lanes 12, 18,39,46,49,51, 64
and 79 ( translating to clones 134-17, 134-25, 134-48, 134-56, 134-60, 134-62 and
134-102) showed a fainter band at the expected length of 665 bp.
These ten clones were grown up, genomic DNA was isolated and further analyzed
by Southern blotting (Fig.22a,b and c). Homologous recombination was confirmed
in clones 134-74 and 134-96, which both had been strongly PCR positive. For none
of the weakly positive ES cell clones the genomic 5.3 kb EcoRI restriction
fragment diagnostic for homologous recombination was detected by hybridisation
with flanking probe A (Fig.22a), thus identifying the faint PCR band amplified
from these randomly integrated clones as an artifact. The hybridisation of the filter
with the exon 3 probe B (Fig.22b) internal to the construct and with probe C
complementary to the neomycin resistance gene (Fig.22c) confirmed homologous
recombination in clones 134-74 and 134-96 by the detection of the 5.3 kb and 11.9
kb bands, respectively of only the 11.9 kb band (see also Fig.20).
In all of the 8 clones characterized by only a faint PCR band, the absence of
homologous recombination was proven by the failure of flanking probe A to detect
the 5.3 kb band. Hybridisation with probe C documented the presence of the
randomly integrated construct DNA in those clones. In some clones the introduced
DNA remained intact while in others it was partially degraded. Except for clone
134-60, integration of a single copy of the construct was demonstrated in these
91
clones. The complex pattern of bands with varying intensities detected by probe B
and probe C in clone 134-60 probably resulted from multiple random integrations
of construct fragments displaying varying patterns of exonuclear degradation.
The result of the genomic Southern blot was confirmed also by PCR analysis of 2
/ig purified genomic DNA, an amount of PCR template that allows the detection
of the diagnostic PCR fragment directly on the ethidiumbromide stained agarose
gel (Fig.23). The targeted mutation in clones 134-74 and 134-96 was confirmed by
this test.
Thus, two homologous recombinant clones were isolated from 83 neo resistant ES
cell colonies and their genotype was confirmed by genomic Southern blotting and
PCR amplification specific for the targeted mutation.
92
Ol Ol
665 bp —
no m m7!O Ol 00 T3
2 <o
COo U1
Olo
665 bp —' t» #
Ol O) CD
-go oi~J -«J 00 3 ^O Ol O ?TTJ
665 bp —
93
Fig.21. (previous page) PCR Screening of an electroporation experiment with exon 3
derived construct pICI.HK2. G418 resistant colonies electroporated with linearized
pICI.HK2 were screened by PCR with primer pair a/b (Fig.20). Homologous
recombinant clones are expected to yield a fragment of 665 bp length. VKp
denotes amplification of the 665 bp fragment from 50 pg of the plasmid control
pTZ19R.VKp (Fig.17). Two clearly positive clones(lane 59 and 75, translating to
clones 134-74 and 134-96) and 8 weakly positive clones (lanes 12,18,39, 46,49,51,
64 and 79, translating to clones 134-17,134-25,134-48,134-56,134-60, 134-62,134-
81 and 134-102) were expanded for further analysis by genomic Southern blotting.
94
CM
r^-mootoocM^ti-ooi- tM t W lOlfl h- CO a> t-
I I I I I III
cocoeocoeocoeocococo
Q.
>
Wm
£t- cq oa
16 kb
11.9 kb—
mm -" m*nm- Ml Mil *—«»'-» %» Jtft ««M »
5.3 kb —
*>U(- « .»-- ?*_».
Probe A
CMa.
SIOOIOOMfrlCOlO^r(M*WIO(ONOOO)i- Is- >
i i i i i i i
cocoeocoeocoeocococo oj *° *£
B
16 kb — |
11.9 kb—!
5.3 kb —
Probe B
CM °-
rN^lfllBIONOOOIr h»>• i i i i i i *T
cocoeocoeocoeocococo w 2 »
rrrrrrrrrrrDO
16 kb_
11.9 kb—
5.3 kb —
Probe C
95
Fig.22. (previous page) Genomic Southern blotting ofPCR positive clones (compare
Fig.21). Genomic DNA was digested with EcoRI, blotted to filters and hybridized
with a flanking Probe A (a), stripped and rehybridized with an internal probe B (b)
and a Probe C from the neo coding region (c). For localization of the three probes,
see Fig.20. Expected fragment lengths can be taken from Fig.20 and Tab.4. B6
indicates wildtype C57BL/6 kidney genomic DNA and B6 + VKp means the
former genomic DNA mixed with 250 pg of positive control plasmid pICLVKp.
96
Fig.23: Confirmation of the PCR screening by amplification of clean genomic DNA.
Ten PCR positive clones (Fig.21) were grown up, used for isolation of genomic
DNA and 2 /zg of genomic DNA checked by PCR with primers a and b for
homologous recombination. The signals were strong enough to be detected
without Southern Blotting of the PCR gel. B6 denotes normal genomic DNA from
BL/6-DI ES cells, VKap indicates 50 pg of the positive plasmid control.
Generation ofmice heterozygous for the disruption of the perforin gene by injection of
homologous recombinant ES cells into blastocysts
All blastocyst injections were carried out by Birgit Ledermann and Kurt Burki,
Sandoz AG, Basel.
Homologous recombinant ES cell clones were judged by the morphology of their
colonies in vitro and then injected into blastocysts. The injected blastocysts were
transferred to foster mothers and eventually gave rise to chimeras.
Several homologous recombinant ES cell clones were obtained from the D3
derived ES cell lines (D3H, D3M, D3). But microinjection of these clones did not
yield any germ line chimeras inspite of the considerable effort made.
The further efforts concentrated on the homologous recombinant clones
originating from the BL/6-III ES cell line. BL/6-III ES cell clones were injected
97
into BALB/c blastocysts and transferred to CD1 foster mothers. The male
chimeras with a black and white spotted fur were bred with C57BL/6 females. The
black fur colour of the offsprings indicated germline transmission of the ES cell
genotype. Transmission of the BALB/c genotype yielded brown fured offsprings
because of the presence of the dominant agouti depigmentation allele in the
BALB/c mouse strain.
Of six homologous recombinant BL/6-B3 clones that were used for blastocyst
injection only one gave rise to germline chimeras. In Tab.6 a series of four clones
is shown with results of the blastocyst injection that clearly show their non-
pluripotent phenotype. Hallmarks of non plunpotency were small or missing litters
of the fostermothers, no or a small proportion of chimeras in the litters and
chimeras born already dead or dying in the first week after birth. But from six
tested clones one, 127-76, was pluripotent and 33 chimeras from 19 litters were
derived from this clone (Tab.7). A high 30 to 3 ratio of male to female chimeras, a
hint for the presence of ES cell derived cells in the germ line of the chimeras, was
found.
9 of the 30 male chimeras descending from clone 126-76 proved to be 100%
transmitting germline chimeras, i.e. chimeras derived from a female blastocyst that
were converted during embryonic development by the male ES cells to a male
phenotype (Fig.24). This was reflected by the black coat colour of all pups when
the chimeras were bred to C57BL/6 female mice. Statistically, 50 % of these
offsprings were expected to be heterozygous for the targeted mutation. The
presence of the mutation in the offspring was confirmed by Southern blotting of
genomic DNA and hybridization with flanking probe A (Fig.25). The heterozygous
animals appeared healthy and no difference to their wildtype littermates was
noted.
Tab.8 shows an analysis of fur colour chimerism and the type of germline
transmisssion of 30 male chimeras obtained from the pluripotent clone 127-76. All
100 % germline transmitting chimeras displayed a relatively high degree of
chimerism between 40 and 90% (Fig.24), while the fertile chimeras not
98
transmitting the ES cell derived phenotype showed a clear tendency towards lower
degrees of chimerism. Nonfertile chimeras occurred at all degrees of chimerism.
Genetically the genotype of the offsprings of 100% transmitting chimeras is
completely derived from C57BL/6 mouse strain except for the mutation in the
perforin locus. Therefore, from the breeding of a heterozygous male and female
mouse animals will be obtained, that are deficient for perforin while retaining the
C57BL/6 genotype. The well characterized C57BL/6 genotype will facilitate the
functional analysis in various immunological tests considerably.
99
Tab.6. Result ofthe blastocyst injection offour non-phvripotentES cell clones
Clone Offsprings Chimeras Dead Chimeras
128-13
128-19
128-29
128-52
4 1 1
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
4 1 1
0 0 0
2 0 0
5 1 1
7 3 3
0 0 0
0 0 0
2 1 1
0 0 0
0 0 0
1 0 0
3 2 2
0 0 0
5 4 4
1 1 1
0 0 0
Homologous recombinant clones 128-13, 128-19, 128-29, 128-52 are not
pluripotent The clones were injected into BALB/c blastocysts. 15 -21 injected
blastocysts were then transferred to each fostermother. After birth, the number
of litters, their chimerism and survival rate reflects the pluripotency and abilityof the ES cells to cooperate with blastocyst derived cells in embryonic
development
100
Tab.7. Results ofthe blastocyst injection ofpluripotentEScell clone 127-76
Chimeras
Clone Pups Born Dead Alive Male Female 100% GL
0
0
- - - - - -
0
6 3 0 3 3 0 0
10 9 5 4 4 0 3
11
0
2
0
4
9 0 9 8 1 1
2 0 2 2 0 1
4 4 0 • . •
5 4 2 2 2 0 2
9 9 6 3 2 1 1
3 2 0 2 2 0 0
7 5 0 5 5 0 1
9 9 6 3 2 1 0
3
0
3 3 - - - -
0
6 6 6
- - - -
Total 75 65 32 33 30 3 9
The targeted BL/6-m Gone 127-76 was injected into BALB/c blastocysts. 15-21
injected blastocysts were then transferred to each fostermother. After birth, the number
of litters, their chimerism and survival rate reflects their pluripotency and ability to
cooperate with blastocyst derived cells in embryonic development
101
Fig. 24. Three germline transmitting chimeras derived from ES cell clone 126-76. ES
cell contribution results in black spots in the white fur. From left to right: Goliath,
Ben Hur and Al Sabah. Note that inspite of the varying degree of coat colour
chimerism all three mice passed the ES cell phenotype to 100% of their offsprings.
16 kb
5.3 kb
- fJ^ <D «>
*•*•' «»«• «•**» |p 4^.
Fig.25. Southern blot analysis of 9 offsprings from a 100% germline chimera.
Genomic DNA was isolated from tail biopsies, digested with EcoRI and hybridized
to Probe A (Fig.20) after separation and blotting. 6 of 9 litters are heterozygot for
the disruption of the perforin gene resulting in the appearance of the diagnostic 5.3
kb band.
305
20
00
20
10
00
(inf
erti
leN
13
00
20
32
31
20
NonGL
31
20
00
00
00
0GL
Partial
90
21
23
10
00
0%GL
100
Total
-100%
90
-90%
80
-80%
70
-70%
60
-60%
50
40-50%
-40%
30
-30%
20
-20%
10
-10%
0Transmission
chimerism
colour
Fur
Germline)
(GL:
127-76
clone
cell
ES
targ
eted
the
jrom
generated
chimeras
male
30
the
ofchimerism
colour
withfitr
transmission
o/ge
rmli
neCorrelation
8.Tab.
103
Discussion
In order to clarify the role of perforin for cytolysis, mice with a mutation of the
perforin gene were produced by gene targeting. To this end, ES cells were
transfected with five different targeting constructs. The first construct was designed
to disrupt the perforin gene by homologous recombination to exon 2. A second
group of four related constructs targeted exon 3. In both cases, a neomycin
resistance gene with translation stop codons in all three reading frames was
introduced into the perforin gene, thus abrogating the production of functional
perforin protein in cells with both alleles disrupted. Homologous recombination
was obtained with two of the exon 3 derived constructs at different frequencies,
while no homologous recombinant ES cell clone could be isolated from
transfection experiments with the exon 2 derived construct. Subsequently, several
male chimeras transmitting the targeted mutation to their offsprings were obtained
from blastocysts injected with pluripotent targeted ES cells. By breeding these
heterozygous offsprings, homozygous perforin deficient mice can be obtained.
Their phenotype will then allow an assessment of the natural function of perforin
in vivo.
Two targeting constructs derived from perforin exon 3 successfully and consistently
induced homologous recombination in ES cells. Although they differed only in the
direction of the neo resistance gene relative to the homologous sequence, the
observed recombination frequencies differed considerably: The first construct
recombined homologously at a frequency of 1 in 30 G418 resistant colonies, while
for the second a frequency of 1 in 240 was observed. How can this significant
difference be explained considering that the two constructs differ only in the
orientation of the neomycine resistance gene? First, a conformational difference
between these two constructs, which is caused by the orientation of the cassette,
could either favour or inhibit homologous recombination. Alternatively, only one
direction of the neo gene in exon 3 of perforin could allow the expression of the
phosphotransferase. The construct carrying the neomycine resistance gene in the
same direction as the endogenous locus yielded homologous recombination
104
frequencies eight times higher than the construct with the resistance gene in
reverse direction. This finding is explained by an inhibition of the
phosphotransferase promoter when it is placed in reverse direction to the
endogenous promoter of the targeted gene. Consequently, the homologous
recombinant cells are more susceptible towards the selecting agent G418 and
would, at least partly, not survive the selection procedure. It is noteworthy, that the
inhibition of homologous recombination with a vector carrying the neo gene in
reverse direction to the homologous sequence was reported before for the fl2-
microglobulin gene (Mansour et al., 1988).
A first attempt to induce homologous recombination with a construct derived from
exon 2 was not successful. The failure to generate homologous recombinant ES-
cell clones with construct pICLPHR could be explained by the following reasons:
(i) Exon 2 might not lead to homologous recombination, either because the
chromosomal DNA is in a unfavorable secondary structure, or alternatively
because the access of introduced construct DNA to the genomic locus is blocked
by DNA binding proteins, (ii) Only 1.5 kb homologous sequence were contained in
the targeting construct pICLPHR, compared to typically more than 5 kb of
homology reported in successful recombination experiments. Possibly, more than
1.5 kb sequence homology is required for recombination of perforin exon 2. (iii)
Homologous recombination events actually occurred but were not detected,
because the screening procedure was not sensitive enough.
All targeting vectors used for transfection experiments contained a point mutation
known to decrease the frequency of G418 resistant cells (Koller and Smithies,
1989; Ledermann, 1991). Electroporation experiments with two targeting
constructs containing this mutation (pICLPHR and an IL-3 construct) and with the
unmutated pMCI.neoPA vector showed, that both mutated constructs yielded
about ten times less G418 resistant colonies (Yenofsky et al, 1990). Thus, the
presence of the point mutation in the perforin targeting constructs leads probably
to a reduction of G418 resistant ES cells in these experiments. But since the
homologous recombination frequency of construct pICLHK2 was relatively high,
105
this was not a substantial problem. However, there is a possibility that the
mutation not only influenced the frequency of overall G418 resistant cells but also
the frequency of homologous recombinant colonies. Two cases can be
distinguished: First, the homologous genomic locus might be favorable for
expression of transfected gene, either because of its open secondary structure, or
alternatively because of adjacent promoters and enhancers. A less efficient
resistance gene expression vector leads in this case to a reduction of non-desired,
random integrants, because the expression of the neomycin resistance gene at the
random genomic locus is dependent on additional endogenous transcription
promoting elements. Consequently, the combination of a weak expression vector
and a targeted locus favoring expression can result in an increase of the frequency
of homologous recombination. Essentially, this phenomenon has been exploited to
target genes, that are expressed in ES cells by applying constructs with the
neomycine resistance coding sequence without a promoter (Ledermann, 1991) or
alternatively by using the neo cassette without a polyadenylation signal
(Schwartzberg et al., 1990; Sedivy and Sharp, 1989). On the other hand, a weakly
expressing neomycine gene is disadvantageous in the case of recombination to a
genomic locus unfavorable for gene expression. Homologous recombinant cells
would not be G418 resistant in this case, while all the viable G418 resistant
colonies would consist of randomly recombined cells, in which the expression of
the resistance gene is enhanced by the environment of the random integration site.
No direct evidence for such an effect was obtained, but eventually the successful
constructs pICI.HK2 and pICI.HK4 profited from this mechanism while some
unknown property of the non-successful constructs pICLHKl and pICI.HK3
inhibited a favorable interaction.
The experience of seven years of research in homologous recombination in ES
cells have left the researchers with very little knowledge about the parameters
influencing the recombination process, since this subject suffers from a lack of
systematic investigations. Current knowledge ( or current state of ignorance) can
be summarized as follows: Expression or lack of expression of a certain gene does
106
neither preclude nor reduce homologous recombination (Donehower et aL, 1992;
Joyner et aL, 1989). The frequency of homologous recombination seems to be
favorably influenced by the length of homology between the construct and
endogenous targeted gene in the range between 0.5 and 7 kb of homology
(Johnson et aL, 1989). Recombination frequencies are neither decreased by a
disruption of the homologous sequence by a stretch of non homologous sequence
nor by one end of the construct having only little homology as long as it is more
than 0.5 kb (Hasty et aL, 1992; Thomas and Capecchi, 1987). In addition, it is
believed that the use of isogenic genomic DNA in the targeting vectors favours
homologous recombination, i.e. that the homologous genomic sequence in the
construct preferentially originates from the same inbred mouse strain as the
electroporated ES cell line (Hasty et aL, 1992). Although these days, most
approaches follow these guidelines, the reported frequencies of homologous
recombination still vary in a wide range between 1 in 20 and 1 in 10000 G418
resistant colonies. However, it is not clear if some of the reported low frequencies
reflect real frequencies or are merely caused by an unreliable screening scheme.
That the above guidelines are not absolute "conditiones sine quae non" is
illustrated by the fact, that the homologous sequence in the successful construct
pICI.HK2 was shorter than the reported optimal length of homologous sequence
and was not isogenic.
The exemplary screening procedure described above illustrated the isolation of 8
clones with a random integration of the targeting construct despite the, although
weak, positive result from the PCR screening (Fig.21 and Fig.22). The faint, false-
positive band observed after PCR-amplification of crude cell lysates can be
explained by the synthesis of so called "shuffle" or hybrid templates. Hybrid
templates can result from a primer that has been partially extended on a random
integrated construct switching to the perforin locus in a subsequent cycle. This is
possible because of the homologous sequence present in the random integrated
construct, which provides the perforin specific sequence to the partially extended
primer. The artificial "shuffle" template is amplified in the further course of the
107
PCR to a fragment identical to the fragment diagnostic for homologous
recombination, as it has been reported before (Tibulewicz et al., 1992). Since it is
not always easy to differentiate a true PCR band from a fainter false positive,
especially in the absence of a homologous recombinant colony, only subsequent
genomic Southern blotting confirms homologous recombination with high
confidence.
It is known that the genomic integration of more than one construct copy is only
rarely observed in electroporated eukaryotic cells (Kim and Smithies, 1992): Of 10
clones checked in this experiment only one carried more than a single copy of the
construct (Fig.22b). Interestingly, the 1.5 kb EcoRI fragment of the linearized
plasmid showed up in all but one randomly integrated clones, while the 4.2 kb
fragment, arising from the more 5' region of the construct, was found in only 2 of
the random recombinant clones. This finding could result because the 4.2 kb 5'-
restriction fragment is protected by only 200 bp of flanking sequence against
exonuclease activity (Fig. 15). This is in contrast to the 1.5 kb fragment, which is
protected by 2.9 kb of vector sequence. The absence of the 4.2 kb band is
explained by the presence of intracellular DNAse activity, which might degrade
transfected DNA in the cytoplasm before stable integration into the genome takes
place. This intracellular DNA degradation might well be the reason for the higher
frequencies of homologous recombination found with longer constructs. Higher
molecular weight could protect the construct against degradation and
consequently, the higher recombination frequencies with longer constructs would
just reflect the higher stability of a transfected long DNA construct.
Of the six ES cell clones injected into blastocysts only two proved to have
preserved pluripotency and consequently gave rise to germline chimeras. The most
frequent cause for the loss of the pluripotent phenotype of ES cell lines lies in non-
optimal cell culture conditions. The finding that clones from some electroporation
experiments lost pluripotency while others conserved it, was correlated with a
change from senile, slowly growing embryonic feeder cells to a more rapidly
proliferating batch of feeder cells. Of course, this point was not investigated
108
systematically and the additional cell culture conditions were difficult to keep
absolutely constant; but senile feeder cells seem to be at least partially responsible
for the loss of pluripotency during electroporation and selection. This finding was
further consolidated by the identification of another pluripotent targeted ES cell
clone in a subsequent electroporation experiment, for which the same batch of
fresh feeder cells was used. This additional pluripotent ES cell clone will be used
to generate a second line of mice lacking the perforin gene, which can serve as a
control to exclude an unlikely mutation cosegregating with the perforin gene in the
first line of deficient mice.
In conclusion, mice carrying a heterozygous mutation in exon 3 of the perforin
gene were produced by homologous recombination with relatively high efficiency
in ES cells. The perforin deficient mice obtained by breeding the heterozygous
animals will be a valuable tool to get further insight into the mechanism of
cytotoxicity by lymphocytes.
109
8. General Discussion
In the first part of this thesis, evidence for an involvement of the cytolytic protein
perforin in cytolysis of virus infected cells by specific cytolytic T cells was provided.
A close kinetic correlation and histological association between perforin
expression and antiviral cytolytic activity was consistent with the concept, that
perforin expression indeed is required for cytolytic activity. In order to test this
hypothesis further, mice with a disruption of the perforin gene were generated by
gene targeting as described in chapter 2. These mice will allow a direct assessment
of the role of perforin in cytolysis by lymphocytes as well as the evaluation of so far
unknown aspects of perforin functions in vivo.
The analysis of perforin expression during an acute infection of mice with
lymphocytic choriomeningitis virus presented in chapter 1 confirm the granule
exocytosis model for cytotoxicity in four important points in vivo: First, perforin
expression does occur in a physiological situation, namely during acute infection
with the noncytopathic LCM virus, in which the prominent role of lytic cytotoxic T
cells in mediating not only virus elimination but also immunopathological
symptoms is well established. This directly contradicts previous claims (Potter et
al., 1984) stating that perforin expression is an artifact observed only in in vitro
cultivated cytotoxic T cells as a result of the exposition of T cells to excessive
amounts of IL-2. Since tissue culture media complemented with supernatant from
concanavalin A stimulated rat spleen cells as a source of IL-2 are required to
maintain cloned cytotoxic T cells in culture, high expression of perforin in cultured
cytolytic cells was claimed to be caused by overstimulation with excessive amounts
of IL-2. Second, the close histological association of perforin expressing
lymphocytes with virus infected liver and brain cells strongly suggests an
involvement of perforin expressing cells in virus elimination. Third, the
comparison of serial sections stained with an antibody specific for the CD8 surface
molecule was consistent with the previous finding, that perforin is expressed in the
CD8 positive cytolytic T cell subset; and fourth, the 48 hours delay between
110
cytolytic activity and anti-viral cytotoxicity after infection again is arguing for an
involvement of perforin in the generation of MHC restricted anti-viral cytotoxicity.
Several subsequent studies conducted by other groups confirmed the expression of
perforin in vivo in other experimental systems with an involvement of cytotoxic
lymphocytes. Perforin expression has been demonstrated in pancreata of NOD
mice developing spontaneous diabetes (Clark, 1988), in human heart transplants
(Young et al., 1989) and in lymphocytes from the synovial fluid of patients with
rheumatoid arthritis (Griffiths et al., 1991). Although these observations provide
strong evidence for a role of perforin in natural antiviral, autoimmune and
allospecific cytotoxicity, they can not be taken as a formal direct proof. Especially,
they fail to answer the intriguing question, whether perforin is the relevant effector
molecule or merely expressed as a side effect of cytotoxic lymphocyte activation.
Assuming that perforin is an important effector molecule able to induce lysis of
target cells by itself, it would be important to know if additional redundant and/or
synergistic lytic mechanisms do exist. These questions were addressed by two
recent experimental approaches. In the first, CTLs were treated with antisense
perforin oligonucleotides during activation in vitro with an anti-CD3 antibody to
prevent perforin synthesis (Griffiths et al., 1992). The partial inhibition of perforin
synthesis led to a reduction of cytolytic activity. In the second, a noncytotoxic rat
mast cell tumor line, equipped with a functional secretory machinery, was
transfected with a perforin expression gene construct (Acha-Orbea et al., 1990).
The complementation with perforin rendered the mast cell line cytolytic. This
strongly argues for perforin being the sole effector molecule for cytolysis, although
both studies are compromised to some degree by the fact, that the cytotoxic
activity investigated was somewhat artificial antibody mediated activity in vitro
instead of natural cytotoxicity in vivo.
Consequently, the mechanism of cytotoxicity remained a controversial issue and it
became clear, that perforin deficient mice would be extremely useful to test the
different prevailing concepts and models for cytotoxicity. The recent establishment
of gene targeting by homologous recombination in embryonic stem cells allows the
111
generation of mice deficient in any known gene, provided it can be isolated from
genomic DNA (Shiver et al., 1992; Shiver and Henkart, 1991). The application of
this new powerful technique to the perforin gene generated mice carrying one
disrupted perforin allele as described in chapter 2. These heterzygot animals will
then be used to breed inter se to generate homozygousity for the deficient perforin
allele. First investigations of these mice will focus on the following aspects: The
capability to mount an effective, antiviral, alloreactive and NK cytolytic response,
the ability to eliminate the LCM virus from the organism after inoculation and the
consequences of the lack of perforin on the fertility of female mice as suggested by
reports of perforin expression in granulated metrial gland cells of the placenta.
The different prevailing theoretical models to account for lymphocyte mediated
cytotoxicity can be used to make predictions about the consequences of the
deletion of the perforin gene:
1. If it is assumed, that the granule exocytosis model is true and that perforin is the
sole effector molecule, then the deficient mice should not be able to generate any
cytotoxic activity against virus-infected, allogeneic or tumor cells. Concomitantly,
the immunopathological symptoms during LCMV infection would fail to develop
as well, i.e. intracerebrally infected perforin deficient mice would not succumb to
the choriomeningitis observed in fully immunocompetent animals. However, if
some of the immunopathology still develops in the absence of any detectable
cytotoxic activity, these symptoms would not be the consequence of cell destruction
but more likely of the lymphokines secreted during the inflammatory process in
the infected tissues. Assuming, that perforin deficient mice lack cytotoxic activity in
the cytotoxic T cell and the NK compartment, they could provide a new
interesting model system to assess the role of both cytolytic cells during infection
with viruses and in a number of other physiological situations.
2. A second concept involves the notion, that perforin and an alternative
mechanism (or mechanisms) is synergizing in a redundant manner to generate cell
lysis (Bradley et al., 1984; Thomas and Capecchi, 1987; Evans and Kaufman, 1981).
In this case, one has to consider that complete elimination of perforin already
112
during development would probably lead to a stimulation of the second pathway
resulting in a compensation of the perforin depletion. Such an effect has been
observed in CD4 deficient mice, in which a compensating helper cell activity is
able to induce antiviral immunoglobulin G after infection with vesicular stomatitis
virus, while animals depleted of CD4 positive T cells with an anti-CD4 antibody do
not develop immunoglobulin of this T-help dependent subclass (TKundig,
personal communication).
In this case, one would expect cytotoxic activity still to be detectable in perforin
deficient mice, although possibly at a lower level. However, the alternative
pathway of lysis would result in different kinetics and morphological appearance of
target cell lysis and the susceptibility patterns of different target cells would be
altered as well. It is well conceivable, that some target cells are susceptible to one
effector molecule, while others are resistant to this effector and sensitive to a
second, or to comphcate things further, succumb only to the synergistic attack of
two effector molecules. Assuming that alternative mechanisms do exist, mice
deficient for one main effector molecule would facilitate the identification of the
additional effector mechanism significantly.
3. It was postulated that apoptosis it the sole mechanism responsible for cytolysis
by lymphocytes (Baenziger et al., 1988; Lancki et al., 1991; Young et al., 1988;
Taylor and Cohen, 1992; Berke, 1991; Young and Liu, 1988). In this case, the
elimination of perforin would not influence at all the immunological competence
of the mutated mice, i.e. they would be perfectly healthy and cytolytic activity of
their CTLs and NK cells would be normal. In addition, no kinetic or morphological
deviation in target cell killing compared to normal mice should be observed.
4. Recently, three proteins associated with the granula of cytotoxic cells, namely
granzyme A, a rat NK granule serine protease and the polyadenylate binding
protein TIA-1, have been shown to induce DNA fragmentation in permeabilized
target cells (Russell, 1983; Golstein, 1987; Martz and Howell, 1989; Meyer et al.,
1984). These findings suggest a slight modification of the original granule
exocytosis model, in that access of these cosecreted molecules to the target cell via
113
the poly-perforin pores would result in the apoptosis observed during lysis of
certain target cells by a yet unknown effect on the target cell nucleus. This
modification would account for the lack of target cell DNA fragmentation
observed in cells lysed by purified perforin protein, a finding that has been often
taken as evidence against the granule exocytosis model. But since all three DNA
fragmentation inducing molecules require a prior permeabilisation step to gain
access to the target cell, the perforin deficient mouse is still expected to be devoid
of cytolytic activity.
Depending on the outcome of the functional characterisation of perforin deficient
mice, the relevance of the candidate models and hypothesis for cytolysis by
lymphocytes will be evaluated. Despite the considerable evidence for a role of
perforin in lymphocyte-mediated cytotoxicity the controversy about this issue is
continuing to be unresolved. It seems, that with the availability of perforin
deficient mice, it will be possible to shed some light on a discussion, that
sometimes has been rather dogmatic than enlightened.
114
9. References
Acha-Orbea, H., Groscurth, P., Lang, R., Stitz, L., and Hengartner, H. (1983).Characterization of cloned cytotoxic lymphocytes with NK-like activity. /.
Immunol 130,2952-2959.
Acha-Orbea, H., Scarpellino, L., Hertig, S., Dupuis, M., and Tschopp, J. (1990).
Inhibition of lymphocyte mediated cytotoxicity by perforin antisense
oligonucleotides. EMBOJ. 9,3815-3819.
Algire, G.H. (1957). Studies on tissue homotransplantation in mice using diffusion-
chamber methods.Ann. N. Y. Acad. Set 64,1009.
Allison, J.P. and Lanier, L.L. (1985). Identification of antigen receptor-associated
structures on murine T cells. Nature 314,107-109.
Alonso, S., Minty, A., Bourlet, Y., and Buckingham, M. (1986). Comparison of
three actin-coding sequences in the mouse: Evolutionary relationships between the
actin genes of warm-blooded vertebrates./. Mol Evol 23,11-22.
Baenziger, J., Hengartner, H., Zinkernagel, R.M., and Groscurth, P. (1988).
Distinct patterns of virus specific T cell mediated cytolysis of transformed vs.
primary target cells. Scand. J. Immunol 28,411-423.
Beck, E., Ludwig, G., Auserwald, E.A., Reiss, B., and Schaller, H. (1982).
Nucleotide sequence and exact localization of the neomycin phosphotransferase
gene from transposon Tn5. Gene 19,327-336.
Berke, G. (1983). Cytotoxic T lymphocytes. How do they function. Immunol Rev.
72, 5-42.
Berke, G. (1991). T cell mediated cytotoxicity. Curr. Opin. Immunol 3,320-325.
Berke, G. and Rosen, D. (1987). Are lytic granules and perforin 1 thereof involved
in lysis induced by in vivo primed peritoneal exudate CTL? Transplant. Proc 19,
412-416.
Berke, G. and Rosen, D. (1988). Highly lytic in vivo primed cytolytic T
lymphocytes devoid of lytic granules and BLT-esterase activity acquire these
constituents in the presence of T cell growth factors upon blast transformation in
vitro. /Immunol 141,1429-1436.
Biggers, J.D., Whitten, W.K., and Whittingham, D.G. (1971). In Methods in
Mammalian Embryology, J.C. Daniel, ed. (San Francisco: Freeman), pp. 86-116.
Bjorkman, PJ., Saper, M.A., Samraoui, B., Bennett, W.S., Strominger, J.L., and
Wiley, D.C. (1987a). Structure of the human class I histocompatibility antigen,
HLA-A2. Nature 329, 506-511.
115
Bjorkman, PJ., Saper, MA, Samraoui, B., Bennett, W.S., Strominger, J.L., and
Wiley, D.C. (1987b). The foreign antigen binding site and T cell recognition
regions of class I histocompatibility antigens. Nature 329,512-518.
Blanden, R.V. (1974). T cell response to viral and bacterial infection. Transplant.
Rev. 19,56-84.
Bradley, A., Evans, M., Kaufman, M.H., and Robertson, E. (1984). Formation of
germ-line chimaeras from embryo derived teratocarcinoma cell lines. Nature 309,
255-256.
Brunet, J.F., Denizot, F., Suzan, M., Haas, W., Mencia Huerta, J.M., Berke, G.,
Luciani, M.F., and Golstein, P. (1987). CTLA-1 and CTLA-3 serine esterase
transcripts are detected mostly in cytotoxic T cells, but not only and not always. /.
Immunol 138,4102-4105.
Bykovskaja, S.N., Rytenko, A.N., Rauschenbach, M.O., and Bykovsky, A.F.
(1978a). Ultrastnicural alteration of cytolytic T lymphocytes following their
interaction with target cells. I. Hypertrophy and change of orientation of the Golgi
aparatus. Cell Immunol 40,164.
Bykovskaja, S.N., Rytenko, AN., Rauschenbach, M.O., and Bykovsky, A.F.
(1978b). Ultrastructural alteration of cytolytic T lymphocytes following their
interaction with target cells. II. Morphogenesis of secretory granules and
intracellular vacuoles. Cell Immunol 40,175.
Cantor, H. and Boyse, E.A. (1975). Functional subclasses of T lymphocytes bearingdifferent Ly antigens I. The generation of functionally distinct T-cell subclasses is a
differentiative process independent of antigen./. Exp. Med. 141,1376-1389.
Capecchi, M.R. (1989a). Altering the genome by homologous recombination.
Science 244,1288-1292.
Capecchi, M.R. (1989b). The new mouse genetics: Altering the genome by gene
targeting. Trends in Genetics 5,70-76.
Cerottini, J.C. and Brunner, K.T. (1974). Cell-mediated cytotoxicity, allograft
rejection and tumor immunity. Adv. Immunol 18, 67-132.
Cerrotini, J.C, Nordin, A.A., and Brunner, K.T. (1970). Specific in vitro
cytotoxicity of thymus derived lymphocytes sensitized to alloantigens. Nature 228,
1308-1309.
Chang, T.W. and Eisen, H.N. (1980). Effects of N*-tosyl-L-lysyl-
chloromethylketone on the activity of cytotoxic T lymphocytes. /. Immunol 124,
1028-1033.
116
Chirgwin, T.M., Przybyla, AE., MacDonald, RJ., and Rutter, WJ. (1979).Isolation of biologically active ribonucleic acid from sources enriched in
ribonuclease. Biochemistry 18,5294-5299.
Chisaka, O., Musci, T.S., and Capecchi, M.R. (1992). Developmental defects of the
ear, cranial nerves and hindbrain resulting from the targeted disruption of the
mouse homeobox gene Hox-1.6. Nature 355,516-520.
Clark, W., Ostergaard, H., Gorman, K., and Torbett, B. (1988). Molecular
mechanisms of CTL-mediated lysis: A cellular perspective. Immunol Rev. 103, 37-
51.
Clark, W.R. (1988). Perforin - a primary or auxiliary lytic mechanism. Immunol
Today 9,101-104.
Cole, GJl., Gilden, D.H., Monjan, AA., and Nathanson, N. (1971). Lymphocytic
choriomeningitis virus: Pathogenesis of acute central nervous system disease.
Federation Proa 30,1831-1841.
DeChiara, T.M., Efstratiadis, A., and Robertson, EJ. (1990). A growth-deficiency
phenotype in heterozygous mice carrying an insulin-like growth factor II gene
disrupted by targeting. Nature 345,78-80.
Dennert, G., Anderson, C.G., and Prochazka, G. (1987). High activity of N*-
benzyloxycarbonyl-L-lysine thiobenzyl ester serine esterase and cytolytic perforin
in cloned cell lines is not demonstrable in in vivo-induced cytotoxic effector cells.
Proc Natl Acad. Set USA 84,5004-5008.
Dennert, G. and Podack, E.R. (1983). Cytolysis by H-2-specific T killer cells. /.
Exp. Med. 157,1483-1495.
Di Virgilio, F., Pizzo, P., Zanovello, P., Bronte, V., and Collavo, D. (1990).
Extracellular ATP as a possible mediator of cell-mediated cytotoxicity. Immunol
Today 11,274-276.
Doetschmann, T., Gregg, R.G., Maeda, N., Hooper, ML., Melton, D.W.,
Thompson, S., and Smithies, O. (1987). Targeted correction of a mutant HPRT
gene in mouse embryonic stem cells. Nature 330,576-578.
Doherty, P.C. and Zinkernagel, R.M. (1974). T cell-mediated immunopathology in
viral infections. Transplant. Rev. 19, 89-120.
Donehower, L.A., Harvey, M., Slagle, B.L., McArthur, MJ., Montgomery, Ch.A.,
Butel, J.S., and Bradley, A (1992). Mice deficient for p53 are developmentally
normal but susceptible to spontaneaous tumours. Nature 356, 215-221.
117
Dourmashkin, R.R., Deteix, P., Simone, C.B., and Henkart, P. (1980). Electron
microscopic demonstration of lesions on target cells membranes associated with
antibody-dependent cellular cytotoxicity. Clin. Exp. Immunol 42,554.
Evans, MJ. and Kaufman, M.H. (1981). Establishment in culture of pluripotential
cells from mouse embryos. Nature 292,154-156.
Falk, K., Rotzschke, O., Stevanovic, S., Jung, G., and Rammensee, H.-G. (1991).
Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC
molecules. Nature 351,290-296.
Farrar, JJ., Fuller-Farrar, J., Simon, P., Hilfiker, M.L., Stadler, B.M., and Farrar,
W.L. (1980). Thymoma production of T cell growth factor (Interleukin 2). /.
Immunol 125,2555-2558.
Fujimura, F.K., Deininger, P.L., Friedmann, T., and Linney, E. (1992). Mutation
near the polyoma DNA replication origin permits productive infection of F9
embryonal carcinoma cells. Cell 23,809-814.
Fung-Leung, W-P. and Mak, T.W. (1992). Embryonic stem cells and homologous
recombination. Curr Opinion in Immunol 4,189-194.
Garcia-Sanz, J.A., Plaetnick, G., Velotti, F., Masson, D., Tschopp, J., MacDonald,
H.R., and Nabholz, M. (1987). Perforin is present only in normal activated Lyt2+
lymphocytes and not in L3T4+ cells, but the serine peotease granzyme A is made
by both subsets. EMBOJ. 6, 933-938.
Garrett, T.PJ., Saper, M.A., Bjorkman, P.J., Strominger, J.L., and Wiley, D.C.
(1989). Specificity pockets for the side chains of peptide antigens in HLA-Aw68.
Nature 342, 692-696.
Germain, R.N. (1986). The ins and outs of antigen processing and presentation.
Nature 322,687-689.
Gershenfeld, H.K. and Weissman, I.L. (1986). Cloning of a cDNA for a T-cell-
specific serine protease from a cytotoxic T lymphocyte. Science 232,854-858.
Golstein, P. (1987). Cytolytic T-cell melodrama. Nature 327,12.
Gossler, A., Doetschman, T., Korn, R., Serfling, E., and Kemler, R. (1986).
Transgenesis by means of blastocyst-derived embryonic stem cell lines. Proc Natl
Acad. Set USA 83,9065-9069.
Govaertz, A. (1960). Cellular antibodies in kidney homotransplantation. /.
Immunol 85, 516.
118
Granger, G.A. and Kolb, W.P. (1968). Lymphocyte in vitro cytotoxicity:mechanisms of immune and nonimmune small lymphocyte mediated target L cell
destruction. J Immunol 101, 111.
Griffiths, G.M., Namikawa, R., Mueller, C, Liu, C.-G, Young, J.D.-E., Billingham,M., and Weissmann, I. (1991). Granzyme A and perforin as markers for rejectionin cardiac transplantation. Eur. J. Immunol 21,687-692.
Griffiths, G.M., Alpert, S., Lambert, E., McGuire, J., and Weissman, I.L. (1992).Perforin and granzyme A expression identifying cytolytic lymphocytes in
rheumatoid arthritis. Proc Natl Acad. Sci USA 89,549-553.
Groscurth, P., Qiao, B.Y., Podack, E.R., and Hengartner, H. (1987). Cellular
localization of perforin 1 in murine cloned cytotoxic T lymphocytes. /. Immunol
138,2749-2752.
Gubler, U. and Hoffman, BJ. (1983). A simple and very efficient method for
generating cDNA libraries. Gene 25,263.
Hasty, P., Ramirez-Sou's, R., Krumlauf, R., and Bradley, A. (1991). Introduction of
a subtle mutation into the Hox-2.6 locus in embryonic stem cells. Nature 350, 243-
246.
Hasty, P., Rivera-Perez, J., and Bradley, A. (1992). The Length of homology
required for gene targeting in embryonic stem cells. Mol Cell Biol 11,5586-5591.
Hayes, M.P., Berrebi, G.A., and Henkart, P.A. (1989). Induction of target cell
DNA release by the cytotoxic T lymphocyte granule protease granzyme A. /. Exp.Med. 170,933-946.
Hedrick, S.M., Cohen, D.I., Nielsen, E.A., and Davis, M.M. (1984). Isolation of
cDNA clones encoding T cell-specific membrane-associated proteins. Nature 308,
149-158.
Henkart, P.A., Millard, PJ., Reynolds, C.W., and Henkart, M.P. (1984). Cytolytic
activity of purified cytoplasmic granules from cytotoxic rat large granular
lymphocyte tumors./. Exp. Med 160,75-93.
Henkart, P.A. (1985). Mechanism of lymphocyte-mediated cytotoxicity. Ann. Rev.
Immunol 3,31-58.
Herberman, R.B., Nunn, M.E., and Lavrin, D.H. (1975). Natural cytotoxic
reactivity of mouse lymphoid cells against syngeneic and allogeneic tumors. I.
Distribution of reactivity and specificity. Int. J. Cancer 16,216-239.
119
Hogan, B., Constantini, F., and Lacy, E. (1986). Manipulating the mouse embryo:
A laboratory manual. (Cold Spring Harbor, NY: Cold Spring Harbor Labratory).
Hogan, B. and Lyons, K. (1988). Getting nearer the mark. Nature 336,304-305.
Hooper, M., Hardy, K., Handyside, A., Huster, S., and Monk, M. (1987). HPRT-
deficient (Lesch-Nyhan) mouse embryos derived from germline colonization by
cultured cells. Nature 326,292-295.
Hotchin, J. (1971). Persistent and slow virus infections. Monogr. Virol 3,1-211.
Hudig, D., Haverty, T., Fulcher, C, Redelmann, D., and Mendelssohn, J. (1981).
Inhibition of human natural cytotoxicity by macromolecular antiproteases. /.
Immunol 126,1569-1574.
Jardetzky, T.S., Lane, W.S., Robinson, RA, Madden, D.R., and Wiley, D.C.
(1991). Identification of self peptides bound to purified HLA-B27. Nature 353,326-
329.
Jenne, D., Rey, C, Haefliger, J.A., Qiao, B.Y., Groscurth, P., and Tschopp, J.
(1988). Identification and sequencing of cDNA clones encoding the granule-
associated serine proteases granzymes D, E, and F of cytolytic T lymphocytes. Proc
Natl Acad. Scl USA 85,4814-4818.
Jenne, D.E. and Tschopp, J. (1988). Granzymes, a family of serine proteases
released from granules of cytolytic T lymphocytes upon T cell receptor stimulation.
Immunol Rev. 103,53-71.
Johnson, R.,S., Sheng, M., Greenberg, M.E., Kolodner, R.D., Papaiannou, V.E.,
and Spiegelmann, B.M. (1989). Targeting of nonexpressed genes in embryonic
stem cells via homologous recombination. Science 245,1234-1236.
Joyner, A.L, Skames, W.C, and Rossant, J. (1989). Production of a mutation in
mouse En-2 gene by homologous recombination in embryonic stem cells. Nature
338,153-156.
Kawasaki, A., Shinkai, Y., Yagita, H., and Okumura, K. (1992). Expression of
perforin in murine natural killer cells and cytotoxic T lymphocytes in vivo. Eur. J.
Immunol 22,1215-1219.
Kim, H.-S. and Smithies, O. (1992). Recombinant fragment assay for gene
targeting based on the polymerase chain reaction. Nucleic Acids Res. 16,8887-8903.
Koller, B.H. and Smithies, O. (1989). Inactivating the B2-microglobulin locus in
mouse embryonic stem cells by homologous recombination. Proc. Natl Acad. Scl
USA 86, 8932-8935.
120
Lachmann, PJ. (1983). Are complement lysis and lymphotoxicity analogous?
Nature 305,473.
Lancki, D.W., Hsieh, C-S., and Fitch, F.W. (1991). Mechanisms of lysis by cytotoxic
T lymphocyte clones./. Immunol 146,3242-3249.
Landsteiner, K. and Chase, M.W. (1942). Experiments on transfer of cutaneous
sensitivity to simple compounds. Proc Soc. Exp. Biol Med. 49,688-690.
Ledermann, B. (1991). Isolierung, Charakterisierung und Transfektion muriner
embryonaler Stammzellinien des C57B1/6 Inzuchtstammes. Dissertation an der
Universitat Basel. (
Ledermann, B. and Burki, K. (1991). Establishment of a germ-line competent
C57B1/6 embryonic stem cell line.Exp. Cell Res. 197,254-258.
Lehmann-Grube, F. (1971). Lymphocytic choriomeningitis virus. Virol Monogr. 10,
1-173.
Lehmann-Grube, F., Assmann, U., Loliger, C, Moskophidis, D., and Lohler, J.
(1985). Mechanism of recovery from acute virus infection. I. Role of T lymphocytes
in the clearance of lymphocytic choriomeningitis virus from spleens of mice. /.
Immunol 134, 608-615.
Lehmann-Grube, F. and Lohler, J. (1988). Immunopathologic alteration of
lymphatic tissues of mice infected with lymphocytic choriomeningitis virus II.
Pathogenetic mechanism. Lab. Invest. 44,205-213.
Lichtenheld, M., Lu, P., Olsen, KJ., Lowrey, D.M., Hameed, A., Hengartner, H.,
and Podack, E.R. (1988). Structure and Function of human perforin. Nature 335,
448-451.
Lichtenheld, M.G. and Podack, E.R. (1989). Structure of the human perforin gene:
A simple gene organization with interesting potential regulatory sequences. /.
Immunol 143,42674274.
Liu, C.C., Perussia, B., Conn, Z.A., and Young, J.D.E. (1986). Identification and
characterization of a pore-forming protein of human peripheral blood natural
killer cells. /. Exp. Med. 164,2061-2076.
Liu, C.C., Steffen, M., King, F., and Young, J.D. (1987). Identification, isolation,
and characterization of a novel cytotoxin in murine cytolytic lymphocytes. Cell 51,
393-403.
Lobe, C.G., Finlay, B.B., Paranchych, W., Paetkau, V.H., and Bleackley, R.C.
(1986). Novel serine proteases encoded by two cytotoxic T lymphocyte-specific
genes. Science 232,858-861.
121
Lowrey, D.M., Aebischer, T., Olsen, KJ., Lichtenheld, M., Rupp, F., Hengartner,H., and Podack, E.R. (1989). Cloning analysis and expression of murine perforin 1
(PI) cDNA, a component of cytolytic T-cell granules with homology to
complement component C9. Proc. Natl Acad. Set USA 86,247-251.
Madden, D.R., Gorga, J.C., Strominger, J.L., and Wiley, D.C. (1991). The structure
of HLA-B27 reveals nonamer self-peptides bound in an extended conformation.
Nature 353,321-325.
Mansour, S.L., Thomas, K.R., and Capecchi, M.R. (1988). Disruption of the proto-
oncogene int-2 in mouse embryo-derived stem cells: A general strategy for
targeting mutations to non-selectable genes. Nature 336,348-352.
Marker, O. and Volkert, M. (1973). Studies on cell-mediated immunity to
lymphocytic choriomeningitis virus in mice./. Exp. Med. 137,1511-1525.
Martz, E. (1976). Early steps in specific tumor cell lysis by sensitized mouse T
lymphocytes: n. Electrolyte permeability increase in the target cell membrane
concomitant with programming for lysis./. Immunol 111, 1023-1027.
Martz, E. and Howell, D.M. (1989). CTL: Are they virus control cells first, and
cytolytic cells second? DNA fragmentation, apoptosis, and the prelytic halt
hypothesis. Immunol Today 10,70-80.
Masson, D., Nabholz, M., Estrade, C, and Tschopp, J. (1986). Granules of cytolytic
T-lymphocytes contain two serine esterases. EMBOJ. 5,1595-1600.
Masson, D. and Tschopp, J. (1985). Isolation of a lytic, pore-forming protein
(perforin) from cytolytic T-lymphocytes./. Biol Chem. 260, 9069-9072.
Masson, D. and Tschopp, J. (1987). A family of serine esterases in lytic granules of
cytolytic T lymphocytes. Cell 49,679-685.
Matteucci, M.D. and Caruthers, M.H. (1981). Synthesis of deoxy-oligo nucleotides
on a solid polymer support. Journal of the American Chemical Society 103, 3185-
3191.
Mclntyre, K.W. and Welsh, R.M. .(1986). Accumulation of natural killer and
cytotoxic T large granular lymphocytes in the liver during virus infection. /. Exp.Med. 164,1667-1681.
McKnight, S. (1992). The nucleotide sequence and transcript map of the herpes
simplex virus thymidine kinase gene. Nucleic Acids Res. 8,5949-5964.
McMahon, AP. and Bradley, A. (1990). The Wnt-1 (int-1) proto-oncogene is
required for development of a large region of the mouse brain. Cell 62,1073-1085.
122
Meyer, P.R., Yanagihara, E.T., Parker, J.W., and Lukes, RJ. (1984). A distinctive
follicular hyperplasia in AIDS and the AIDS-related-complex. Hematol Oncol 2,
319-347.
Miller, J.F.A.P. (1961). Immunological function of the thymus. Lancet 2,748-749.
Mitchison, N.A. (1954). Passive transfer of transplantation immunity. Proc. R. Soc
Lond. Biol 142,72.
Mombaerts, P., Iacomini, J., Johnson, R.S., Herrup, K., Tonegawa, S., and
Papaioannou, V.E. (1992). RAG-1-deficient mice have no mature B and T
lymphocytes. Cell 68,869-877.
Morrison, L*A., Lukacher, A.E., Braciale, V.L., Fan, D.P., and Braciale, TJ.
(1986). Differences in antigen presentation to MHC class I- and class H-restricted
influenza virus specific cytolytic T lymphocyte clones./. Exp. Med. 163,903-921.
Moller, G. (1988). Molecular mechanisms of T cell mediated lysis.
Immunol Rev., Vol.103
Mueller, C, Gershenfeld, H.K., Lobe, C.G., Okada, C.Y., Bleackley, R.C., and
Weissman, I.L. (1988). A high proportion of T-lymphocytes that infiltrate H-2
incompatible heart allografts in vivo express genes encoding cytotoxic cell specific
serine proteases, but do not express the MEL-14 defined lymph node homing
receptor./. Exp. Med. 167,1124-1136.
Miiller, Ch., Kagi, D., Aebischer, T., Odermatt, B., Held, W., Podack, E.R..,
Zinkernagel, R.M., and Hengartner, H. (1989). Detection of perforin and
granzyme A mRNAs in infiltrating cells during LCMV infection of mice. Eur. J.
Immunol 19,1253-1259.
Nagler Anderson, G, Lichtenheld, M., Eisen, H.N., and Podack, E.R. (1989).
Perforin mRNA in primary peritoneal exudate cytotoxic T lymphocytes. /.
Immunol 143, 3440-3443.
Nakata, M., Kawasaki, A., Azuma, M., Tsuji, K., Matsuda, H., Shinkai, Y., Yagita,
H., and Okumura, K. (1992). Expression of perforin and cytolytic potential of
human peripheral blood lymphocyte subpopulations. Int. Immunol 4,1049-1054.
Nickoloff, J.A. and Reynolds, RJ. (1990). Transcription stimulates homologous
recombination in mammalian cells. Mol Cell Biol 10,4837-4845.
Ojcius, D.M., Persechini, P.M., Zheng, L.M., Notaroberto, P.C., Adeodato, S.C.,
and Young, J.D. (1991). Cytolytic and ion channel-forming properties of the N-
terminus of lymphocyte perforin. Proc. Natl Acad. Set USA 88,4621-4625.
Ostergaard, H. and Clark, W.R. (1987). The role of Ca2+ in activation of mature
cytotoxic T lymphocytes for lysis. /. Immunol 139,3573-3579.
123
Ostergaard, H.L., Kane, K.P., Mescher, M.F., and Clark, W.R. (1987). Cytotoxic T
lymphocyte mediated lysis without release of serine esterase. Nature 330,71-72.
Pasternack, M.S., Verret, C.R., Liu, MA, and Eisen, H. (1986). Serine esterase in
cytolytic T lymphocytes. Nature 322,740-743.
Pasternack, M.S. and Eisen, H. (1985). A novel serine esterase expressed by
cytotoxic T lymphocytes. Nature 314,743-745.
Peters, PJ., Geuze, HJ., Van der, Donk.H.A., Slot, J.W., Griffith, J.M., Stam, NJ.,
Clevers, H.C., and Borst, J. (1989). Molecules relevant for T cell-target cell
interaction are present in cytolytic granules of human T lymphocytes. Eur. J.
Immunol 19,1469-1475.
Pierres, M., Goridis, C, and Golstein, P. (1982). Inhibition of murine T cell-
mediated cytolysis and T cell proliferation by a rat monoclonal antibody
immunoprecipitating two lymphoid cell surface polypeptides of 94kD and 180kD
molecular weight. Eur. J. Immunol 12,60-69.
Podack, E.R. (1983). Assembly of two types of tubules with putative cytolytic
function by cloned natural killer cells. Nature 302,442-445.
Podack, E.R. (1985). The molecular mechanism of lymphocyte-mediated tumor
cell lysis. Immunol Today 6,21-27.
Podack, E.R., Ding-E Young, J., and Cohn, Z.A. (1985). Isolation and biochemical
and functional characterization of perforin 1 from cytolytic T-cell granules. Proc
Natl Acad. Set USA 82,8629-8633.
Podack, E.R., Lowrey, D.M., Lichtenheld, M., Olsen, KJ., Aebischer, T., Binder,
D., Rupp, F., and Hengartner, H. (1988). Structure, function and expression of
murine and human perforin 1 (PI). Immunol Rev. 103, 203-211.
Podack, E.R., Hengartner, H., and Lichtenheld, M.G. (1991). A central role of
perforin in cytolysis. Annu. Rev. Immunol 9,129-157.
Podack, E.R. and Konigsberg, P.J. (1984). Cytolytic T cell granules: Isolation,
structural, biochemical, and functional characterization./. Exp. Med. 160,695-710.
Potter, H., Weir, L., and Leder, P. (1984). Enhancer-dependent expression of
human kappa immunoglobulin. Proc. Natl Acad. Set USA 81,7161-7165.
Redelman, D. and Hudig, D. (1980). The mechanism of cell-mediated cytotoxicity.
I. Killing by murine cytotoxic T lymphocytes requires cell surface thiols and
activated proteases. /. Immunol 124,870-878.
Robertson, E.J. (1986). Pluripotential stem cell lines as a route into the mouse
germ line. Trends in Genetics Januar, 9-13.
124
Rossant, J. and Joyner, A.L. (1989). Towards a molecular-genetic analysis of
mammalian development. Trends in Genetics 5,277-283.
Ruddle, N.H. and Waksman, B.H. (1968). Cytotoxicity mediated by soluble antigenand lymphocytes in delayed hypersensitivity UJ. Analysis of mechanism. /. Exp.Med. 128,1267.
Russell, J.H. (1983). Internal disintegration model of cytotoxic lymphocyte-induced
target damage. Immunol Rev. 72,97-118.
Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, SJ., Higuchi, R., Horn, G.T., Mullis,
K.B., and Erlich, HA (1988). Primer-directed enzymatic amplification of DNA
with a thermostable DNA polymerase. Science 239,487-491.
Schwartzberg, P.L., Robertson, E.J., and Goff, S.P. (1990). Targeted disruption of
the endogenous c-abl locus by homologous recombination with DNA encoding a
selectable fusion protein. Proc. Natl Acad. Set USA 87,3210-3214.
Sedivy, J.M. and Sharp, P.A. (1989). Positive genetic selection for gene disruptionin mammalian cells by homologous recombination. Proc. Nail Acad. Sci USA 86,
227-231.
Shi, L., Kraut, R.P., Aebersold, R., and Greenberg, A.H. (1992). A natural killer
cell granule protein that induces DNA fragmentation and apoptosis. /. Exp. Med.
175,553-566.
Shinkai, Y., Takio, K., and Okumura, K. (1988). Homology of perforin to the ninth
component of complement (C9). Nature 334,525-527.
Shinkai, Y., Yoshida, M.C., Maeda, K., Kobata, T., Maruyama, K., Yodoi, J.,
Yagita, H., and Okumura, K. (1989). Molecular cloning and chromosomal
assignment of a human perforin (PFP) gene. Immunogenetics 30,452-457.
Shiver, J.W., Su, L., and Henkart, P.A. (1992). Cytotoxicity with target DNA
breakdown by rat basophilic leukemia cells expressing both cytolysin and granzyme
A. Cell 71,315-321.
Shiver, J.W. and Henkart, P.A. (1991). A noncytotoxic mast cell tumor line exhibits
potent IgE-dependent cytotoxicity after transfection with the cytolysin perforin
gene. Cell 64,1175-1181.
Sim, G.K., Yague, J., Nelson, J., Marrack, P., Palmer, E., Augustin, A., and
Kappler, J.W. (1984). Primary structure of human T-cell receptor a-chain. Nature
312, 771-775.
Simon, M.M., Hoschiitzky, H., Fruth, U., Simon, H.G., and Kramer, M.D. (1986).
Purification and characterization of a T-cell specific serine proteinase (TSP-1)
from cloned cytolytic T lymphocytes. EMBOJ. 5, 3267-3274.
125
Smithies, O., Gregg, R.G., Boggs, S.S., Koralewski, M.A., and Kucherlapati, R.S.
(1985). Insertion of DNA sequences into the human chromosomal beta-globinlocus by homologous recombination. Nature 317,230-234.
Smyth, MJ., Ortaldo, J.R., Bere, W., Yagita, H., Okumura, K., and Young, H.A.
(1990a). IL-2 and IL-6 synergize to augment the pore-forming protein gene
expression and cytotoxic potential of human peripheral blood T cells. /. Immunol
145,1159-1166.
Smyth, MJ., Ortaldo, J.R., Shinkai, Y.-I., Yagita, H., Nakata, M., Okumura, K.,
and Young, HA (1990b). Interleukin 2 induction of pore-forming protein gene
expression in human peripheral blood CD8+ T cells./. Exp. Med. 171,1269-1281.
Sprent, J. (1978). Role of H-2 gene products in the function of T helper cells from
normal and chimeric mice in vitro. Immunol Rev. 42,108-142.
Springer, T.A. (1990). Adhesion receptors of the immune system. Nature 346, 425-
433.
Stewart, C.L., Schuetze, S., Vanek, M., and Wagner, E.F. (1987). Expression of
retroviral vectors in transgenic mice obtained by embryo infection. EMBO J. 6,
383-388.
Taylor, M.K. and Cohen, JJ. (1992). Cell mediated cytotoxicity. Curr Opinion in
Immunol 4,338-343.
Thomas, K.R. and Capecchi, M.R. (1987). Site directed mutagenesis by gene
targeting in mouse embryo-derived stem cells. Cell 51,503-512.
Thomas, K.R. and Capecchi, M.R. (1990). Targeted disruption of the murine Int-1
proto-oncogene resulting in severe abnormalities in midbrain and cerebellar
development. Nature 346,847-850.
Thompson, S., Clarke, A.R,, Pow, A.M., Hooper, M.L., and Melton, D.W. (1989).
Germ line transmission and expression of a corrected HPRT gene produced by
gene targeting in embryonic stem cells. Cell 56, 313-321.
Thorn, R. and Henney, C. (1988). Studies on the mechanism of lymphocyte
mediated cytolysis./. Immunol 116,146-149.
Tian, Q., Streuli, M., Saito, H., Schlossman, S.F., and Anderson, P. (1991). A
polyadenylate binding protein localized to the granules of cytolytic lymphocytes
induces DNA fragmentation in target cells. Cell 67, 629-639.
126
Tibulewicz, V.LJ., Tremblay, M.L., LaMarca, M.E., Willemsen, R., Stubblefield,
B.K., Winfield, S., Zablocka, B., Sidransky, E., Martin, B.M., Huang, S.P., Mintzer,
KA., Westphal, H., Mulligan, R.C., and Ginns, E.I. (1992). Animal model of
Gaucher's disease from targeted disruption of the mouse glucosecerebrosidase
gene. Nature 357,407-410.
Timonen, T., Ortaldo, J.R., and Herberman, R.B. (1988). Characteristics of human
large granular lymphocytes and relationship to natural killer and K cells. /. Exp.
Med. 153,569-582.
Townsend, A.R.M., Rothbard, J., Gotch, F.M., Bahadur, G., Wraith, D., and
McMichael, AJ. (1986). The epitopes of influenza nucleoprotein recognized by
cytotoxic T lymphocytes can be defined with short synthetic peptides. Cell 44, 959-
968.
Trapani, JA., Kwon, B.Y., Kozak, CA., Chintamaneni, G, Young, J.D.-E., and
Dupont, B. (1989). Genomic organization of the mouse pore-forming protein
(perforin) gene and localization to chromosome 10: Similarities to and differences
from C9.7. Exp. Med. Ill, 545-557.
Traub, E. (1936). Persistence of lymphocytic choriomeningitis virus in immune
animals and its relation to immunity./. Exp. Med. 63,847-861.
Trenn, G., Takayama, H., and Sitkovsky, M.V. (1987). Exocytosis of cytolytic
granules may not be required for target cell lysis by cytotoxic T-lymphocytes.
Nature 330,72-74.
Tschopp, J. and Nabholz, M. (1990). Perforin-mediated target cell lysis by cytotoxic
T lymphocytes. Ann. Rev. Immunol 8,279-302.
Ucker, D.S. (1987). Cytotoxic T lymphocytes and glucocorticoids activate an
endogenous suicide process in target cells. Nature 327,62-64.
Versteeg, R. (1992). NK cells and T cells: mirror images? Immunol Today 13, 244-
247.
Weber, S., Traunecker, A., Oliveri, R, Gerhard, W., and Karjalainen, K. (1992).
Specific low-affinity recognition of major histocompatibility complex plus peptide
by soluble T-cell receptor. Nature 356,793-796.
Yenofsky, R.L., Fine, M., and Pellow, J.W. (1990). A mutant neomycin
phosphotransferase II gene reduces the resistance of transformants to antibiotic
selection pressure. Proc. Natl Acad. Sci USA 87,3435-3439.
Yoshikai, Y., Anatoniou, D., Clark, S.P., Yanagi, Y., Sangster, R., Van den Elsen,
P., Terhorst, T, and Mak, T.W. (1984). Sequence and expression of transcripts of
the human T-cell receptor 6-chain genes. Nature 312,521-524.
127
Youn, B.S., Liu, C.C., Kim, K.K., Young, J.D.E., Kwon, M.H., and Kwon, B.S.
(1991). Structure of the mouse pore-forming protein (perforin) gene: Analysis of
transcription initiation site, 5' flanking sequence, and alternative splicing of 5'
untranslated regions./. Exp. Med. 173,813-822.
Young, J.D., Damiano, A., DiNome, M.A., Leong, L.G., and Cohn, Z.A. (1987).
Dissociation of membrane binding and lytic activities of the lymphocyte pore-
forming protein (perforin)./. Exp. Med. 165,1371-1382.
Young, J.D., Liu, C.C., Persechini, P.M., and Cohn, Z.A. (1988). Perforin-
dependent and -independent pathways of cytotoxicity mediated by lymphocytes.
Immunol Rev. 103,161-202.
Young, J.D.-E., Clark, W.R., Liu, C.-G, and Cohn, Z.A. (1987). A calcium- and
perforin-independent pathway of killing mediated by murine cytolytic lymphocytes.
/. Exp. Med. 166,1894-1899.
Young, J.D.-E. and Liu, C.-C. (1988). Multiple mechanisms of lymphocyte-
mediated killing. Immunol Today 9,140-144.
Young, J.D.E., Cohn, Z.A., and Podack, E.R. (1986a). The ninth component of
complement and the pore-forming protein (perforin 1) from cytotoxic T cells:
Structural, immunological, and functional similarities. Science 233,184-190.
Young, J.D.E., Hengartner, H., Podack, E.R., and Cohn, Z.A. (1986b). Purification
and characterization of a cytolytic pore-forming protein from granules of cloned
lymphocytes with natural killer activity. Cell 44, 849-859.
Young, J.D.E., Nathan, C.F., Podack, E.R., Pelladino, MA, and Cohn, Z.A.
(1986c). Functional channel formation associated with cytotoxic T-cell granules.
Proc. Natl Acad. Set USA 83,150-154.
Young, J.D.E., Podack, E.R., and Cohn, Z.A. (1986d). Properties of a purified
pore-forming protein (perforin 1) isolated from H-2-restricted cytotoxic T cell
granules./. Exp. Med. 164,144-155.
Young, L.H., Peterson, L.B., Wicker, L.S., Persechini, P.M., and Young, J.D.
(1989). In vivo expression of perforin by CD8+ lymphocytes in autoimmune
disease. Studies on spontaneous and adoptively transferred diabetes in nonobese
diabetic mice. /. Immunol 143,3994-3999.
Young, L.H., Joag, S.V., Zheng, L.M., Lee, C.P., Lee, Y.S., and Young, J.D. (1990).
Perforin-mediated myocardial damage in acute myocarditis. Lancet 336, 1019-
1021.
128
Yue, CC, Reynolds, CW., and Henkart, P.A. (1990). Inhibition of cytolysin
activity in large granular lymphocyte granules by lipids: Evidence for a membrane
insertion mechanism of lysis. Mol Immunol 24,647-653.
Zheng, L.M., Liu, CC, Ojcius, D.M., and Young, J.D. (1991). Expression of
lymphocyte perforin in the mouse uterus during pregnancy./. Cell Set 99,317-323.
Zijlstra, M., Bix, M., Simister, N.E., Loring, J.M., Raulet, D.H., and Jaenisch, R.
(1990). Beta 2-microglobulin deficient mice lack CD4-8+ cytolytic T cells (see
comments). Nature 344,742-746.
Zink, T.M., Zimmer, M., Weitz, S., Tschopp, J., Jenne, D.E., and Lichter, P.
(1992). Human perforin (PRF1) maps to 10q22, a region that is syntenic with
mouse chromosome 10. Genomics 13,1300-1302.
Zinkernagel, R.M., Haenseler, E., Leist, T.P., Cerny, A., Hengartner, H., and
Althage, A. (1986). T cell mediated hepatitis in mice infected with lymphocytic
choriomeningitis virus./. Exp. Med. 164,1075-1092.
Zinkernagel, R.M. and Doherty, P.C (1974). Restriction of in vitro T cell
mediated cytotoxicity in lymphocytic choriomeningitis within a syngeneic or
semiallogeneic system. Nature 248,701-702.
129
Folgende Personen haben alle in der einen oder anderen Art ganz entscheidenden
Anteil am Gelingen dieser Arbeit gehabt. Danken mdchte ich deshalb:
Hans Hengartner fiir die Betreuung meiner Arbeit und dafiir, dass er
immer ein offenes Ohr fiir alle Sorgen und Probleme
hatte.
Rolf Zinkernagel fiir die intellektuellen Anregungen und die
Moglichkeit diese Arbeit in seiner Arbeitsgruppe
durchfiihren zu konnen.
Birgit Ledermann und Kurt Biirki fiir die grosse Arbeit, die sie in die Arbeit mit
der BL/6-III Zellinie und die Blastocysteninjektion
investiert haben, und die langfristige Zusammenarbeit
auch in Zeiten, in denen nicht alles rund lief.
Peter Seiler fur die Herstellung der Rekombinationskonstrukte
wahrend seiner Diplomarbeit.
Alana Althage und Rosmarie Lang fiir die tatkraftige Hilfe und die geduldige
Einweihung in die praktischen Aspekte der zellularen
Immunologie.
Allen Mitarbeitera des Institutes fur die unzahligen Gesprache nicht nur fachiicher
Art und die kollegiale Arbeitsatmosphare.
Meinen Eltern fur ihre stete Unterstiitzung.
Curriculum vitae
Personalien
Name: Kagi
Vorname: David
Titel Dipl.sc.nat.
Adresse G: Institut fur experimentelle Immunologie,
Umversitatsspital, 8091 Zurich
Adresse P: Winterthurerstrasse 641,8051 Zurich
Heimatorte: Zurich, Wallisellen, Baretswil
Zivilstand: Ledig
Schqlbildung
1970 -1976 Primarschule in Hinteregg/ZH
1976 -1982 Realgymnasium Ramibuhl, Zurich, Abschluss mit der
eidgenossischen Matura Typus B, (Humanistisches
Gymnasium)
Studium
1983-1988
1986
1986-1987
1987- 1988
1988
Studium an der Abteilung fur Naturwissenschaften, ETH
Zurich, biologisch-chemische Teilrichtung
Praktikum an der Harvard Medical School, Brigham and
Women's Hospital, Hematology Division, Boston MA,
USA, Prof. H.F. Bunn
Vollpraktikum am Institut fur Biochemie, ETH Zurich,
Prof. G. Semenza
Diplomarbeit am Institut fur Pathlogiem Abteilung fur
experimentelle Pathologie, Umversitatsspital Zurich, Prof.
H. Hengartner und Prof. R.M. Zinkernagel
Titel:" Expression Lymphocyten-spezifischer Gene bei
viraler und bakterieller Infektion"
Diplom der ETH Zurich
1988 Beginn einer Doktorarbeit an der Abteilung fur
Naturwissenschaften der ETH Zurich, ausgefiihrt am
Institut fur experimentelle Immunologie,
Umversitatsspital Zurich, Prof. H. Hengartner und Prof.
R.M. Zinkernagel
1990 Aufenthalt am Ontario Cancer Institute, Toronto, Canada,
Prof. T.W. Mak
1990 -1992 Zusammenarbeit mit der Embryonic Stem Cell Unit,
Praklinische Forschung, Sandoz AG, Dr. B. Ledermann
und Dr. K. Burki
Zurich, den 13.12.1992