University of Southern California Dissertations and Theses

Molecular mechanism of genomic imprinting in somatic cell cultures

(USC Thesis Other)

Molecular mechanism of genomic imprinting in somatic cell cultures

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may
be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, prim bieedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorised copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g^ maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand corner and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in
reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6 " x 9 " black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly
to order.
A Bell & Howell Information Company
300 North Zeeb Road. Ann Arbor M l 48106-1346 USA
313'761-4700 800/521-0600
MOLECULAR MECHANISM OF GENOMIC IMPRINTING
IN SOMATIC CELL CULTURES
by
Pamela Louise Eversole-Cire
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Biochemistry and Molecular Biology)
September 1994
Copyright 1994 Pamela Louise Eversole-Cire
OMI Number: 9600972
UMI Microform 9600972
Copyright 1995* by OKI Company. All rights r•served.
This microform edition is protected against unauthorized
copying under Title 17* United States Code.
UMI
300 North Xeeb Road
Ann Arbor* HI 40103
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90007
This dissertation, written by
Pamala Louisa Evsrsola-CIra
under the direction of h.ff. Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
Dean of Graduate Studies
Date ....Q c £ Q b ftc .,3 & ...1 3 2 4
DISSERTATION COMMITTEE
Chairperson
DEDICATION
To Bill Eversole, my father, who taught me that any goal was
attainable if pursued with vigor.
To Wanda Eversole, my mother, who has always set an example
for me to emulate.
To Frank Cire, my husband and loving friend, without his
unselfish support, understanding, and patience this goal
would never have been achieved.
ACKNOWLEDGEMENTS
My sincere appreciation goes to Peter Jones, my thesis advisor
and mentor, for his support and guidance throughout the
course of this work. His valued criticism and advice allowed
me to successfully complete this dissertation and achieve my
educational goals.
Felicidad Gonzales deserves a special thank you for her
technical assistance and help with various aspects of this work.
The friendship of the many members of the Jones' laboratory
helped to make the difficult moments of this work more
enjoyable.
The critical evaluation and advice of the members of my
advisory committee, Debra Johnson, Ph.D., Robert Deans, Ph.D.,
Michael Stallcup, Ph.D., Robert Stellwagen, Ph.D., and Peter
Jones, Ph.D., are gratefully acknowledged.
TABLE OF CONTENTS
DEDICATION........................................................................................................ ii
ACKNOWLEDGEMENTS...................................................................................iii
LIST OF FIGURES............................................................................................. vii
LIST OF TABLES.................................................................................................x
ABSTRACT.............................................................................................................xi
CHAPTER I
INTRODUCTION TO GENOMIC IMPRINTING...........................................1
Evidence for Genomic Imprinting in the M ouse.................................1
Possible Mechanisms for Genomic Imprinting
DNA M ethylation.........................................................................................4
DNA methylation during Mouse Embryonic
Development...................................................................................................... 6
Chromatin C onform ation......................................................................... 8
Imprinted G enes...............................................................................................10
I g f-2 ........................................................................................................................13
H 1 9 ..........................................................................................................................17
Genomic Imprinting and Human D isease.............................................21
5-Azacytidine......................................................................................................... 23
5 -B ro m o d e o x y u rid in e ................................................................................... 24
SUMMARY............................................................................................................25
CHAPTER 2
Establishment of Mouse Somatic Cell Cultures from
Embryos with Maternal Uniparental Disomy for Distal
Chromosome 7 .................................................................................................. 28
INTRODUCTION................................................................................................. 28
MATERIALS AND METHODS
C ells........................................................................................................................ 31
Isolation of Subclones...........................................................................................32
Isolation of Nucleic Acids................................................................................... 33
Probes......................................................................................................................34
RESULTS
Growth of Cultures......................................................................................... 34
Genomic Structure of Igf-2 ........................................................................35
Expression of Igf-2 ........................................................................................ 38
Methylation of Ig f-2 .....................................................................................45
iv
DISCUSSION 67
CHAPTER 3
Activation of an Imprinted Igf-2 Gene in Mouse Somatic
Cell Cultures.................................................................................................... 70
INTRODUCTION..............................................................................................70
MATERIALS AND METHODS
Cell Culture...................................................................................................... 72
Isolation of Subclones....................................................................................... 72
Isolation of Nucleic Acids................................................................................ 73
T reatm en ts
Nucleoside Analogs................................................................................74
Calcium Ionophore.................................................................................74
Heat or Cold Shock.................................................................................75
Sodium Butyrate.....................................................................................75
Reverse Transcription (RT)-Polymerase Chain Reaction
(PCR)....................................................................................................................75
RESULTS
RT-PCR Assay.................................................................................................78
Effects of Agents and Treatments Known to Alter Gene
Expression Patterns.....................................................................................85
Isolation of MatDi7 Subclones which Express Igf-2 ..................... 98
DISCUSSION...................................................................................................101
CHAPTER 4
Coordinate Regulation of lgf-2 and H19 in Cultured Cells
Indicates that these Genes may be Regulated by a
Common Regulatory Mechanism................................................................. 1 1 0
INTRODUCTION 1 10
MATERIALS AND METHODS
Cell Culture.....................................................................................................1 1 2
Expression Studies.....................................................................................113
DNA Synthesis.................................................................................................. 1 1 4
Southern Blot Analysis.................................................................................... 1 1 5
Probes................................................................................................................ 115
Reverse Transcription (RT)-Polymerase Chain Reaction
(PCR).................................................................................................................. 1 15
RESULTS
Expression of Igf-2 and H19 are Extinguished in Unison 116
Coordinate Growth Regulation of Igf-2 and H I 9 ......................... 120
Reactivation of the Maternal Igf-2 Allele is Associated
with Increased Methylation of the 5' Region of Igf-2 and
of the HI9 P ro m o te r.................................................................................133
DISCUSSION...................................................................................................148
CHAPTER 5
Activation of an Imprinted Human H I9 in Human-Mouse
Somatic Cell Hybrids.................................................................................159
INTRODUCTION............................................................................................159
MATERIALS AND METHODS
Cell Culture....................................................................................................161
BUdR Treatment.............................................................................................. 1 6 1
Isolation of Nucleic Acids..............................................................................1 6 2
Polymerase Chain Reaction (PCR).......................................................162
Reverse Transcription (RT)-Polymerase Chain Reaction
(PCR)................................................................................................................163
O ligom ers..................................................................................................... 1 65
RESULTS
Somatic Cell Hybrids Containing Single Copies of the
Human Igf-2/HI9 Locus.........................................................................166
Retention of Human Igf-2 in Microcell Hybrids................................1 7 0
Expression of the Mouse Igf-2 and H I9 in Microcell
H y b rid s..........................................................................................................1 7 3
The Level of Human Igf-2 Expression is Greater than that
of Human H I9 in Somatic Cell Hybrids.............................................178
Activation of an Imprinted Human H I9 Gene upon BUdR
T re a tm e n t.....................................................................................................1 8 2
DISCUSSION..................................................................................................191
CHAPTER 6
Summary and Conclusions............................................................................ 1 9 6
LIST OF REFERENCES.................................................................................2 0 4
APPENDIX.....................................................................................................2 20
vi
LIST OF FIGURES
CHAPTER 1
Figure 1-1. Summary of endogenous imprinted genes................ 11
Figure 1-2. Genomic structure of the region containing
the mouse insulin-2, Igf-2, and H I9 ............................................................ 15
Figure 1-3. Reciprocal imprinting of Igf-2 and H J 9 .....................19
CHAPTER 2
Figure 2-1, Genomic structure of the mouse Igf-2 and
summary of methylation analysis of Igf-2 .......................................36
Figure 2-2. Expression of Igf-2 in low-passage cultured
cells.......................................................................................................................39
Figure 2-3. Expression of Igf-2 in high-passage cells................ 42
Figure 2-4, Methylation analysis of the Igf-2 promoter 1
and the region 5' in normal and MatDi7 cells..................................48
Figure 2-5. Methylation analysis of the Igf-2 promoter 1
region in MatDi7 subclones and in embryonic DNA......................50
Figure 2-6. Methylation analysis of the Igf-2 promoter 1
and the region 3' in normal and MatDi7 cells..................................52
Figure 2-7. Methylation analysis of the Igf-2 promoter 2
region in normal and MatDi7 cells........................................................ 56
Figure 2-8. Methylation analysis of the Igf-2 promoter 3
region in normal and MatDi7 cells........................................................ 58
Figure 2-9. Methylation analysis of the Igf-2 coding
region in normal and MatDi7 cells........................................................61
Figure 2-10. Igf-2 can become methylated in permanent
cell lines...............................................................................................................65
CHAPTER 3
Figure 3-1. Schematic of the primer design for the
RT-PCR assay.................................................................................................. 79
Figure 3-2. Demonstration of RT-PCR assay with respect
to the number of PCR cycles.................................................................... 81
Figure 3-3. Demonstration of RT-PCR assay with respect
to the number of input molecules................................................................ 83
Figure 3-4. Induction of Igf-2 expression in low-passage
MatDi7 cells upon 5-aza-CdR treatment................................................... 88
Figure 3-5. Induction of Igf-2 expression in MatDi7
immortalized cells does not result from a selection
p ro c e ss................................................................................................................. 90
Figure 3-6. Induction of Igf-2 expression in a MatDi7
subclone upon BUdR treatment.......................................................................93
Figure 3-7. Time course of induction of Igf-2 expression
in MatDi7 subclone upon BUdR treatment........................................96
Figure 3-8. Expression of Igf-2 in second-passage MatDi7
cells treated with 5-aza-CdR and BUdR.............................................. 99
Figure 3-9. Growth-regulated expression of Igf-2 in
subclone MatDi7 Aza Cl L..............................................................................1 0 2
Figure 3-10. Growth-regulated expression of Igf-2 in
MatDi7 1-la subclone...................................................................................... 1 04
CHAPTER 4
Figure 4-1. Expression of Igf-2 and H I9 in subclones of
normal cells........................................................................................................118
Figure 4-2. Coordinate expression of Igf-2 and H19 in
Normal Cl 1 cells..........................................................................................1 2 1
Figure 4-3A. H19 expression in MatDi7 l-6b cells
becom es deregulated................................................................................ 1 24
Figure 4-3B. t3H]Thymidine incorporation in
MatDi7 l-6b cells induced to become quiescent either by
confluence arrest or serum starvation.....................................................12 6
Figure 4-4A. Igf-2 and H I9 demonstrate reciprocal
expression patterns in MatDi7 1-la cells......................................... 129
Figure 4-4B. [3H]Thymidine incorporation in MatDi7 1-la
cells induced to become quiescent either by confluence
arrest or serum starvation.....................................................................131
Figure 4-5. Map of region 5' to Igf-2 and the H I 9
promoter and summary of the methylation analysis of
the region 5' to Igf-2 and the H19 p ro m o te r................................ 134
Figure 4-6. Increased methylation of the region 5' to the
Igf-2 promoters correlates with reactivation of the
maternal Igf-2 allele.........................................................................................1 3 6
Figure 4-7. Methylation analysis of exon 1 of H I 9 in
normal and MatDi7 cell lines........................................................................1 4 0
Figure 4-8. Methylation analysis of the H I9 promoter in
normal and MatDi7 cell lines........................................................................1 4 3
Figure 4-9. Increased methylation of the maternal H 19
promoter is associated with reactivation of the maternal
Igf-2 allele........................................................................................................ 145
Figure 4-10. Proposed model for mechanism of Igf-2 and
H19 genomic imprints.............................................................................15 1
CHAPTER 5
Figure 5-1. Diagram of human-mouse somatic cell hybrid
generation.........................................................................................................168
Figure 5-2. Human Igf-2 genomic sequence is retained in
human-mouse somatic cell hybrids.................................................. 17 1
Figure 5-3. Expression of the mouse Igf-2, H 19, and
MyoD sequences in human-mouse somatic cell hybrids 1 7 4
Figure 5-4. Expression of human Igf-2 and H19 in
human-mouse somatic cell hybrids.................................................. 179
Figure 5-5. Expression of human Igf-2 is downregulated
while expression of human H19 is upregulated in somatic
cell hybrids upon BUdR treatment.................................................... 184
Figure 5-6. Cytotoxicity study for BUdR treatment of
human-mouse somatic cell hybrids.................................................. 188
LIST OF TABLES
CHAPTER 3
TABLE 3-1. Effects of treatments on Igf-2 expression in
MatDi7 cells...........................................................................................................87
CHAPTER 5
TABLE 5-1. Effect of BUdR treatment on Igf-2 and H I9
expression in microcell hybrids.......................................... 186
ABSTRACT
The mouse insulin-tike growth factor 2 (Igf-2) and the H19
genes, located on distal chromosome 7, are genomicatly
imprinted such that the paternally derived Igf-2 and the
maternally derived H I 9 are preferentially expressed. Igf-2
and HI9 are reciprocally imprinted yet demonstrate a similar
spatio-tem poral pattern of expression during em bryogenesis.
Somatic cell cultures were derived from mouse embryos with
maternal uniparental disomy for distal chromosome 7 (MatDi7)
and normal littermates to examine the molecular basis of
genomic imprinting.
MatDi7 cells maintained the Igf-2 imprint since these cells
generally did not express Igf-2 . Methylation analysis indicated
that extensive de novo methylation of Igf-2 had not occurred
during culturing. Expression of the maternal Igf-2 in MatDi7
cells was increased in a dose-dependent manner by treatment
with 5-aza-2'-deoxycytidine or 5-bromodeoxyuridine. The
mechanism of the Igf-2 imprint may involve methylation
and/or chromatin conformation changes affected by these two
agents.
Igf-2 and H I 9 were expressed in a coordinated manner in
normal cells suggesting a common regulatory mechanism.
xi
Spontaneous activation of Igf-2 in a MatDi7 clone was
associated with de novo methylation of the Igf-2 upstream
region and the H J9 promoter so that the maternal allele(s) had
acquired a methylation pattern characteristic of the paternal
allele. Expression of Igf-2 and H I9 from different parental
chromosomes may result from parental-specific epigenetic
m odifications.
The molecular mechanisms of the human Igf-2 and H 19
im prints were studied using human-mouse somatic cell hybrids
containing a single copy of human chromosome l i p where the
Igf-2/HI9 locus resides. Treatment of the hybrids with
5-bromodeoxyuridine resulted in a dose-dependent increase in
expression of the human H I9 and a dose-dependent decrease
in Igf-2 expression. These results indicate that Igf-2 and H I 9
may compete for transcriptional regulatory element(s) allowing
only Igf-2 or H 19 to be transcribed from a single parental
chromosome. Therefore, the genomic imprints of Igf-2 and
H I9 may be mechanistically and functionally related.
CHAPTER 1
INTRODUCTION TO GENOMIC IMPRINTING
One of the fundamental questions in developmental biology
concerns the different roles of the maternal and paternal
genomes during embryonic development. Several lines of
evidence indicate that the parental genomes do not function
equivalently during mammalian development. Although the
maternal and paternal gametes contribute the same genetic
material, different epigenetic modifications may result in
functional differences between certain chromosomal regions.
DNA methylation and changes in chromatin conformation are
possible epigenetic mechanisms which may affect gene
expression to direct proper development of embryonic cells.
Determining the process by which genes are epigenetically
modified or genomically imprinted is important to
understanding the control of gene expression and regulation of
mammalian development.
Evidence for Genomic Im printing In the Mouse
Complete development in mammals requires the presence of
both the maternal and paternal genomes unlike other species
which successfully undergo parthenogenesis. Pronuclear
1
transplantation experim ents have dem onstrated that both
parental genomes are required for proper development of the
mouse conceptus (McGrath and Solter, 1984). Complete
duplication of parental genomes has shown that the maternal
genome contributes mainly to the development of the embryo
whereas the paternal genome contributes to development of
the extra-embryonic lineages (Barton et al., 1984; Surani et al.,
1984). Therefore, the individual contributions of the parental
genomes to the developing conceptus appear to be different
suggesting that the chromosomes may be imprinted differently
in the parents to affect gene activity in the zygote.
Genetic complementation studies have also indicated that
the parental genomes do not function equivalently during
mouse development. Analysis of specific chromosomal
duplications using offspring of mice carrying different
chromosomal translocations or different metacentric
chromosomes has shown that certain combinations of
duplicated regions may produce abnormal phenotypes or
lethality in the progeny (Cattanach and Kirk, 198S). These
results suggest that the duplicated regions of the chromosomes
contain genes whose expression is determined by parental
origin. The functional differences between the maternal and
paternal genomes indicate that some type of chromosomal
imprinting event must affect the activity of certain genes
2
within these duplicated chromosomal regions. Thus, certain
regions of autosomes must be represented by both parental
genomes for normal phenotype and development of the mouse.
The preferential inactivation of the paternally inherited
X-chromosome in the extra-embryonic lineage of the mouse
also indicates that the genetic material contributed by the
gametes is functionally different during embryogenesis (Monk
and Grant, 1990; Takagi and Sasaki, 1975). Together these
findings suggest that during the initial stages of development
there exists a mechanism which allows distinction between the
parental chromosomes.
Several criteria must be met for a genomic imprinting
process. The imprint must be established prior to or during
gametogenesis, be faithfully inherited, and persist throughout
DNA replication and cell division. The imprint must also be
reversible: it must be removed in the germ line and be
re-established in the gametes. Additionally, the imprint must
act in a cis manner since an active and an inactive allele are
present within a single cell. DNA methylation and chromatin
conformation changes are two epigenetic modifications which
may play roles in the genomic imprinting process (Surani et al.,
1990).
3
P ossib le M echanism s for G enom ic Im p rin ting
DNA M ethylation
One molecular mechanism of genomic imprinting in the
mammalian genome has been hypothesized to be DNA
methylation (Holliday, 1989). The transfer of a methyl group
from S-adenosyl-L-methionine to cytosine residues present in
CpG dinucleotides is catalyzed by the enzyme DNA
methyltransferase. The modification of cytosine to
5-methylcytosine occurs on approximately 3% of the cytosine
residues in the mammalian genome (Doeffler, 1983; Riggs and
Jones, 1983). The distribution of CpG dinucleotides is
nonrandom in the vertebrate genome and clusters known as
CpG islands exist in concert with other regions of the genome
which are CpG depleted. CpG islands are characterized by a
higher guanine and cytosine content than is present in the
remainder of the genome and are regions generally devoid of
methylation (Tykocinski and Max, 1984; Bird, 1986; Gardiner-
Garden and Frommer, 1987). Recently it has been shown that
CpG islands are located in regions of the genome which
replicate early in S phase (Craig and Bickmore, 1994). CpG
islands are usually found in the S' promoter regions of
housekeeping or tissue-specific genes (Bird, 1987) but can also
be found within the body of genes (Jones et al., 1990) and
4
within introns (Stoger et al., 1993). The methylation status of
some CpG islands are altered in cell culture affecting the
expression of the genes with which they are associated (Jones
et al., 1990; Antequera et al., 1990).
Many properties of DNA methylation are consistent with the
features expected for an imprinting process. Methylation
patterns are established post-replicatively and once
established are stably inherited. DNA methylation has been
implicated in controlling the regulation of eucaryotic gene
expression (Cedar, 1988). It is possible that DNA methylation
may inhibit the transcription of genes to negatively regulate
their expression by altering chromatin structure (Keshet et a).,
1986) or by preventing the interaction between a transcription
factor with its proper binding site (Becker et al., 1987).
Methylation may also positively affect the expression of certain
genes by preventing the binding of a repressor molecule
(Stoger et al., 1993) or by increasing the binding of an
activator. Therefore, differentia] imprinting of parental
genomes may possibly be achieved by chemical modification,
such as DNA methylation, of specific DNA sequences.
5
DNA M ethylation during M ouse Embryonic
D e v e lo p m e n t
There is considerable evidence which supports DNA
methylation as a mechanism for genomic imprinting in the
mouse. The methylation pattern of transgenes randomly
inserted into the mouse genome have been examined to
determine if the parental genomes contain differentially
modified regions. Results of the transgenetic mouse
experiments indicate that the pattern of methylation of some
transgenes is influenced by parental inheritance (Reik et at.,
1987; Sapienza et al., 1987; Ueda et al., 1992). The imprinted
methylation pattern of a specific transgene, c-myc, has also
been correlated with its expression (Swain et al., 1987).
Differential methylation of transgenes suggests that
endogenous genes may also be subject to similar modifications
affecting their expression during development.
Differences in the overall levels of DNA methylation exist
between the gametes. The overall level of methylation in
sperm DNA is greater that the level in oocyte DNA in the mouse
(Monk et al., 1987). In addition, certain repetitive and low
copy DNA sequences were found to be highly methylated
during spermatogenesis and undermethylated during oogenesis
(Sanford et al., 1987). The differential methylation of the
6
gametic DNA may provide a mechanism for imprinting of the
parental genomes. The initial methylation differences between
the germ lines may influence the onset of gene expression of
the parental alleles during early developmental stages.
Although global methylation differences exist between the
gametes, it is also possible that the mechanism may involve the
modulation of methylation patterns along specific regions of a
gene. This hypothesis is supported by the observation that CpG
islands of certain genes on the inactive X-chromosome have
higher levels of methylation than the complementary island on
the active X-chromosome and is generally associated with
transcriptional silencing (Wolf and Migeon, 1985; Toniolo et al.,
1988; Keith et al., 1986). If modulation of DNA methylation
patterns is the determinant of genomic imprinting, these
patterns could be established during gametogenesis by the
methyltransferase and maintained by the enzyme throughout
early embryonic development and into the adult state.
An analysis of the methylation levels in various tissues of
the mouse conceptus has also shown that both spatial and
temporal changes in DNA methylation occur (Monk et al.,
1987). Global methylation differences between the embryonic,
extra-embryonic, and germ cell lineages have been shown to
occur during various stages of development. An overall
7
decrease in the methylation of the developing embryo occurs
during preimplantation development (Kafri et al., 1992;
Hewlett and Reik, 1991) with lineage specific de novo
methylation beginning to reestablish DNA methylation patterns
during the gastrulation stage. A recent study has shown that
the mechanism for the observed genome-wide demethylation
is an active process which recognizes specific sites to remove
methyl groups (Kafri et al., 1993). If specific sites which were
differentially methylated in the parental genomes escape
demethylation, these sites could serve as primary imprinting
signals. In fact, recent studies report the existence of a
mechanism which can be acquired post fertilization and
functions in the early embryo to protect specific sites from the
general demethylation event during early embryonic
development (Kafri et al., 1993). These results strongly
support a role for DNA methylation in the genomic imprinting
process.
C hrom atin C onform ation
Chromatin structure is another epigenetic mechanism which
may play a role in the genomic imprinting process. It has been
proposed that the inheritance of chromatin patterns from the
parental genomes may serve as a primary imprinting signal
(Groudine and Conkin, 1985; Weintraub, 1985). Many features
8
of the structure of chromatin satisfy the requirements for an
imprinting mechanism. Chromatin conformation is stably
maintained, is heritable, and can regulate the expression of
genes in cis.
Chromatin is a nucleoprotein complex formed between DNA
and histones which are subject to post-translational
modification by acetylation, ADP-ribosylation, phosphorylation,
and methylation (Stryer, 1988). Modification of the histone
molecules may affect DNA packaging and thereby affect
replication as well as the transcriptional activity of genes.
Regions of chromatin which contain actively transcribed genes
usually have fewer methylated cytosine residues and exhibit
increased sensitivity to nucleases such as DNase 1 (Keshet et
a)., 1986; Stryer, 1988) which may allow access of transcription
factors. Although DNA methylation may be involved in the
formation of inactive chromatin, it has recently been
demonstrated that inactive chromatin can also spread from a
highly methylated region to a region which is not methylated
(Kass et al., 1993). Therefore, changes in the pattern of DNA
methylation and chromatin conformation during developm ent
may be responsible for the differential functions of the
parental genomes.
9
Im p rin ted G enes
Several endogenous genes have recently been identified
which are believed to be genomically imprinted in the mouse
(Fig. 1-1). The insulin-like growth factor 2 receptor (lgf-2r ;
Barlow et al., 1991), insulin-like growth factor 2 (Igf-2\
DeChiara et al., 1991), H19 (Bartolomei et al., 1991), small
ribonucleoprotein polypeptide n (Snrpn', Leff et al., 1992;
Cattanach et al., 1992) and the U2af-binding protein related
sequence (U2afbp-rs\ Hayashizaki et al., 1994) have been
shown to exhibit parental-specific expression. Igf-2, Snrpn,
and U2afbp-rs are primarily transcribed from the paternally
inherited allele whereas Igf-2r a n d / / / 9 are primarily
transcribed from the maternally inherited allele. The
parental-specific expression of Igf-2, H I 9, and Snrpn has also
been observed in humans (Glaser et al., 1989; Ohlsson et al.,
1993; Giannoukakis et al., 1993; Zhang et al., 1993; Glenn et al.,
1993; Reed and Leff, 1994). The lgf-2r is, however, not
imprinted in human (Kalscheuer et al., 1993).
A general characteristic of these genomically imprinted
genes is allele-specific methylation. Methylation of a region 5'
to the mouse Igf-2 (Sasaki et al, 1992; Brandeis et al., 1993)
and intron 5 of the human Snrpn occurs on the active paternal
allele. Methylation of the promoter of both the mouse and
1 0
Figure 1-1. Sum m ary o f endogenous im printed genes.
An open box represents the entire gene sequence for Igf-2 ,
H J 9 , and U2afbp-rs and partial exonic sequences for !gf-2r and
Snrpn. Hatched boxes represent intron regions. Slashed lines
represent 27 kb of intervening sequence between the Igf-2r
promoter and an intron. The lack of imprinting of Igf-2r in
humans is represented by dashes. Allele-specific expression is
designated by a female symbol for maternal and by a male
symbol for paternal. Arrows signify transcription.
1 1
Gene Parental-Specific Expression
and Methytation
tgf-2r
E
lgf-2
H19
Snrpn
U2afbp-rs
o o • •
• • ^
OO
• •
_F -.
o o o
r *
I I
• •
_Pn
9
cf
Expressed allele
Mouse Human
?
cf c f
9 9
c r c f
ct
human H19 (Bartolomei et al., 1993; Ferguson-Smith et al.,
1993; Brandeis et al, 1993) occurs on the inactive paternal
allele and methylation of a single Not I site located 5' to the
mouse U2afbp-rs (Hayashizaki et al., 1994) occurs on the
inactive maternal allele. Interestingly, there are two regions of
methylation of Igf-2r\ the first region is located within the
promoter on the inactive paternal allele and the second region
is located within an intron located 27 kb downstream of the
transcriptional start on the active maternal allele (Stoger et al.,
1993). Support that these parental-specific methylation
differences are involved in the imprinting process comes from
a recent study in which mice deficient in DNA
methyltransferase no longer expressed fgf-2, H 1 9, or Igf-2r in
an allele-specific manner (Li et al., 1993). Although,
parental-specific methylation differences have been identified
for many of the endogenous imprinted genes, whether or not
DNA methylation serves as a primary imprinting signal or the
exact mechanism(s) by which methylation may regulate
allele-specific expression of these genes is not known.
l g f ’ 2
The mouse insulin-like growth factor (Igf-2) gene is located
on the distal portion of chromosome 7, approximately 18 kb
distal to the insulin gene (Rotwein and Hall, 1990) and 90 kb
13
proximal to H19 (Zemel et al., 1992), in the same
transcriptional orientation (Fig. 1-2). The genomic structure of
Igf-2 is complex and spans over 12 kb of genomic DNA
sequence. The gene is composed of 6 exons with exons 1
through 3 encoding different S' untranslated regions each
transcribed by a different promoter, one of which is a CpG
island, while exons 4 through 6 encode a 180 amino acid
precursor (Rotwein and Hall, 1990). Upstream of exon 1 are
two pseudoexons which exhibit significant homology to exons 2
and 3 of the human gene. Three Igf-2 transcripts are produced
which contain different 5’ untranslated regions but the same
coding sequence and 3' untranslated regions as a result of
differential promoter usage and RNA processing. Igf-2 is
parentally imprinted such that the gene is primarily
transcribed from the paternal allele (DeChiara et al., 1991).
The complexity of Igf-2 suggests that the expression of this
gene is subject to multiple controls. Since the gene is
transcribed by multiple promoters, the genomic imprinting
event which silences the gene must repress transcription from
each promoter on the maternal allele.
Several lines of evidence suggest that Igf-2 functions as an
important growth factor during embryonic development of the
mouse. Igf-2, which encodes a 67 amino acid protein, is
synthesized primarily during the fetal and neonatal periods of
1 4
Figure 1-2. Genom ic structure o f the region containing
the m ouse in su lin -2 , Igf-2, and HI9,
Exons and pseudoexons (\|/) are shown as solid filled boxes. The
two enhancer sequences downstream of H19, designated El and
Eli, are shown as stippled boxes. The three promoters of Igf-2
(PI, P2, and P3) are indicated. Transcriptional initiation sites
are depicted as arrows. The translational initiation site (ATG),
the termination codon (TGA), and the major polyadenylation
site (poly A) are shown for Igf-2. Slashed lines represent
approximately 90 kb of intervening DNA sequence between
lgf-2 and HI9. Scale, in kb, is shown.
1 5
tnsuhn-2 tgf-2
yl y ’ PI P2
ft- - - - - - - - - - - - - - - - - - - - - - -1 — 1 - - - - - - - - - - - - - - -
Exons 12 3 1 2
H19
2 k b
l_... J
ATG TGA poly A
E l E ll
early mouse development (Adams, 1983). Igf-2 transcripts are
first detected during the two cell stage of development
(Rappolee et ah, 1992) with the transcript levels increasing
through the blastocyst stage. The gene is expressed
abundantly in the extra-embryonic and embryonic tissues such
as cells of mesodermal derivatives including skeletal muscle
and chondroblasts in post-implantation embryos (Lee et al.,
1990). Gene knockout experiments have shown that
heterozygous mice containing a disrupted Igf-2 exhibit a
growth deficiency phenotype (DeChiara et al., 1990). Chimeric
embryos containing cells with a paternal duplication/maternal
deficiency of chromosome 7 and therefore approximately a
two-fold increase in the level of the Igf-2 gene product show a
marked enhancement of embryonic growth (Ferguson-Smith et
al., 1991). Therefore, the developmental pattern of expression
as well as its association with growth control suggests that the
Igf-2 gene product is an important embryonic mitogen and
that dosage control of this gene by genomic imprinting is
essential for normal development.
H 1 9
H J9 lies approximately 90 kb distal to Igf-2 (see Fig. 1-2;
Zemel et a)., 1992). The gene is composed of 5 exons which
range between approximately 0.1 kb and 1.3 kb within a 3 kb
1 7
region (Pachnis et al., 1988). The promoter region of H19
contains a CpG island (Ferguson-Smith et al., 1993) and two
enhancers are located between 5 and 6.5 kb downstream of the
coding region (Yoo-Warren et al., 1988). H I9 is conserved
among many mammalian species (Brannan et al., 1990). HI9 is
oppositely imprinted as Igf-2 (Fig. 1-3) with transcription
occurring primarily from the maternal allele during
development (Bartolomei et al., 1991). H19 is transcribed by
RNA polymerase U (Brannan et al., 1990) and the 2.5 kb
message is processed by splicing and polyadenylation.
However, the H19 message contains multiple translation
termination signals and lacks a conserved open reading frame
indicating that the gene may not encode a functional protein
(Pachnis et al., 1988; Brannan et al., 1990). The H I9 mRNA has
been localized to the cytoplasm associated with a 7S
ribonucleotide particle (Brannan et al., 1990). It has, therefore,
been postulated that H19 may function at the mRNA level and
there is recent evidence which demonstrates that the H I9 RNA
exhibits tumor suppressor activity (Hao et al., 1993).
H I9 exhibits a similar spatial and temporal expression
pattern during embryogenesis as Igf-2 with high expression in
endodermal and mesodermal tissues of the mouse embryo
(Poirier et al., 1991; Lee et al., 1990). The two genes are
initially activated in the extra-embryonic tissues soon after
1 8
Figure 1-3. Reciprocal imprinting of igf-2 and H 19.
Open rectangles represent the Igf-2 and H19 genes and the
stippled polygons represent the two H I9 enhancer sequences.
Arrows signify transcription. Igf-2 and HI9 are imprinted
such that Igf-2 is expressed primarily from the paternal allele
and H I9 primarily from the maternal allele.
19
rw i F f ? 9 |
implantation and are expressed in similar tissues of the
embryo proper. Expression of H I 9 and Igf-2 are repressed in
most tissues after birth; H19 is expressed predominately in
skeletal muscle and Igf-2 is biallelically expressed in the
choroid plexus and leptomeninges in the adult (DeChiara et al.,
1991). Coexpression of HI9 and Igf-2 suggests that these
genes may share transcriptional regulatory elements.
Conservation of H I9 and its developm ental^ regulated
pattern of expression suggests that H19 may serve an
important role during murine development. This suggestion is
supported by the observation that ectopic expression of the
wildtype H I9 in transgenic mice results in prenatal lethality
between e l4 and birth (Brunkow and Tilghman, 1991). This
finding suggests that the gene dosage of H I9 is highly
regulated and imprinting is necessary to control levels of H 19
expression for proper mouse development.
Genomic Imprinting and Human Disease
Genomic imprinting may be involved in the genesis of
several types of human diseases. The Beckwith-W iedemann
syndrome, for example, which is characterized by the
overgrowth of many tissues and a predisposition to embryonal
tumors such as Wilms' tumor of the kidney, has been linked to
21
I lp l5 (Waziri et al., 1983; Turleau et al., 1984). The sporadic
cases of this growth abnormality are believed to result from
paternal duplication of human llp l S while the familial cases
of this disease are related to maternal transmission of
chromosome 11 (Mannens et al., 1988; Henry et al., 1991).
This region of human chromosome 11 is syntenic with the
region of mouse chromosome 7 which harbors Igf-2 and H I9.
(Glaser et al., 1989). Recently the human Igf-2 and HI9 were
found to exhibit parental-specific expression similar to the
mouse homologs (Glaser et al., 1989; Ohlsson et al., 1993;
Giannoukakis et al., 1993; Zhang et al., 1993). In addition to
the above mentioned genetic alterations, "loss of imprinting” of
Igf-2 and H I9 has been observed for some cases of Wilms'
tumor (Rainer et al., 1993; Ogawa et al., 1993). Thus genomic
imprinting may play a role in the genesis of specific types of
tum or syndromes.
Conservation of genomic imprinting of Igf-2 and H19 am o n g
human and mouse allows one to study the imprinting process
using mouse models. Such models have proven to be more
amenable and less limiting than using human subjects. The
presence of both an active and an inactive allele within a
normal cell complicates the study of genomic imprinting.
However, it is possible to obtain mice with a duplication of the
distal portion of the maternal chromosome 7 (which harbors
22
the igf-2/HI9 locus) and without the paternal complement
(Cattanach and Kirk, 1985; Searle and Beechey, 1990). Somatic
cell cultures established from these mice will create a useful in
vitro model system with which to study the molecular
mechanism of genomic imprinting. Treatment of such cell lines
with agents known to affect the methylation profile and/or
chromatin structure of genes, such as 5-azacytidine and
5-bromodeoxyuridine, two epigenetic modifications thought to
be involved in the genomic imprinting process, may be helpful
in determining the nature of the imprint imposed on the
Igf-2/HI 9 locus.
5 - A z a c y tid ln e
5-Azacytidine (5-aza-CR) and 5-aza-2'-deoxycytidine
(5-aza-CdR) are base analogs which have been shown to have
profound effects on the expression of many genes in tissue
culture (Jones, 1985). The nucleoside analog differs from
cytidine in the substitution of a nitrogen atom for a carbon
atom at the 5 position of the pyrimidine ring. The analog is
believed to exert its action(s) via incorporation into DNA
sequences causing demethylation by irreversibly binding to
the DNA methyltransferase enzyme thereby inhibiting its
subsequent action (Creusot et al., 1982; Taylor and Jones,
1982). Inhibition of the DNA methyltransferase causes an
23
alteration of methylation patterns which may result in the
induction of many different genes in culture (Jones, 1985).
5 -B rom odeoxy u ridine
5-Bromodeoxyuridine (BUdR) is a thymidine analog which
incorporates into DNA and affects the activities of certain
cellular functions. The substitution of a bromine atom at the 5
position of the pyrimidine ring results in the analog being more
electronegative than thymidine and may be the direct cause of
some of the observed effects. Treatment with BUdR during
DNA replication is necessary to produce effects in some
systems (Bischoff and Holtzer, 1970; Miura and Wilt, 1971;
Weintraub et al., 1972, Stellwagen and Tomkins, 1971A and
197IB) but it has also been shown that BUdR-induced effects
can be generated in the absence of cell division indicating that
the mechanism of BUdR action is not always dependent on DNA
incorporation (Schubert and Jacob, 1970).
The effects of BUdR are numerous (for review see Rutter et
al., 1973). Treatment of cells with BUdR has been shown to
reversibly inhibit the differentiation process without unduly
affecting the growth rate or viability of cells. In some systems,
exposure to BUdR results in the repression of certain genes
(Tapscott et al., 1989) while in other systems certain genes are
24
induced. BUdR exposure may affect gene activity by inhibiting
DNA processing by decreasing the rate of RNA synthesis as well
as by altering the types of mRNA species produced.
Incorporation of BUdR into the DNA may also alter the
interaction between regulatory molecules with the DNA
sequence. The melting profile of DNA containing BUdR also
changes which may affect protein:DNA complexes necessary to
maintain proper chromatin structure (Rutter et al., 1973).
Sum m ary
The functional differences of the maternal and paternal
genomes during early embryonic development may be the
result of a genomic imprinting event. Genomic imprinting is
thought to be an epigenetic process by which homologous
alleles are marked or modified so that their genetic expression
is a result of parental inheritance. DNA methylation and
chromatin conformation are two epigenetic modifications which
have been proposed to be involved in the process of genomic
imprinting (Holliday, 1989; Groudine and Conkin, 1985;
Weintraub, 1985). The differential modification of the parental
genomes may occur so that during mammalian development
there will be preferential expression of alleles to ensure proper
development of the zygote. In the adult, imprinting may
25
control the expression of genes involved in promoting cell
growth (Steenman et al., 1994).
Igf-2 and HI 9, two endogenous imprinted genes, are
thought to be involved in the regulation of cell growth. The
mouse Igf-2 may encode an important embryonic mitogen and
H i9 may function as a tumor suppressor gene (DeChiara, et al.,
1990; Ferguson-Smith et al., 1991; Hao, et al., 1993). The
mouse Igf-2 and HI9 are very similar to the human homologs.
Sequence as well as the direction of the imprints imposed on
these genes have been conserved (Rotwein and Hall, 1990;
DeChiara et al., 1991; Bartolomei et al., 1991; Brannan et al.,
1990; Glaser et al., 1989; Ohlsson et al., 1993; Giannoukakis et
al., 1993; Zhang et al., 1993). In the human, Igf-2 and H I9 are
believed to be involved in the genesis of the
Beckwith-Weidemann syndrome (Ohlsson et al., 1993). Loss of
imprinting resulting in biallelic expression of both Igf-2 and
H19 has been observed in Wilms' tumor of the kidney which is
often associated with the Beckwith-Weidemann syndrome
(Rainer et al., 1993; Ogawa et al., 1993). Biallelic expression of
either Igf-2 or H I9 may result in the promotion of cell growth
and lead to the formation of tumors.
It has been hypothesized that the imprints of Igf-2 and H 1 9
may be mechanistically related (Sasaki et al., 1992; Surani et
26
al., 1993; Bartolomei et al., 1993). The similar spatio-temporal
expression patterns of Igf-2 and H19 during mouse embryonic
development (Lee et al., 1990; Poirier et al., 1991) has lead to
the hypothesis that these two genes are coordinately regulated.
The reciprocal imprinting of these genes may have evolved to
ensure their coordinate expression. Coordinate expression of
Igf-2 and H19 may help to control normal cell growth since
Igf-2 is thought to promote cell growth and H I9 is thought to
inhibit cell growth (Steenman et al., 1994). Somatic cell
cultures which are maternally disomic for the Igf-2/HI 9
chromosomal region may be an important in vitro mouse
model system to study the nature of the imprints as well as the
regulation of these genes and may yield information as to how
these genes control cell growth.
Therefore, a better understanding of the genomic imprinting
process is important to elucidate the molecular and genetic
mechanisms that control cell growth and differentiation and
will therefore enhance our knowledge concerning human
genetic disorders thought to involve the differential imprinting
of alleles.
27
CHAPTER 2
Establishment of Mouse Somatic Cell Cultures
from Embryos with Maternal Uniparental
Disomy for Distal Chromosome 7
INTRODUCTION
Genetic complementation studies have shown that certain
autosomal regions must be represented by both parental
genomes for normal development in mice (Cattanach and
Beechey, 1990; Cattanach and Kirk, 1985; Searle and Beechey,
1990). For example, both the maternal and paternal copies of
distal chromosome 7 are required for proper development of
the mouse conceptus because of differential epigenetic
modification which affects the expression of some homologous
alleles. At least two genes which are located on this region of
chromosome 7 can be examined for the molecular basis of
genomic imprinting; the paternal copy of the insulin-like
growth factor 2 gene (Igf-2) and the maternal copy of H 19 are
preferentially expressed (Bartolomei et al., 1991; DeChiara et
al., 1991; Ferguson-Smith et al., 1991).
The Igf-2 product functions as an important growth factor
during early embryonic development, and the 67 amino acid
28
protein is synthesized primarily during the fetal and neonatal
periods (Adams et al., 1983). The gene is expressed during
preim plantation developm ent (Rappolee, 1992) and
abundantly in both the extra-embryonic and embryonic
tissues, for example, in cells of mesodermal derivatives
including skeletal muscle and chondroblasts (Lee et al., 1990).
Inactivation of the paternal Igf-2 by homologous
recombination in mice causes animals to be growth-deficient
(DeChiara et al., 1990). Chimeric embryos containing cells with
a paternal duplication-maternal deficiency of chromosome 7
(PatDi7), and therefore an expected twofold increase in the
level of the Igf-2 product, show a marked enhancement of
embryonic growth (Ferguson-Smith et al., 1991). Hence, the
Igf-2 product is an important embryonic mitogen.
The genomic structure of mouse Igf-2 is complex, with three
distinct promoters, one of which is a CpG island, distributed
over approximately 5 kb of DNA. Three transcripts which
contain independent leader exons but contain a common
3,045-nucleotide 3'-untranslated region are produced as a
result of differential promoter usage and RNA splicing (Rotwein
and Hall, 1990). At least two or three of these major
transcripts have been detected in embryonic tissues (Ikejiri et
al., 1991; Rotwein and Hall, 1990; Stempien et al., 1986). There
is also evidence for antisense transcripts from the Igf-2 locus
29
in the chicken (Taylor et al., 1991) as well as in the mouse
(Rivkin et al., 1993). The complexity of igf-2 implies that its
expression is subject to multiple controls. Presumably,
repression of the maternal Igf-2 allele must be achieved by a
mechanism capable of overriding any of the potential
activating signals for any of the promoters.
It is possible to obtain mice with a duplication of the
maternal chromosome 7 harboring Igf-2 without the paternal
complement by genetic manipulation ([MatDi7]; Cattanach and
Kirk, 1985; Searle and Beechey, 1990), and such embryos do
not express RNA for the growth factor (Ferguson-Smith, et al.,
1991). The mechanism by which the maternally derived Igf-2
allele is repressed is not known, but the chromatin
configuration and methylation status of the repressed Igf-2
allele are potentially active and indistinguishable from those of
the active paternal gene (Sasaki, et al., 1992). In this study,
fibroblast cell cultures from MatDi7 embryos and their normal
littermates were established and examined to determine the
stability of repression of Igf-2 in culture. Igf-2 was not
expressed in the MatDi7 cells, except very infrequently when
there was spontaneous reactivation (see Chapter 3).
Methylation analysis of Igf-2 also showed that extensive
methylation of the gene had not occurred during the
immortalization process. The MatDi7 cell line, therefore, is a
30
useful and amenable system to complement in vivo studies on
the molecular basis of imprinting.
MATERIALS AND M ETHODS
C ells
Normal and MatDi7 embryos were obtained at day 15 of
gestation as previously described (Ferguson-Smith et al.( 1991).
MatDi7 embryos were distinguished by a lack of eye pigment
indicated by an albino locus (c) located on distal chromosome 7.
One embryo of each type was decapitated and degutted, and
the carcasses were minced between scalpel blades, washed
with phosphate-buffered saline (PBS) and trypsinized for 30
min at 37°C with 0.05% trypsin dissolved in PBS (1:250; Difco,
Detroit, Mich.). The tissue was further disaggregated by
vigorous pipetting, and the cells were pelleted by
centrifugation. The cell pellets were resuspended in 25 ml of
Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10%
bovine calf serum, penicillin and streptomycin. The cells from
each embryo were plated into two T75 tissue culture flasks
and routinely passaged at a split ratio of approximately 1 to 4
by trypsinization when in late exponential phase of growth. In
later experiments, the cells were cultured in Eagle’s Minimum
Essential Medium (MEM) containing 10% heat-inactivated fetal
calf serum (Tissue Culture Biologicals), penicillin (100 U/ml)
31
and streptomycin (100 ug/ml) (GIBCO). (Initial establishment
of somatic cell cultures was performed by P. A. Jones.
Secondary cultures analyzed in Chapter 3 were derived by P. L.
Eversole-Cire). Cells of the C3H10T1/2 Cl 8 (10T1/2; Reznikoff
et al.. 1973A) and MCA Cl 15 Cl (MCA; Reznikoff et al.. 1973B)
lines were grown in Eagle’s basal medium supplemented with
10% fetal calf serum, penicillin and streptomycin.
Isolation of Subclones
Ring clones (Normal Cl 1, Normal Cl 4, and MatDi7 Cl 4) were
isolated from normal and MatDi7 cells by the following
procedure. Normal cells at passage 24 and MatDi7 cells at
passage 25 were seeded at 104/100-mm diameter dish and
10 3/ 100 mm-dish, respectively. Ring clones were selected
after discreet colonies had formed by placing a glass ring which
had been dipped in sterile vacuum grease around the colony.
Cells within the ring were trypsinized in a small volume of
trypsin and resuspended in medium. Cells were seeded in a
35-mm dish after centrifugation and resuspension in fresh
medium. Subclones were then passaged and amplified. Stocks
were stored in 10% dimethylsulfoxide and stored in liquid
nitrogen.
32
Isolation of Nucleic Acids
Total cellular RNA was isolated from growing or stationary
cultures of cells by the method of Chomczynski and Sacchi
(1987). Usually one or two 100-mm dishes of cells were used
per condition and 10 ug of RNA per lane was analyzed on 1%
formaldehyde gels, transferred to nylon membranes
(Amersham, Hibond N+) then UV cross-linked (Church and
Gilbert, 1984) and hybridized to 32P-radiolabelled riboprobes
(Maniatis, et al., 1982). Care was taken that cultures were in
the same growth phase for comparative studies. DNA was
isolated from cells in late log phase of growth, dissolved in TE
(10 mM TRIS-HC1 [pH 8.0], 1 mM EDTA [pH 8.0]), and digested
with various restriction enzymes in accordance with the
manufacturers' (Amersham and Boehringer Mannheim)
recommendations. Digestions were incubated overnight at the
recommended temperature for each enzyme, either 25°C or
37°C, with 4 to 5 U of each enzyme per ug of DNA in a reaction
volume of 400 ul. Restricted DNA was purified by phenol
extraction, precipitated with ethanol, electrophoresed on 1%
agarose gels, blotted onto nylon membranes (Amersham,
Hibond N+ or Dupont GeneScreen Plus) and cross-linked by UV
(Church and Gilbert, 1984).
33
P robes
A mouse Igf-2 genomic clone was kindly provided by Dr,
Peter Rotwein, Washington University, St. Louis, Mo. For a map
of the clone see Fig. 1. Probes B2 (Bam H l, 2.6 kb), C (Bam H l-
Bglll 1.5 kb), D {Sall-BstEll 1.7 kb), E (BamHl, 2.3 kb),
F (BsfEII, 1.0 kb), H (BamHl, 2.4 kb), and I (BamHl, 2.3 kb)
were subcloned into pBluescript II KS+ (Stratagene, La Jolla,
Calif). Probes used for the analysis of DNA methylation
patterns were excised from the plasmid, gel purified and
labeled by the random primer method of Feinberg and
Vogelstein (1984). An antisense riboprobe was prepared for
probe I by cutting the 5' end of the insert with K pn I and then
filling the resultant overhang with T4 DNA polymerase. The
riboprobe was then synthesized using the kit provided by
Stratagene and used as a probe for Northern (RNA) analysis.
RESULTS
Growth of Cultures
Cultures of cells derived from normal and MatDi7 embryo
littermates were passaged when approximately 80% confluent,
and stock cultures cryopreserved in 10% dimethlysulfoxide at
various passages. Details of growth kinetics were not kept
routinely, but in general it appeared that cells derived from
34
normal embryos had a shorter doubling time than did cells
from MatDi7 embryos until later stages of cell culture. The
growth of both types of cells slowed down at about passage 12
after explant and subsequently accelerated, presumably as
immortalized cells began to dominate the cultures. It was
possible to isolate subclones of cells after passage 24 by plating
the cells at low density (approximately 1,000/100-mm dish)
and ring isolating individual colonies of rapidly growing cells.
Genomic Structure of Igf-2
The genomic structure of Igf-2 y which spans over 12 kb of
genomic DNA, is illustrated in Fig. 2-1. The gene is composed
of six exons, with exons 1 through 3 encoding different
5'-untranslated regions which are transcribed by three
different promoters, one of which (promoter 2) is a CpG island.
Three Igf-2 transcripts are produced which contain distinct
leader sequences but the same coding sequence and
3'-untranslated region as a result of differential promoter
usage and RNA processing (Ikejiri et al., 1991; Rotwein and
Hall, 1990). Since the gene is transcribed by all three
promoters, the imprinting event which silences the maternal
allele must repress transcription from all three promoters.
Figure 2-1 also shows the locations and sizes of the probes
used in the methylation analysis of Igf-2 . For the expression
35
Figure 2-1. Genomic structure of the mouse Igf-2 and
summary of methylatlon analysis of lgf~2
A. Genomic structure of the mouse lgf-2. Exons are shown as
solid black boxes. The three promoters of the gene (PI, P2, and
P3) are indicated, as are the major transcriptional initiation
sites (arrows). The translational initiation site (ATG), the
termination codon (TGA), and the major polyadenylation site
(poly A) are also shown. The probes used in this study and
their locations are shown as stippled boxes (see Materials and
Methods for additional probe information). The positions of
and distances between the BamYW sites are shown. The
restriction sites of the methylation-sensitive enzymes used in
the methylation analysis are depicted as follows: Hhal (Hh),
Hpall (Hp), Sail (Sa), Smal (Sm), Xho\ (Xh). The positions of
most of the restriction sites were previously published (Sasaki
et al, 1992). B. Summary of methylation analysis of lgf-2 . An
open circle indicates an unmethylated CpG site, and a closed
circle indicates a fully methylated site. Partial methylation is
indicated by either a half-closed or a quarter-closed circle. The
methylation status of sites without circles could not be
determined by the analysis used here.
36
A
Restriction
Sites
K jt h ;
B
l l i l - W V
CO
studies using Northern analysis, an antisense RNA probe (I)
was used since this sequence hybridizes to all RNA transcripts
containing the lgf-2 coding sequences.
Expression of lg f-2
Cell culturing of fibroblasts derived from normal and MatDi7
embryos was performed to study the regulation of lgf-2
expression. Expression of lgf-2 in low-passage cultures of
fibroblasts derived from normal and MatDi7 embryo
littermates is shown in Fig. 2-2. (Northern analysis was
performed by P. A. Jones). Three transcripts hybridizing to
probe I were observed in confluent second-passage normal
cells (lane 1). The highest-molecular-weight transcript
corresponds in size to an mRNA containing exon 2, whereas the
predominant species, which migrates slightly faster than 28S
RNA, corresponds to an exon 3-containing mRNA (Rotwein and
Hall, 1990). The low-molecular-weight (approximately 1.8-kb)
transcript may represent a cleavage product containing the 3'
end of the lgf-2 message (Meinsma et al., 1992). None of these
transcripts were present in RNA obtained from cells derived
from MatDi7 embryos (lane 2). Figure 2-2 also shows that the
level of RNA for lgf-2 in normal cells was growth-regulated,
since transcripts were not easily visible in actively dividing
fourth-passage cells (lane 3) but were clearly visible in
38
Figure 2-2. Expression of lgf-2 in low-passage
cultured cells.
Total cellular RNA was extracted from stationary-phase (lanes
1, 2, 5, and 6) or logarithmic-phase (lanes 3 and 4) cells, and
10 ug were analyzed by Northern analysis using an antisense
RNA probe to probe (I). Lanes 1 and 2 contained RNAs
extracted from second-passage cells derived from normal and
MatDi7 embryos, respectively. Lanes 3 and 5 contained RNAs
from fourth-passage normal cells, and lanes 4 and 6 contained
RNAs from fourth-passage MatDi7 cells. The presence of
equivalent amounts of RNA in the lanes was verified by
examination of the intensities of ethidium bromide staining of
28S and 18S rRNA bands (lower panel).
39
1 2 3 4 5 6
28S —
18S —
28S —
18S —
• * * -
40
confluent cultures (lane 5). There also appeared to be a
decrease in the overall level of expression with increasing time
in culture (compare lanes 1 and 5). However, the imprint was
apparently stable since cells derived from MatDi7 embryos did
not express detectable levels of lgf-2 mRNA under any growth
conditions (lanes 4 and 6).
The extent of lgf-2 mRNA expression was also examined in
stationary-phase cells at passage 17 which had been
independently derived from separate embryos (Fig. 2-3).
(Northern analysis was performed by F. A. Gonzales). At the
higher cell passage, lgf-2 was still expressed by the normal
cells whereas the cells derived from MatDi7 embryos did not
express lgf-2 (lanes 1 and 2).
In normal passage 39 cells, which had presumably
undergone immortalization, lgf-2 expression was also observed
(lane 3). The transcript which corresponds to an mRNA
containing exon 2, however, was present at a lower level in
RNA isolated from higher-passage cells. Therefore, continued
culturing of the cells derived from the normal embryos
appeared to affect the activities of the different promoters to
various degrees. The overall level of lgf-2 expression also
decreased upon prolonged cell culturing, lgf-2 expression was
not observed in cells derived from MatDi7 embryos at any
41
Figure 2-3. Expression of lgf-2 in high-passage
cells.
Northern blot analysis of total cellular RNA isolated from
stationary-phase normal (lanes 1 and 3) or MatDi7 (lanes 2 and
4) cells in passage 17 (lanes 1 and 2) or 39 (lanes 3 and 4).
Lanes 5 to 7 contained total RNAs extracted from ring clones
Normal Cl 1. Normal Cl 4, and MatDi7 Cl 4, respectively. RNAs
isolated from C3H10T1/2 Cl 8 (lane 8) and MCA Cl 15 Cl (lane
9) cells were also analyzed. The probe used was antisense
probe I. The migration positions of the 28S and 18S rRNAs, are
indicated, and the intensities of ethidium bromide staining of
these bands were used to verify the presence of equivalent
amounts of RNA (lower panel).
42
I 2 3 4 5 ft 7 X ’>
passage analyzed (lanes 2 and 4). Therefore, expression of
lgf-2 remained repressed in fibroblast cultures derived from
MatDi7 embryos upon prolonged culturing.
Igf-2 expression was also analyzed in subclones isolated
from the mass culture of cells derived from normal embryos at
passage 24 and from cells derived from MatDi7 embryos at
passage 25. One subclone isolated from normal cells,
Normal Cl 1, was found to express lgf-2 whereas a second
subclone, Normal Cl 4, did not express lgf-2 (Fig. 2-3, lanes 5
and 6). Therefore, it was possible to isolate subclones from the
mass cultures which had lost the ability to express lgf-2. The
presence of cells which did or did not express lgf-2 within the
mass culture of normal cells may explain the decrease in the
overall level of lgf-2 expression upon continued culturing of
cells isolated from normal embryos. However, an insufficient
number of clones was examined to determine the frequency
with which extinction occurred in normal cells. A subclone
isolated from MatDi7 embryos, MatDi7 Cl 4, did not express
lgf-2 (lane 7). Hybrids formed between mRNA sequences and
the antisense riboprobe (probe I) were shown to be specific for
lgf-2 since these hybrids survived treatment with RNase A.
Hybrids formed with a sense riboprobe, however, were found
to result from nonspecific hybridizations, since none of these
hybrids survived RNase A treatment, and therefore no
44
evidence for antisense transcripts in this region was found
(data not shown).
The pattern of tgf-2 expression was also determined for the
C3H10T1/2 Cl 8 line (10T1/2) of mouse embryo fibroblasts
(Reznikoff et al.t 1973A) and an oncogenic derivative of this
cell line, MCA Cl 15 Cl (MCA) (Reznikoff et al., 1973B), which
had been chemically transformed by the carcinogen
3-methylcholanthrene (Fig. 2-3). 10T1/2, but not the MCA,
cells expressed lgf-2 RNA in the stationary growth phase
(lanes 8 and 9); however the higher-molecular-weight
transcript which contains exon 2 was not present in the
immortal 10T1/2 cell line. Other experiments (data not shown)
confirmed that the gene was not expressed in growing cultures
of 10T1/2 or MCA cells, suggesting that the growth-regulated
expression of lgf-2 was similar to that seen in the newly
isolated normal cells described here.
Methylation of lgf-2
The in vivo expression pattern of certain genes is thought to
be altered by aberrant methylation during in vitro cell
culturing (Jones et al., 1990). Previously, the in vivo
methylation pattern of lgf-2 was analyzed by using DNAs
isolated from normal and MatDi7 embryos at e l 5 (Sasaki et al.,
45
1992). None of the sites which were analyzed in the promoter
regions or the coding region of the gene showed parental
gene-specific differences in methylation status. Parental
gene-specific methylation differences were, however,
discovered several kilobases upstream of the lgf-2 promoter
region. An extensive methylation analysis was performed on
two clones derived from normal and MatDi7 cells to determine
whether aberrant methylation of lgf-2 had occurred during the
in vitro culturing of the MatDi7 cells. For the methylation
analysis, DNAs isolated from subclones Normal Cl 1 and
MatDi7 Cl 4 were digested with Bam\\\ to generate
specific-size fragments, then double digested with a
methylation-sensitive restriction enzyme (Hha\, Hpall, Mspl,
Sal I, Sma I, or Xho\). The probes used in this analysis are
indicated in Fig. 2-1, and the Southern blot analyses are given
in Figs. 2-4 through 2-9. The methylation status of particular
sites of the gene were determined by comparing the band sizes
generated in the samples restricted with Bam HI and a
methylation-sensitive enzyme with the band size produced for
S am HI digestion alone. A summary of the methylation results
is presented in Fig. 2-1. (Some of the methylation analyses
were performed by F. A. Gonzales).
The first area analyzed included most of promoter 1 of lgf-2
and the region upstream. The 2.6 kb Bam HI fragment which
4 6
hybridized to probe B2 was completely cleaved by HpaU,
Mspl, and S m a I generating a 2.1 kb and a 0.S4 kb fragment in
DNA isolated from Normal Cl 1 cells (Fig. 2-4; lanes 2, 3, and 4).
This result shows that the Sml site was unmethylated in these
cells. However, in DNA isolated from MatDi7 Cl 4 cells neither
HpaU nor Sma I cleaved the 2.6 kb BamHI fragment indicating
that the site was fully methylated (lanes 6 and 8). To
determine if the Sml site which appeared to be differentially
methylated in the normal and MatDi7 cells was also
differentially methylated in vivo, methylation of this site was
analyzed in DNA extracted from normal and MatDi7 embryos at
e l 5 (Fig. 2-5, lanes 13 and 14). This site was found to be
nearly completely unmethylated in both the normal and
MatDi7 embryonic DNA and completely unmethylated in three
additional MatDi7 subclones, one of which spontaneously
expressed lgf-2 (MatDi7 1-la; Fig. 2-5, lanes 1 through 12).
Therefore, the methylation differences observed in the DNA
isolated from the Normal Cl 1 and MatDi7 Cl 4 cells resulted
from growth in culture and did not reflect the methylation
pattern in vivo.
Probe C was used to analyze the methylation status of the
5.6 kb Bam HI fragment which contains the remaining portion
of PI as well as promoters 2 and 3 (Fig. 2-6). Digestion of
DNAs isolated from both Normal Cl 1 and MatDi7 Cl 4 with
47
Figure 2-4. Methylation analysis of the igf-2
promoter I and the region 5' in normal and MatD17
cells.
The autoradiograph shown demonstrates the methylation
status of lgf~2 in subclones Normal Cl 1 and MatDi7 Cl 4.
Isolated DNAs were restricted with Bam HI and either the
methylation-sensitive enzyme HpaU or Sma I. Mspl, an
enzyme insensitive to the methylation status of the internal
cytosine yet sensitive to the external cytosine residue of the
Hpa II sequence (CCGG) was used as a control. The probe used
in the Southern analysis was Probe B2 (see Fig. 2-1) and the
result of this and other blots (not shown) are summarized in
Fig. 2-1.
46
ro ro
^ b>
I I
f t
t
t
*
C O
cr
Bam HI
Bam Hl/Hpa II
Bam Hl/Msp I
Bam Hl/Sma I
Bam HI
Bam Hl/Hpa II
Bam Hl/Msp I
Bam Hl/Sma I
z
o
3
»
o 7
— o
— cr
C D
a
5
" ■ J
O
G O
r o
Figure 2-5. Methylation analysis of the lgf-2
promoter 1 region In MatDI7 subclones and In
embryonic DNA.
Methylation status of lgf-2 promoter 1 region was determined
for the MatDi7 1-la, MatDi7 l-6b and MatDi7 4-2c subclones
and for normal (N) and MatDi7 (Di) embryonic DNA. DNAs
were isolated and digested with Bam\\\ and either the
methylation-sensitive enzyme Sma I or Hpa II. Mspl was used
as a control. Southern blot analysis was performed using Probe
B2 (see Fig. 2-1) as a radiolabeled probe and the
autoradiograph is shown. The sizes of the fragments generated
are indicated.
5 0
w
•
t
C J 1
* ' ^ t ' «
4
o>
• - v t
m
C O
"• r**% a m
w
o
♦ * « .
A
*
f O
•
«
J J I
i
-* -» fo f O
tab o
o w
b f r
1
1
r
u
*
1
■
tx
Bam H I
£
a
Bam Hl/Sma 1
5
Bam Hl/Hpa I I
r
Bam HIM ap 1
a
Bam H I
£
a
Bam Hl/Sma 1
5
xi
Bam H l/H pa I I
i
Bam H IM ap 1
c r
Bam H I
£
•
Bam Hl/Sma 1 5
■ n j
Bam H l/H pa I I
*
to
Bam H IM ap 1
o
Bam Hl/Sma I I z
Bam Hl/Sma I | p
Figure 2-6. Methylation analysis of the lg f-2
promoter 1 and the region 3' in normal and MatDi7
c e lls.
Methylation status of the lgf-2 promoter 1 and 3' region was
determined for Normal Cl 1 and MatDi7 Cl 4. DNAs were
extracted and digested with BamHX and one of the following
m ethylation-sensitive enzymes; Hpa II, Mspl, Salh Sma I, or
Xhol. The probe used in the Southern analysis was Probe C
(see Fig. 2-1) and the autoradiograph is shown. The sizes of
the fragments generated are indicated and the results are
summarized in Fig. 2-1.
52
o o o o - * ■ ro ro
C O U ) ■ ni io 00 o W
U 1 ro o o o
1 1 1 1 1 t 1
ro
( ■ »
*
tn
0 1
O B
ID
*
• »
N
U 1
co
c n
b>
Bam H I
Bam Hl/Hpa II
Bam Hl/Msp I
Bam Hl/Sal I
Bam Hl/Sma I
Bam Hl/Xho I
Bam HI
Bam Hl/Hpa II
Bam Hl/Msp I
Bam Hl/Sal I
Bam Hl/Sma I
Bam Hl/Xho I
Z
o
3
»
O
-0
o
cr
w
s °
Q J
6
-si
o
HpaU yielded a 0.52 and a 0.35 kb band. The same banding
pattern was obtained for M spl digestion indicating that the
HpaU sites within this region were unmethylated (lanes 2, 8
and 3, 9, respectively). Sma I digestion yielded bands of 0.90,
0.78, 0.52, and 0.35 kb for Normal Cl 1 DNA. The same banding
pattern was obtained for MatDi7 DNA with the exception of the
0.90 kb band (lanes 5 and 11). Therefore, all of the Sma I sites
within this area were unmethylated (Sm2 to Sm5) in the
MatDi7 cells. One of the Sma I sites, most probably Sm4,
appeared to be slightly methylated in the normal cells as
evidenced by the appearance of the 0.90 kb band. Sal I
cleavage gave rise to a 2.5 and a 2.0 kb band for Normal Cl 1
DNA indicating partial methylation of the Sal site and no
methylation of the Sa2 site. Sail cleavage of MatDi7 Cl 4 DNA,
however, yielded predominantly a 2.0 kb band with a faint
band at 2.5 kb indicating that the Sal site was less methylated
in these cells compared to the normal cells and the Sa2 site was
unmethylated (compare lanes 4 and 10). X h o I digestion was
complete for DNA samples extracted from both cell lines,
generating only a 1.8 kb band, suggesting that the Xhl site was
unmethylated in both the normal and MatDi7 cell lines (lanes 6
and 12).
The methylation status of some of the sites analyzed with
probe C was confirmed by hybridizing a similar blot with
5 4
probe E (Fig. 2-7). HpaU digestion for both Normal Cl 1 and
MatDi7 Cl 4 DNA yielded bands of 0.78 and 0.60 kb which was
the same pattern generated with Mspl digestion (lanes 2, 3 and
8, 9). Therefore, all of the HpaU sites within this region were
unmethylated in both cell lines. Smal digestion of DNA isolated
from both cell lines gave rise to a 3.1 kb band suggesting that
the Sm6 site was unmethylated (lanes 5 and 11). Digestion
with Sa/I and X ho I were also complete, generating a 3.0 and a
3.5 kb band, respectively, in DNAs extracted from both the
normal and MatDi7 cells (lanes 4, 10 and 6, 12). Therefore,
both the Sa2 and Xhl sites were unmethylated, thus
confirming results obtained with probe C.
DNA was digested with BamHl and Hhal and hybridized
with probe F to analyze the area surrounding promoter 3,
(Fig. 2-8). The Hhal digest for Normal Cl 1 DNA gave rise to
2.4, 1.9, and 1.6 kb bands indicating that the Hhl site was
unmethylated, Hh2 and Hh4 sites were methylated, and Hh3
and Hh5 sites were partially methylated (lane 2). Digestion of
MatDi7 Cl 4 DNA with Hhal yielded bands of 1.9, 1.6 and 1.3 kb
which indicated that the Hh3 site was unmethylated with
partial methylation at sites Hh4 and Hh5 (lane 4). Since the
methylation patterns within this region were different
between the two cell lines, the same analysis was performed
on normal and MatDi7 el5 embryonic DNA (data not shown).
55
Figure 2-7. Methylation analysis of the lgf-2
promoter 2 region in normal and MatDi7 cells.
The autoradiograph shown demonstrates the methylation
status of the lgf-2 promoter 2 region in the subclones.
Normal Cl 1 and MatDi7 Cl 4. Isolated DNAs were digested
with BamHl and one of the methylation-sensitive enzymes
Hpa II, Mspl, Sail, Sma I, or Xho I. Southern analysis was
performed using Probe E (see Fig. 2-1) and the autoradiograph
is shown. The sizes of the fragments generated are indicated
and the results are summarized in Fig. 2-1.
56
< • >
Jk
W i
0 3
■ k j
O D
V
C D -si
O 0
I I
ro
tn
" •J
U U U U1
o j in b)
"ii r i
t
BamHl
Bam Hl/Hpa II
Z
o
Bam Hl/Msp 1
3
tt
Bam Hl/Sal 1
Bam Hl/Sma 1
O
Bam Hl/Xho 1
_*
Bam H I
*w
Bam Hl/Hpa I I
2
O J
Bam HlAtep 1 6
Bam Hl/Sal 1
Bam Hl/Sma 1
'-■ j
o
Bam Hl/Xho 1
A
- * 0
o
cr
®
m
Figure 2-8. Methylation analysis of the l g f ' 2
promoter 3 region in normal and MatDt7 cells.
Methylation status of the lgf-2 promoter 3 region was
determined for the subclones Normal Cl 1 and MatDi7 Cl 4.
DNAs were isolated and restricted with BamHl and the
m ethylation-sensitive enzyme Hhal. The probe used in the
Southern blot analysis was Probe F (see Fig. 2-1). The sizes of
the fragments generated are indicated and the results for the
subclones are summarized in Fig. 2-1.
58
ro
co
r 1 , r 1 , r 1 , r °
00 O) b
1 I I I
• I
I
tn
b>
I
*
• •
cn
10
7T
O'
, Bam HI
Bam Hl/Hha I
Bam HI
| Bam Hl/Hha I
0 5
1 3
C D
O to
* 9 .
*v|
o
O'
< D
The methylation pattern was found to be the same for DNA
isolated from normal and MatDi7 el5 embryos and was
identical to the pattern for the MatDi7 Cl 4 DNA.
Probe H, which hybridizes to a 2.4 kb BamHl fragment, was
used to analyze the methylation status of the coding region of
lgf-2 (Fig 2-9). HpaU digestion of the Normal Cl 1 DNA gave
rise to 1.4, 1.2, 0.92, 0.58, and 0.47 kb bands. Since Mspl
digestion of the normal DNA only yielded a 0.92, 0.58 and a
0.47 kb band, there was partial methylation at a subset of the
HpaW sites which resulted in the formation of the larger-sized
fragments (lanes 2 and 3). HpaW digestion of the MatDi7 Cl 4
DNA yielded the same pattern as the Mspl digest with the
exception of a faint band at 1.2 kb which suggested that only
one HpaU site was slightly methylated (lanes 8 and 9). Sma I
digestion for normal DNA resulted in predominantly a 2.0 kb
band with a faint band at 2.4 kb (lane 5). Therefore, the Sm7
site was partially methylated, consistent with the Hpa II
results. 5m al digestion of the MatDi7 DNA resulted in only a
2.0 kb band indicating that the Sm7 site was unmethylated in
this sample (lane 11). Sal I digestion yielded a 2.4 kb band for
both normal and MatDi7 DNA (lanes 4 and 10). In the area of
known sequence, there do not appear to be any Sal I sites in
this region, which was confirmed by restriction mapping and
this analysis. Xhol digestion gave rise to a 1.4 and a 0.92 kb
60
Figure 2-9. Methylatlon analysis of the Igf-2 coding
region in normal and MatD17 cells.
Methylation status of the coding region of Igf-2 was
determined for the subclones Normal Cl 1 and MatDi7 Cl 4.
DNAs were isolated and digested with Bam HI and with one of
the methylation-sensitive enzymes Hpa II, Mspl, Sal I, Sma I, or
Xhol. Southern blot analysis was performed using Probe H (see
Fig. 2-1) as a radiolabeled probe and the autoradiograph is
shown. The sizes of the fragments generated are indicated and
the results are summarized in Fig. 2-1.
61
ro
w
*
w
o >
0 0
« D
O O O
-v l
I
tn
a>
to to ro a
IV ) 4^
I 1 l I I
i- o
a>
ro
rv) ro
© -u
I I
?r
O
Bam HI
Bam Hl/Hpa I I
Bam Hl/Msp I
Bam Hl/Sal I
Bam Hl/Sma I
Bam Hl/Xho I
Bam HI
Bam Hl/Hpa II
Bam Hl/Msp I
Bam Hl/Sal I
Bam Hl/Sma I
BamHVXho I
to
g
o
o
cr
a >
band for both the normal and MatDi7 DNA samples indicating
that the Xh2 was unmethylated in both cell lines (lanes 6 and
12).
Probe I which hybridizes to a 2.3 kb Bam HI fragment was
used to determine the methylation status of the 3' untranslated
region of Igf-2 gene (data not shown). Analysis after HpaW
digestion was difficult to interpret, however Xho\ gave rise to
two bands in the normal DNA sample but failed to cleave the
MatDi7 DNA sample. This result indicated partial methylation
of the Xh3 site in the normal cells and complete methylation in
MatDi7 cells. (Some of the above methylation analyses were
performed by F. A. Gonzales).
Overall, igf-2 was found to be largely unmethylated in the
normal and MatDi7 cell lines. Minor differences between the
methylation patterns of the gene in the normal and MatDi7
subclones did not correlate with the differential Igf-2
expression from the parental alleles in the animal, since these
differences were not present in DNAs isolated from normal and
MatDi7 embryos and were not present in other clones analyzed
(some data not shown). The methylation status of one site in
particular (Sm l) initially seemed to be consistent with the idea
that methylation plays a role in the control of Igf-2 expression
in culture since it was unmethylated in the Normal Cl 1
63
subclone and fully methylated in the MatDi7 Cl 4 subclone
(compare lanes 4 and 8 in Fig. 2-4). However, this site was
shown to be unmethylated in three other MatDi7 subclones,
one of which spontaneously expressed Igf-2, and in normal and
MatDi7 embryonic DNA as well (see Fig. 2-5). Although these
methylation differences did not correlate with the differential
expression of the Igf-2 alleles in vivo, it was possible that
differences in methylation at sites not examined were involved
in Igf-2 regulation in cell culture.
Results from the methylation analyses of Igf-2 in normal
and MatDi7 cells indicated that the gene did undergo some
epigenetic modifications during the process of culturing;
however, extensive methylation of the maternal Igf-2 allele
did not occur. Methylation of Igf-2 was also analyzed in the
10T1/2 cell line and its transformed derivative MCA to
determine whether the lack of extensive methylation within
the CpG island of Igf-2 was specific to these newly derived cell
cultures. (Methylation analysis was performed by P. A. Jones).
Igf-2 contained significant methylation within and around the
CpG island, and several Sma 1 sites were partially methylated in
10T1/2 cells, as indicated by incomplete digestion of the
5.6-kb Bam HI fragment and the appearance of intermediate
sized fragments in the various digests (Fig. 2-10A). One Sma\
site, however, was unmethylated in the 10T1/2 cells, since a
64
Figure 2-10. igf~2 can become methylated In
perm anent cell lines.
DNA extracted from 10T1/2 or MCA cells was analyzed on
Southern blots hybridized with probe C (A) or D (B). Lanes 1
and 5 contained Bam HI (B) digests; the remaining lanes
contained double digests with BamH\ and
methylation-sensitive enzymes Sail (B+Sa), Sm a I (B+Sm), and
Xho\ (B+X), respectively.
65
kb
5 .6 -
5 .6 -
Probe C
10T1/2 M C A
i i i
B B B B B B
+ + + + + +
B Sa Sm X B Sa Sm X
Hlttii
2 .1 -
1.8 -
I .
i
B Probe 0
3 8 -
3 2 - * -
66
3.2-kb band was present in Sma\~BamH\ double digests when
they were hybridized to probe D (Fig. 2-10B, lane 3). In
contrast, all of the restriction sites which were examined in
DNA isolated from MCA cells were more than 90% methylated
since none of the methylation-sensitive enzymes were capable
of restricting the DNA (Fig. 2-10A and B; methylation analysis
was performed by P. A. Jones). Therefore, the lack of extensive
methylation of Igf-2 in MatDi7 cells did not result from an
inherent inability of the CpG island to become methylated.
DISCUSSION
This study demonstrated that Igf-2 is repressed in MatDi7
cell lines, but expressed in cell lines derived from normal
littermates, albeit at a reduced level relative to the freshly
explanted cultures. Although the molecular mechanism of
repression of the maternal Igf-2 gene is not known, it
apparently does not involve extensive methylation of the CpG
island within the gene or heterochromatinization of the locus
(Sasaki et al., 1992). Furthermore, the gene remains
transcriptionally silent in long-term somatic cell cultures
without extensive de novo methylation occurring during the
immortalization process. Cell cultures which appear to have
retained some of the essential features of control of an
67
imprinted gene should be a useful system to complement
in vivo studies on the molecular basis of imprinting.
The expression of the active allele of Igf-2 in normal cells
showed an increase during growth arrest. It was previously
reported that expression of Igf-2 exhibits a 25-fold increase
during the differentiation of mouse myoblasts in culture
(Tollefsen et al., 1989), which was attributed to cell
differentiation, but it is possible that the increase in expression
was a result of growth arrest since fused myoblasts are, by
definition, postmitotic. It has recently been reported that
human Igf-2 can be regulated by retinoic acid (Matsumoto, et
al., 1992) and the increase in expression induced by retinoids
corresponded to inhibition of DNA synthesis. Thus, while the
maternally inherited Igf-2 is kept silent by an unidentified
mechanism, the paternally inherited allele is either active or
inactive, depending on the physiological state of the cell.
Analysis of the methylation of Igf-2 showed that the gene
was considerably unmethylated in the normal and MatDi7 cell
lines at high passage so that extensive de novo methylation did
not occur in cultures. By contrast, methylation of the CpG
island in MCA tumorigenic cells was observed. The lack of
methylation of the maternal allele does not, therefore, appear
to result from an inherent inability of Igf-2 promoter 2 to
66
become methylated; however, extensive methylation of the CpG
island within this gene is not the mechanism used to maintain
the maternal allele of Igf-2 in a repressed state in culture.
Therefore, the mechanism for silencing Igf-2 differs from the
silencing mechanism of X inactivation (Pfeifer et al., 1990; Yen
et al., 1984). The small changes in methylation observed in the
ring clones also did not appear to be responsible for
differential expression of the Igf-2 allele in vivo or in vitro,
since the changes were not mirrored in the uncultured
embryos or in other clones. However it is possible that the
mechanism of Igf-2 repression may involve (i) changes in the
methylation status of possible key regulatory sites within the
gene not included in this analysis or perhaps (ii) changes in the
methylation status of important sites outside the gene. This
newly derived cell line provides an excellent system in which
the process by which endogenous genes become imprinted may
be further studied.
6 9
CHAPTER 3
Activation of an Imprinted Igf-2 Gene in Mouse
Somatic Cell Cultures
INTRODUCTION
The establishment and characterization of a somatic cell line
which apparently maintains the genomic imprint of an
endogenous gene, namely Igf-2, was described in Chapter 2.
Igf-2 was shown to be repressed in the cell lines derived from
MatDi7 embryos whereas it was expressed in cell lines derived
from normal littermates. Although the molecular mechanism
of the Igf-2 imprint is currently not known, it has recently
been shown that the imprinting event does not involve
extensive methylation of the CpG island within the gene or
heterochromatinization of the gene as evidenced by nuclease
sensitivity assays (Sasaki et al., 1992). Igf-2 was also found to
be in a transcriptionally competent state since a low level of
expression was observed (Sasaki et al., 1992). Therefore, the
mechanism which represses Igf-2 does not completely silence
the gene.
Studying the molecular mechanism of genomic imprinting is
complicated by the presence of both an active and an inactive
70
allele within the same cell. Therefore, a comparative analysis
of differential epigenetic modification is difficult in animals.
The establishment of a cell line which is maternally disomic for
distal chromosome 7 allows one to more easily study the
mechanism which represses Igf-2.
The expression pattern of many genes can be altered in
tissue culture upon treatment with specific agents. The
calcium ionophore A23187, for example, can induce the
expression of some genes by depleting intracellular calcium
stores (Drummond et al., 1987). Whereas sodium butyrate
alters gene expression patterns in tissue culture by a variety of
mechanisms including altered histone acetylation and changes
in chromatin conformation (Kruh, 1982). Gene activity may
also be affected by inhibition of DNA synthesis or by
incorporation into DNA, actions of the nucleoside analog
1-B-D-arabinofuranosylcytosine (Benedict and Jones, 1979).
S-Aza-cytidine is thought to alter methylation profiles and
possibly the chromatin structure of genes by inhibiting DNA
methyltransferase (Taylor and Jones, 1982).
5-Bromodeoxyuridine may alter gene expression patterns by
incorporation into the DNA resulting in an altered chromatin
structure (Tapscott et al., 1989). Induced expression of the
normally silent Igf-2 with an agent with a known mechanism
71
of action may yield insights into the nature of the mechanism
responsible for suppressing the maternal lgf-2.
Therefore, treatment with agents known to alter gene
expression patterns in culture were performed on fibroblast
cell cultures from MatDi7 embryos in an attempt to induce
expression of the repressed maternal lgf-2. Increased
expression of lgf-2 was only observed in MatDi7 cells treated
with 5-aza-cytidine or 5~bromodeoxyuridine. Therefore, only
treatments which may alter the methylation pattern and/or
chromatin conformation of Igf-2, two epigenetic modifications
which may be involved in the genomic imprinting process
(Surani et al., 1990), substantially increased expression of the
imprinted lgf-2 allele.
MATERIALS AND METHODS
Cell Culture
Conditions for cell culture were previously described in
Chapter 2.
Isolation of Subclones
Subclones derived from single cells were isolated from
normal and MatDi7 cells at passage 20. Cells in late log phase
were trypsinized, rinsed and resuspended in medium at a
72
dilution of 1.5 cells/25 ul. Aliquots (25 ul) of the cell
suspension were dispensed to the center of each well of a
24-well dish. Cells were allowed to attach, and then 0.5 ml of
conditioned medium was added to each well. Plates were
carefully examined under the microscope to determine which
wells contained single cells. Wells were screened for several
days to monitor which wells contained colonies which arose
from single cells. Subclones were then passaged and amplified.
Stocks were stored in 10% dimethylsulfoxide and stored in
liquid nitrogen. Isolation of ring clones (Normal Cl 1,
Normal Cl 4, and MatDi7 Cl 4) was previously described in
Chapter 2. The ring clone (MatDi7 Aza Cl L) was isolated from
MatDi7 cells which had been seeded at lt^/bO-mm dish and
treated 24 h after seeding with 3 x l0 '6 M 5-azacytidine
(5-aza-CR) for 24 h. A muscle-forming colony was isolated by
the ring cloning procedure described in Chapter 2.
Isolation of Nucleic Acids
Total cellular RNA was isolated from growing or stationary
cultures of cells by the method of Chomczynski and Sacchi
(1987) and as described in Chapter 2.
73
T reatm en ts
N ucleoside Analogs
Celts were plated at 0.2x106 to 1.0xl06/100-mm dish and
treated 24 h after seeding with the analog. The concentrations
used were 3 x 1 O'7 M and l.OxlO*6 M for
5-aza-2'-deoxycytidine (5-aza-CdR; Sigma), l.OxlO-6 M and
1.0x1 O'5 M for bromodeoxyuridine (BUdR; Sigma), and
l.OxlO 7, l.OxlO-6, and 1 . 0 x 1 0 - 5 m for
1-B-D-arabinofuranosylcytosine (ara-C; Sigma). The medium
was changed 24 h after addition of the analog and every three
days thereafter. RNA was isolated at 3 and/or 6 days after
treatment. All BUdR treatments were performed in a darkened
room, and the cells were washed twice with unsupplemented
medium to remove the analog. A time course for BUdR
activation was performed as described above, with RNA
samples being isolated at 1, 2, 3 and 6 days post-treatment.
Calcium lonophore
Cells were allowed to become confluent and were then
treated with 7 .4 x l0 '6 M ionophore which had been
resuspended in ethanol (A23187, [Calcimycin]; Sigma) for
either 3 or 15 h. RNA was isolated without any medium
changes. A control dish was treated with an equivalent volume
of ethanol.
74
Heat or Cold Shock
Cells were seeded into a T75 flask for each treatment and
allowed to become 80 to 90% confluent. The medium was
changed several hours prior to treatment, and the sealed flasks
were then preequilibrated in a 37°C water bath. Thermal
shock was achieved by submerging a flask in a water bath at
4°C or 43°C for 30 min. The flasks were then reequilibrated in
a 37°C water bath and returned to the incubator. RNA was
isolated after 12 h of incubation at 37°C.
Sodium Butyrate
Cells were plated at 1.0xl06/100-mm dish and treated 24 h
later with the sodium salt of n -butyric acid (Sigma) at a final
concentration of 10'3 M, 10-2 M, and 10*1 M. Medium was
changed 24 h after treatment and every three days thereafter.
RNA was isolated 6 days after treatment. A stock solution of
sodium butyrate was made up in medium and filtered through
a 0.45-um-pore-size filter for sterilization.
Reverse Transcription (RT)-Polymerase Chain Reaction
(PCR)
Total RNA was isolated and resuspended in
diethylpyrocarbonate-treated double-distilled water. Reverse
transcription of RNA into cDNA was performed as previously
75
described, with minor modifications (Horikoshi, et a)., 1992).
Purified RNA (1 to 5 ug) was added to the following
components in a total reaction volume of SO ul: 50 mM
Tris-HCl (pH 8.3), 75 mM KC1, 3 mM MgCl2, 1.0 U of RNasin
(Promega) per ul, 10 mM dithiothreitol (DTT), each
deoxynucleoside triphosphate (dNTP; Pharmacia) at ImM,
0.075 mg bovine serum albumin per ml, 0.5 optical density
unit of random hexamers (pd(N)6; Pharmacia) per ml, and 10 U
of Moloney murine leukemia virus reverse transcriptase
(MMLV RT; Bethesda Research Laboratories, Inc.; Gaithersburg,
Md.) per ul. The reaction mix was incubated at room
temperature for 10 min and then at 42°C for 45 min. Samples
were then heated to 90°C for 3 min to denature the
heterodimers and then quickly chilled to 0°C. The first-strand
synthesis was repeated to increase cDNA production by adding
an additional aliquot of MMLV RT (1.25 ul, 200 U/ul) and again
incubating the mixture at 42°C for 45 min. A final incubation
at 75°C for 10 min was performed to destroy the DNAse
activity of the MMLV RT. RT products were stored at either
-20°C or -80°C for long term storage.
An aliquot of RT products equivalent to 100 ng of RNA was
mixed with the following reagents in a total volume of 50 ul for
the PCR: Ix PCR buffer (50 mM KC1, 10 mM Tris-Cl, [pH 8.3],
1,5 mM MgCl2, 0.01% [wt/vol] gelatin, [Perkin Elmer Cetus],
76
each dNTP at 200 uM, the 5' and 3' primers at 1 uM). Samples
were heated to 94°C for 10 min prior to the addition of Taq
polymerase (2U; Perkin Elmer Cetus). An overlay of 50 ul of
sterile mineral oil was added to each sample, and the samples
were then subjected to the PCR for a specific number of cycles.
Each PCR cycle consisted of incubation periods of 1 min at 94°C,
30 s at 65°C, and 1 min at 72°C. The PCR products were
analyzed by electrophoresing a 20-ul aliquot on a 2 to 2.5%
agarose gel in lx TBE (0.089 M TRIS-borate, 0.003 M EDTA).
DNA was transferred to a nylon membrane (GeneScreen Plus;
Dupont) via alkaline transfer. Southern blot analysis was then
performed by hybridizing the filters with a radiolabeled or a
digoxigenin labeled probe (fragment H) in accordance to
m anufacturer's (Boehringer Mannheim) instructions.
Quantitation of PCR products was performed by scanning
autoradiographs or lumigrams with an LKB UltroScan XL laser
densitometer. PCR primers used were as follows:
OP1 (exon 3, 5 ) 5’AGCGGCCTCCTTACCCAACTTCAG3’
OP3 (exon 2, 5 ) 5 AGAAGGGCCTTCGCCCGCTGTT3
OP8 (exon 4-exon 5 junction, 3’)
5’ AAGGCCTGCTG AAGT AGAAGCCG3’
OP9 (actin, 5 ) 5 GCTGTGCTATGTTGCTCTAGACTTC3
OP10 (actin, 3’) 5 'CTC AGT A AC AGTCCGCCT AG A AGC3'
77
RESULTS
RT-PCR Assay
Once the somatic cell cultures were established and shown
to maintain the Igf-2 imprint it was next attempted to induce
Igf-2 expression in MatDi7 cells. Since Northern analysis was
not sensitive enough to detect the level of induced expression
in MatDi7 cells, a sensitive and quantitative RT-PCR assay
similar to an RT-PCR assay previously described (Sasaki et al.,
1992) was developed. Primers for the PCR portion of the assay
were designed to allow detection of the different Igf-2
messages transcribed from the different Igf-2 promoters
(Fig. 3-1). The amount of PCR product produced in this assay
was dependent on the number of cycles (Fig. 3-2). Assays
were carried out for a specific number of cycles at which there
appeared to be a linear relationship between the amount of a
PCR product produced and the number of input molecules of
Igf-2 template RNA (Fig. 3-3). Although Northern analysis did
not detect any Igf-2 transcripts for the MatDi7 cells, a low level
of expression of Igf-2 was observed with the RT-PCR assay at
higher cycle numbers (approximately 86- to 140-fold lower
than their normal counterparts). The levels of B-actin
expression were determined to control for equal amounts of
RNA analyzed for Igf-2 expression. Therefore, this assay
78
Figure 3-1. Schematic of the primer design for the
RT-PCR assay.
The positions of the oligonucleotide primers (OP) used in the
PCR are shown as arrows, OP3 is located within exon 2, and
OP1 is located within exon 3. OP8 spans the intron between
exons 4 and 5. PCR using OP3 and OP8 amplified a product
which corresponds to a P2-specific transcript, whereas
amplification using OP1 and OP8 produced a product which
corresponds to a P3-specific transcript. The stippled box
represents the probe used in the Southern blot analysis of the
RT-PCR products.
79
Exon 1
Pi
m
TGA
P2
00
o
Exon 2
P2
Exon 3
P3
Exon 4 Exon 5
ATG TGA
Exon 6
poly A
1
OP3 OP1 OPS
H
poly A
Spliced M e s s a g e PCR Primers PCR Product
poly A
E2 E4 E5 E6 OP3 OP8 246 bp
E3 E4 E5 E6 OP1 OP8 196 bp
Figure 3-2. Demonstration of RT-PCR assay with
respect to the number of PCR cycles.
Total RNA isolated from normal passage 21 (21°) cells was
reverse transcribed, and then 100 ng of template RNA was
subjected to an increasing number of PCR cycles. PCR was
performed with primers specific for P3 transcripts. The cycle
curve for the MatDi7 RNA sample shows that the low level of
expression of lgf-2 can be detected at a higher number of PCR
cycles. PCR products were analyzed by Southern blot analysis
using digoxigenin-labeled fragment H. Quantitation of the PCR
products was performed by densitometric scanning of the
displayed lumigram. The amplification curve demonstrates
that the PCR response increased with respect to the number of
PCR cycles, up to approximately 28 cycles. The amount of the
PCR product produced at 32 cycles was taken as 100%.
81
Response (Percent of Maximum)
00
ro
Figure 3-3. Demonstration of RT-PCR assay with
respect to the number of Input molecules.
Total RNA isolated from normal passage 21 (21°) cells was
reverse transcribed, and then 50, 100, and 200 ng of template
RNA were subjected to 26 cycles of PCR. PCR products were
quantitated as described in Fig. 3-2. The concentration curve
demonstrates that the PCR response was shown to be linear
with respect to the amount of the RNA template in the PCR.
83
(Parcant o f Maximum)
100
•o-
o o
RNA (ng) 50 100 200
Normal 21° P3
4 0 -
aso 100 aoo o
RNA (nanogram*)
84
allowed a quantitative comparison between the relative levels
°f J&/-2 expression in normal and MatDi7 cells.
Effects of Agents and Treatments Known to Alter Gene
E xpression Patterns
Cells derived from MatDi7 embryos were treated with
agents known to alter gene expression patterns in culture in an
attempt to induce expression of Igf-2 (Table 3-1). Treatments
with calcium ionophore, heat shock, cold shock, sodium
butyrate, or ara-C did not increase the low levels of expression
of lgf-2 by the cells. Addition of deoxycytidine to MatDi7 cells
also failed to induce Igf-2 expression. Treatment of cells with
calcium ionophore A23187 has been shown to induce gene
expression by depleting intracellular calcium stores
(Drummond et al., 1987), whereas environmental stress, such
as thermal shock, also affects the transcriptional activity of
genes in cultured cells (Kim and Lee, 1986;
Miiller-Taubenberger et al., 1988). Sodium butyrate alters the
expression patterns of many genes in tissue culture by a
variety of mechanisms, including altered histone acetylation
and changes in chromatin structure (Kruh, 1982). The
nucleoside analog ara-C may affect gene activity by inhibiting
DNA synthesis or by incorporation into DNA (Benedict and
Jones, 1979). Since none of these agents induced Igf-2
85
expression, it seems unlikely that mechanisms related to these
perturbations were responsible for suppressing the maternal
gene.
As indicated in Table 3-1 increased expression of lgf-2 was
only observed in cells treated with 5-aza-CdR or BUdR.
Inhibition of DNA methyltransferase by 5-aza-CdR alters
methylation patterns and possibly the chromatin conformation
of some genes in tissue culture, thereby affecting their
expression (Jones et al., 1990). BUdR is thought to alter gene
expression profiles by DNA incorporation, thus altering
chromatin structure (Tapscott et al., 1989). Therefore, only
treatments which could affect the methylation profile and/or
the chromatin structure of lgf-2, two epigenetic modifications
which may play roles in genomic imprinting (Surani et al.,
1990), substantially increased expression of the repressed
lgf-2 allele.
The mechanism of activation by 5-aza-CdR and BUdR was
explored in greater detail (Figs. 3-4 and 3-5). Treatment of
low- or high-passage mass cultures or subclones of MatDi7 cells
with 5-aza-CdR previously shown to induce the expression of
many genes in tissue culture (Jones, 1985) resulted in a
dose-dependent (two- to ninefold) increase in expression from
lgf-2 promoters 2 and 3 on the maternally derived gene
86
TABLE 3-1. Effects of treatments on lgf-2 expression
in MatDi7 cells
T reatm en t0
Fold increase in
expression^
P rom oter2 P rom oter3
Ara-C, 10-M 0-5 M
* -
5-Aza-CdR, 3x10-7-1x10-6 M
3 -4 2 -4
BUdR, 10-6-10-5 M 2 -6 3 -5
Calcium ionophore, 7.4x10-6M,
3 or 15 h
- -
Deoxycytidine, 2x10-4 M
- -
Sodium butyrate, lO ^-lO -1 M
- -
Thermal shock, 4°C, 30 min
- -
Thermal shock, 43°C, 30 min
- -
a All treatments were performed on normal and MatDi7
cells at passage 21, except for sodium butyrate treatment,
which was performed on cells at passage 22, and heat shock
and cold shock, which were performed on cells at passage 23.
Results are given only for MatDi7 cells, since normal cells did
not show measurable differences in expression (see Fig. 3-4).
The numbers of determinations by PCR analysis were two for
sodium butyrate and thermal shock and three or more for all
other treatments.
^ Levels of lgf-2 expression were determined by
densitometric analysis of Southern blot hybridizations of the
RT-PCR products with probe H. Expression levels of lgf-2
promoters 2 and 3 were calculated as fold increases in
expression in treated samples compared with untreated
samples. -, less-than-twofold increase in expression. Similar
results were obtained for 5-Aza-CdR treatment of four MatDi7
subclones and for BUdR treatment of two MatDi7 subclones.
See Fig. 3-4 and 3-5 for examples.
87
Figure 3-4. Induction of igf-2 expression in
low-passage MatDf7 celts upon 5-aza-CdR treatment.
RT-PCR analysis of lgf-2 expression in seventh-passage (7°)
normal and MatDi7 cells treated with 3x1 O'7 M and IxlO*6 M
5-aza-CdR (lanes 2 and 3 and 5 and 6, respectively). Lanes 1
and 4 contained controls for lgf-2 expression levels in normal
and MatDi7 cells without treatment. Lane 7 shows the results
for a sample in which no cDNA was added prior to the PCR to
assay for contamination in the PCR reagents or during the
setup of samples for the PCR. Expression of promoters 2 (P2)
and 3 (P3) of the lgf-2 gene is shown. PCR was performed for
26 cycles for normal samples and the no-cDNA control sample
and for 33 cycles for the MatDi7 samples. RT-PCR products
were analyzed by Southern blot analysis using a digoxigenin-
labeled fragment (probe H) as the probe. The filter was then
exposed to X-ray film to visualize the results. RT-PCR was also
performed with actin-specific primers to demonstrate
approximately equivalent levels of the RNA template in all of
the cDNA pools used. The amounts of actin-specific PCR
products were quantitated by densitometric scanning of a
negative of the photograph of ethidium bromide-stained
products. The fold differences between the actin samples for
both Normal and MatDi7 passage 7 cells ranged from 1 to 1.3.
88
0 0
CO
"0
r o
*
A za (M)
1
0
z
o
1
3x10 7
3
8L
1
1x10 -«
-n I
o
0 s
0)
1
3x10 7
O
-si
1
1x10 6
-si
o
no cDNA
Figure 3-5. Induction of lgf-2 expression in MatDI7
immortalized cells does not result from a selection
p ro ce ss.
RT-PCR analysis of expression of lgf-2 in mass culture of
MatDi7 cells at passage 21 (21°) and in two subclones isolated
from single MatDi7 (MatDi7 l-6b and MatDi7 4-2c) cells
treated with 3x10*7 M and lxlO '6 M 5-aza-CdR (lanes 2, 5, and
8 and 3, 6. and 9, respectively). Lanes 1, 4, and 7 show
expression levels in untreated MatDi7 cells. Lane 10 contained
a no-cDNA negative control. Fold differences between the actin
control samples varied from 1 to 1.5 for MatDi7 passage 21
cells, from 1 to 1.3 for MatDi7 l-6b, and from 1 to 1.4 for
MatDi7 4-2c. All samples were subjected to 33 cycles of PCR,
and amplified products were analyzed by Southern analysis,
hybridized with digoxigenin-labeled fragment H, and exposed
to X-ray film.
90
“ 0
ro
Aza(M)
3x10 7
1x10 •*
3x10 7
1x10 ■ *
3x10 7
1x10 * ®
no cDNA
M atD i7 2 1 M atD i7 1 -6 b Mat0i7 4-2c
(Figs. 3-4 and 3-5), whereas treatment of the cells derived
from the normal embryos did not significantly alter the
expression of lgf-2 (Fig. 3-4). Expression levels from both
promoters were increased in most cases in the MatDi7 cells.
The increase in lgf-2 expression in three different subclones
derived from single immortalized MatDi7 cells suggested that
the increase in gene expression did not result from a selection
of pre-existing Igf-2 -expressing cells in the immortalized
population. It was also significant that 5-aza-CdR treatment
could increase the level of expression in low-passage cells
before immortalization (Fig. 3-4), suggesting that the results
obtained with immortalized cells might have relevance to in
vivo processes.
BUdR, which has strong effects on gene expression (Tapscott
et al., 1989), was also effective in increasing lgf-2 expression
from the repressed allele in a dose-dependent manner
(Fig. 3-6). Similar results were obtained for the mass culture
of MatDi7 cells and with two subclones isolated from single
MatDi7 cells, which again argued against a selection of cells
expressing high levels of lgf-2 within the mass culture. Since
treatment of MatDi7 cells with ara-C, which acts primarily as
an inhibitor of DNA synthesis, did not alter the levels of lgf-2
expression, the induction which was observed for both
5-aza-CdR and BUdR treatments was probably not a
92
Figure 3-6. Induction of lgf-2 expression in a MatDI7
subclone upon BUdR treatment.
RT-PCR analysis of lgf-2 expression in subclone MatDi7 l-6b
treated with 10'6 M and 10*5 M BUdR (lanes 2 and 3,
respectively). Lane 1 demonstrates the expression level of
lgf-2 in untreated cells. Lane 4 contained a- negative control in
which no cDNA was added to the PCR. PCR was performed for
30 cycles. Quantitation of actin controls indicated that the
amounts of actin PCR products produced differed 1- to 1.2-fold.
Products were analyzed by Southern analysis, hybridized with
digoxigenin-labeled fragment H, and exposed to X-ray film.
93
>
o
~0
CO
ro
CO
t
C O
~0
r> o
BUdR (M)
§ 1x10
0)
-e 3
1x10 5 6)
CT
no cDNA
consequence of cytotoxicity and may have resulted from direct
effects of these drugs on the gene itself.
The ability of BUdR to increase expression of the maternal
lgf-2 allele was examined in greater detail. The MatDi7 l-6b
subclone was treated with BUdR and the expression of lgf-2
analyzed at distinct time points over a period of 144 h
(Fig. 3-7). Expression of lgf-2 from promoter 2 was induced to
high levels within 24 h after treatment with BUdR and
continued to increase until 72 h after treatment at which time
the expression had reached maximal levels. The induced
expression of lgf-2 from promoter 3, however, was
considerably lower than that from promoter 2 with expression
only being detected 48 h (see lane 6) after treatment. The
cause of the observed difference between the levels of
induction of expression from the lgf-2 promoters 2 and 3 is not
known. E t is important to note that the cells were
approximately 85 to 90% confluent at the time of the first RNA
isolation. Since the cells did not undergo many divisions
during the sampling period, the cells may have still contained
both unincorporated BUdR and BUdR incorporated into the
DNA. The possibility that both unincorporated and
incorporated BUdR may have remained in the cells prevented
determination of the heritability of the observed effect.
95
Figure 3-7. Time course of induction of lg f - 2
expression In MatDi7 subclone upon BUdR treatment.
RT-PCR analysis of lgf-2 expression in MatDi7 l-6b subclone
treated with 10'6 and 10-5 M BUdR at one, two, three, and six
days after treatment (lanes 2, 5, 8, 11 and 3, 6, 9, and 12,
respectively). Lanes 1, 4, 7, and 10 demonstrates the
expression levels of lgf-2 in untreated cells. Lane 13 contained
a negative control to assay for possible contamination of the
PCR. PCR was performed for 32 cycles. PCR products were
analyzed by Southern blot analysis, hybridized with a
digoxigenin-labeled fragment H, and exposed to X-ray film.
PCR was also performed with actin-specific primers to
demonstrate similar levels of RNA template in the different
sam ples.
96
L 6
>
o
T >
O )
T >
fO
BUdR (M)
ro
U )
c n
cn
-O
ao
to
ro
co
I
I
0
1x10
1x10
0
1x10
1x10
0
1x10
1x10
•6
■ 5
6
-5
-6
■5
-6
-5
0
1x10
1x10
no cDNA
O
0 )
*<
O
fu
<
ro
O
0 )
-<
CO
O
C D
•<
CD
To determine if the increase in the expression of lgf-2 upon
treatment with 5-aza-CdR or BUdR also occurred in cells at a
lower passage than seven as was previously used, cells were
treated with these drugs immediately after explant (Fig. 3-8).
Interestingly, second-passage MatDi7 cells treated with
5-aza-CdR did not exhibit an increase in the expression of lgf-2
(lanes 7 and 8), whereas cells treated with BUdR did exhibit an
increase in expression (lanes 9 and 10). The failure to observe
an increase in lgf-2 expression in second-passage MatDi7 cells
upon treatment with 5-aza-CdR may reflect differences which
occurred in the cells at higher-passages during the process of
cell culturing. However, the induction of lgf-2 expression in
the higher-passage MatDi7 cells indicates that treatment of
5-aza-CdR does in fact result in increased expression of the
normally repressed maternal lgf-2 allele, regardless of changes
acquired during culturing.
Isolation of MatDI7 Subclones which Express lgf-2
Although, the MatDi7 cells generally maintained lgf-2 in a
repressed state, both chemically-induced and spontaneous
revertants could be isolated. A muscle-forming subclone
isolated from MatDi7 cells (MatDi7 Aza Cl L; isolated by F. A.
Gonzales) after 5-aza-CR treatment expressed lgf-2 at a level
98
Figure 9-8. Expression of lgf-2 in second-passage
MatD17 cells treated with 5-aza-CdR and BUdR.
RT-PCR analysis of lgf-2 expression in second-passage (2°)
normal and MatDi7 cells treated with 3x1 O'7 and lxl0~6 M
5-aza-CdR (lanes 2 and 3 and 7 and 8, respectively) and 10*6
and 10*5 M BUdR (lanes 4 and 5 and 9 and 10, respectively).
Lanes 1 and 6 demonstrates the expression levels of lgf-2 in
untreated normal and MatDi7 cells, respectively. Lane 11
contained a sample in which no cDNA was added to control for
possible contamination. PCR was performed for 26 cycles for
normal samples and 32 cycles for MatDi7 samples. RT-PCR was
also performed with actin-specific primers to demonstrate
similar levels of RNA template in the various cDNA pools used.
99
0(H
>
o
T J
W
"0
D
CO
cn
O )
"4
00
CD
3x10 7
1x10 6
1x10 6
1x10 5
>
N
0 0
i
CD
C
o .
3 )
3x10 7
1x10 6
1x10 6
1x10 5
no cDNA
N
C D
c
Q .
X
approximately 30% of that seen in the normal counterpart
(data not shown). This clone showed that expression from
promoters 2 and 3 was growth-regulated, suggesting that
control of the maternal allele was similar to that of the
paternal allele (Fig. 3-9). It was also found that one of four
randomly picked clones (MatDi7 1-la) derived from the
progeny of single MatDi7 cells at passage 21 expressed the
maternally derived lgf-2 (Fig. 3-10). Thus, although the gene
was found to be repressed generally in culture, spontaneous
and induced expression of the maternal lgf-2 allele could
occur. The relationship between spontaneous activation of
lgf-2 in vitro and the reversible imprinting of the gene in the
germ line, as well as the lack of such imprints in some somatic
tissues, notably the leptomeninges and choroid plexus, is
unknown (DeChiara et al., 1991).
DISCUSSION
Chapter 2 describes the establishment and characterization
of somatic cell lines derived from embryos maternally disomic
for distal chromosome 7, MatDi7, and their normal littermates.
lgf-2 was expressed in a cell culture derived from normal
embryos; this expression was growth-regulated with lgf-2
mRNA levels increasing in the stationary phase of cell growth.
lgf-2 was not expressed in the MatDi7 derived celts except
1 0 1
Figure 3-9. Growth-regulated expression of Igf~2 in
subclone MatD17 Aza Cl L.
RT-PCR analysis of lgf-2 expression in aza-derivative MatDi7
cells. PCR was performed for 26 cycles. An autoradiograph of
the Southern blot hybridized with radiolabeled probe H is
shown. Lanes 1 and 3 show expression of lgf-2 P2 and P3,
respectively, in cells in the logarithmic (log) phase of growth
and lanes 2 and 4 show expression of lgf-2 P2 and P3,
respectively, in confluent (conf) cells.
102
MatDi7 Aza Cl L
S 8 S 8
P2
P3
1 2 3 4
1 03
Figure 3-10. Growth-regulated expression of lgf-2 in
MatDf7 1-la subclone.
RT-PCR analysis of lgf-2 expression in MatDi7 subclone which
spontaneously expressed the maternal lgf-2 allele. Lane 1
shows the expression of lgf-2 in cells in the logarithmic (log)
phase of cell growth and lane 2 shows the expression of lgf-2
in confluent (conf) cells. Lane 3 contained a sample in which
no cDNA was added to assay for possible contamination. PCR
was performed for 33 cycles. Products were analyzed by
Southern blot hybridization with a digoxigenin-labeled
fragment H. RT-PCR was also performed with actin-specific
primers to demonstrate similar levels of RNA template in the
different cDNA pools used.
104
log
conf
no cDNA
. 0)
VI
very infrequently when there was spontaneous reactivation.
The fact that the increased levels of lgf-2 messages were only
found in growth arrested cultures of normal cells and not in
MatDi7 cultures indicates that the imprinted lgf-2 is subject to
regulations within limits imposed by the imprint. Thus the
maternally inherited lgf-2 allele is kept silent by an as yet
unidentified mechanism, whereas the paternally inherited
allele is either active or inactive depending on the physiological
state of the cell. Reversal of the repressed state should
therefore result in the maternal allele being subject to the
same growth regulations as the paternal allele. This is what
was indeed observed as coordinate derepression of both
promoters 2 and 3 occurs both spontaneously and in induced
activation of the repressed maternal lgf-2 . Furthermore both
promoters in the clone derived after 5-aza-CR treatment were
strongly growth-regulated, with a large increase in expression
in monolayer cultures.
An extensive methylation analysis of lgf-2 failed to show
differences in any of the cell lines which were consistent with
methylation playing a role in repression of the maternal allele.
The mechanism of lgf-2 repression may involve (i) changes in
the methylation status of possible key regulatory sites located
within the gene or outside of the gene not included in this
analysis or perhaps (ii) changes in chromatin conformation,
106
since induction of the repressed lgf-2 allele was only observed
when cells were treated with 5-aza-CdR and BUdR which are
known to affect the methylation profiles and/or chromatin
structures of genes.
Although, expression of the silent maternal lgf-2 allele
could be increased upon treatment of both low-passage
(seventh-passage) and immortalized MatDi7 cells with
5-aza-CdR and BUdR, increased expression in second-passage
MatDi7 cells was only observed upon treatment with BUdR.
The observed differences in 5-aza-CdR inducibility of lgf-2
expression between the second- and seventh-passage MatDi7
cells may result from changes which had occurred during the
early stages of culturing. Regardless of these differences, the
induction of lgf-2 expression in the seventh-passage and
immortalized MatDi7 cells indicates that the treatment of
5-aza-CdR does cause an increase in the expression of the
normally repressed maternal lgf-2 allele.
The induction of lgf-2 expression upon 5-aza-CdR and BUdR
treatment is unlikely to be due to the cytotoxic effects of the
drugs, since activation was not observed in MatDi7 cells treated
with ara-C, a drug which interferes primarily with DNA
synthesis. Induction of lgf-2 in MatDi7 cells was not a result of
a selection of cells already expressing high levels of lgf-2
107
within the mass population of cells, since activation was also
demonstrated in three subclones isolated from single MatDi7
cells. The increased sensitivitities of two subclones,
MatDi7 l-6b and MatDi7 4-2c, to 5-aza-CdR treatment may be
due to the incorporation of the drug in more cells which are
actively dividing in these lines which display shorter
generation times (data not shown). Therefore, the activation of
the maternal copy of lgf-2 by 5-aza-CdR and BUdR resulted not
from a selection process but rather from increased expression
of the silent lgf-2 allele. Additional treatments also known to
affect gene expression, i.e., calcium ionophore, heat and cold
shocks, and sodium butyrate, did not induce expression of the
maternal lgf-2 allele.
An additional regulatory mechanism of gene expression in
eucaryotes is the presence of endogenous antisense transcripts
(Hildebrandt and Nellen, 1992). Simultaneous transcription of
both strands of the DNA helix may inhibit transcription
because of physical hindrance of the multiple transcription
complexes. Antisense transcripts may also regulate levels of
gene expression by hybridizing to the mRNA transcripts and
altering their stability or preventing their subsequent
translation. Antisense transcripts for lgf-2 have been detected
in the chicken (Taylor et al., 1991) and in the intergenic region
between the insulin and lgf-2 genes (Rivkin et al., 1993).
108
However, antisense transcripts in the 3' region of lgf-2 do not
appear to be present in the normal or MatDi7 mouse cell lines
used in this study. Therefore, antisense transcription from this
region is probably not involved in the repression of the
maternal lgf-2 allele in culture.
The lgf-2 and H19 genes are closely linked on chromosome
7 yet are oppositely imprinted. These genes appear to be
coordinately regulated in the mouse embryo (Poirier et al.,
1991), except in the choroid plexus and the leptomeninges,
where the H19 gene is not expressed but the maternal copy of
lgf-2 is expressed. If these two genes share an imprint
mechanism which somehow silences the maternal lgf-2 allele
and activates the H19 maternal allele, then perhaps
reactivation of lgf-2 results in inactivation of the H19 gene.
The occurrence of enhancers associated with expression of the
H19 gene (Yoo-Warren et al., 1988) make enhancer
competition between the H19 gene and lgf-2 an attractive
possibility. The failure to find parental gene-specific
methylation differences in the promoters of lgf-2 indicates
that the silencing event occurs outside the gene, perhaps
within silencers or enhancer sequences. This is now a testable
hypothesis, with the derivation of MatDi7 cell lines in which
the maternal lgf-2 allele is stably repressed and lines which
express the maternal lgf-2 allele.
109
CHAPTER 4
Coordinate Regulation of lgf-2 and H 1 9 in
Cultured Cells Indicates that these Genes may
be Regulated by a Common Regulatory
M echanism
INTRODUCTION
The mouse insulin-like growth factor 2 (lgf-2 ) and H19
genes are located within 100 kb of each other on mouse distal
chromosome 7 in the same transcriptional orientation (Zemel et
al., 1992). lgf-2 and H19 are reciprocally imprinted such that
the paternal lgf-2 allele and the maternal H19 allele are
expressed in the embryo (DeChiara et al., 1991; Bartolomei et
al., 1991). Although these genes are transcribed from different
parental chromosomes, the timing and tissue localization of
expression of both genes is generally very similar (Lee et al.,
1990; Poirier et al., 1991). Since the similar spatial and
temporal expression patterns for lgf-2 and H I 9 were
determined using in situ hybridization and Northern analysis it
is not yet known if the expression of these genes is
coordinately controlled. The mechanism underlying reciprocity
of parental origin effects is not understood but could
conceivably be achieved by several different pathways (Sasaki
110
et al., 1992; Bartolomei et al., 1993). Indeed, it has been
suggested that lgf-2 and H 19 may share a common regulatory
mechanism but little direct evidence in support of this has yet
been shown.
The molecular mechanisms responsible for imprinting of
lgf-2 and H19 are not clear, but it is seems likely that DNA
methylation plays an important role since mice deficient in
DNA methyltransferase do not maintain the
parental-origin-specific expression of these genes (Li et al.,
1993). Increased methylation of CpG rich regions upstream of
the unmethylated lgf-2 promoters and within the H19
promoter are found on the paternal chromosome (Sasaki et al..
1992; Ferguson-Smith et al., 1993; Bartolomei et al., 1993).
Since it has been shown directly that methylation of the human
H 19 promoter can lead to its transcriptional inactivation (Zhang
et al., 1993) it is reasonable to suggest that the silencing of the
paternal H19 promoter by imprinting may be achieved, or
reinforced, by DNA methylation.
The establishment of mouse somatic cell lines from normal
embryos or from their littermates which were maternally
disomic and paternally deficient for distal chromosome 7
(MatDi7) where the Igf-2/H19 locus resides was described in
Chapter 2 (Eversole-Cire et al., 1993). It was shown that the
111
normal cells upregulated lgf-2 raRNA levels at growth arrest
and that subclones could be derived which had extinguished
lgf-2 expression. The repression of lgf-2 was quite stable in
the MatDi7 cells, although one subclone was obtained which
began to express the gene spontaneously. These somatic cell
clones as well as newly isolated clones have now been
examined for the expression of both lgf-2 and H I9 and for
methylation of regions previously shown to be of potential
importance in control of gene expression. The genes were
expressed in a highly coordinated manner in the normal cells
suggesting a common control mechanism. Interestingly,
spontaneous activation of lgf-2 in a clone of MatDi7 cells was
associated with de novo methylation of both the lgf-2
upstream region and the HI 9. promoter so that the DNA
methylation pattern on the maternally derived chromosome(s)
resembled that of a paternal chromosome.
MATERIALS AND METHODS
Cell Culture
Derivation of subclones Normal 2-5b, Normal 3-3b,
Normal 3-5b, Normal Cl 1, MatDi7 1-la, and MatDi7 l-6b, cell
culturing conditions, and staging of cultures for RNA isolation
were described in Chapter 3 (Eversole-Cire et al., 1993).
112
Expression Studies
Normal Cl 1 cells were seeded at 0.4 x 106/100-mm dish.
Total RNA was isolated daily for 5 consecutive days (Days 1 to
5) as described by Chomzynski and Sacchi (1987) starting
three days after seeding when cells were 60-70% confluent.
Medium was changed on the remaining dishes three days after
seeding and every second day thereafter. A duplicate dish was
trypsinized on each day and the cells counted in a Coulter
Counter. The cells for the Day 5 sample were confluent as
assessed visually and as indicated from a growth curve. Cells
were maintained in Eagle’s minimal essential medium
supplemented with 10% heat-inactivated fetal calf serum
(HIFCS). MatDi7 1-la and MatDi7 l-6b were seeded at
0.15 x 106/60-mm dish, the same cell density used above for
the normal cells. The first RNA sample (Day 1) was isolated
two days after seeding. RNA was isolated daily for 5
consecutive days and then 9 days after seeding. Medium was
changed on Day 2 and every other day thereafter.
Normal Cl 1 cells were seeded as described above to study
the effect of serum on the expression of lgf-2 and H19> The
serum concentration in the medium was reduced to 0.5% HIFCS
on the first day of RNA isolation (Day 1). RNA was isolated
from dishes at 0, 4, 8, 24, and 48 h after addition of the 0.5%
serum. After cells were maintained in the 0.5% serum for 48 h,
113
the medium was then changed to that containing 20% serum
and RNA isolated at 0, 4, 8, 24, and 48 h later. The cells
appeared to be approximately 85-90% confluent after 48 h in
the presence of 20% serum. MatDi7 cells were seeded as above
and medium containing 0.5% serum was added to the cells two
days after seeding (Day 1). RNA samples were isolated at 0,
24, and 48 h after addition of the low serum. Medium
containing 20% serum was then added and RNA extracted at 24
and 48 h after addition. Cells were maintained in medium
containing 20% serum for an additional 3 days and RNA was
isolated on Day 8.
DNA Synthesis
Cell cycle progression for expression studies using MatDi7
cells was monitored by [3H]Thymidine (TdR) incorporation.
Replicate cultures of cells were pulsed for 30 min with
0.25 uci/ml TdR and then washed twice with PBS. Salmon
sperm DNA, 25 ug/ml, and 1.5 ml of 0.1% SDS were added to
each dish and cells were allowed to lyse for 10 min. Samples
were collected and transferred to polystyrene tubes to which
150 ul of cold 50% trichloroacetic acid (TCA) were added.
Samples were precipitated on ice for a minimum of 10 min and
then filtered through Whatman GF/C filters and washed with 5
ml of cold 5% TCA. Filters were allowed to air dry and then
transferred to scintillation vials and 10 ml of aqueous
114
scintillation fluid added. Amount of radioactivity bound to
each filter was counted for 1 min in a scintillation counter.
Southern Blot Analysis
DNA isolation, enzyme restriction digests, Southern blot
hybridizations, and preparation of radiolabeled probes were
performed as previously described in Chapter 3 (Eversole-Cire
et al., 1993).
P ro b es
The mouse lgf-2 genomic clone used in the methylation
analysis was provided by Peter Rotwein, Washington
University, St. Louis, MO. Probe A was a 4.8 kb BamHX
fragment, B2 was a 2.6 kb Bam HI fragment, and probe Al was
a 0.7 kb EeoRI to Xba I fragment excised from probe A. For
hybridization with HI 9, probe 1 was a 1.0 kb Bam HI to Xba I
fragment and probe 2 was a 1.6 kb Xba I to Bam HI fragment.
These clones were kindly provided by T. Kiode (manuscript in
preparation). Probe 3, a 0.8 kb BcoRI to Bam HI fragment, was
subcloned from a cosmid by Drs. C. J. Chan and P. Rigby.
Reverse Transcription (RT)-Polymerase Chain Reaction
(PCR)
The procedure used for RT-PCR analysis for lgf-2 is
described in Chapter 3 (Eversole-Cire et al., 1993). PCR
115
conditions used for H19 analysis were the same as those used
for lgf-2. Parameters for each PCR cycle for
glyceraldehyde-3-phosphate dehydrogenase (G APD H )
amplification were as follows: 94°C for 1 min, 60°C for 1 min,
and 72°C for 1 min. The number of PCR cycles performed using
each primer set was chosen so that the reactions were carried
out under non-saturating conditions (as determined from cycle
curves). The primers used for PCR were as follows:
H19 OP I (exon 5, inter, oligo) 5’TCA TAG CAC CCA CCC ACC CC3’
H19 OP2 (exon 5, 3' oligo) 5'GGG AAG GCG AGG CCT CAA GC3'
HJ9 OP3 (exon 4, 5’ oligo) 5'GGG CCT TTG A AT CCG GGG AC3'
GAPDH\ (5' oligo) 5’ CAG CCT CGT CCC GTA GAC AAA ATG G3'
GAPDH2 (3* oligo) 5TTC TGG GTG GCA GTG ATG GCA TGG A3'
GAPDH3 (internal oligo) 5'CGG TGC TGA GTA TGT CGT GG3'
RESULTS
Expression of Igf~2 and H19 are Extinguished In
Unison
The reciprocally imprinted lgf-2 and H19 exhibit similar
temporal and spatial patterns of expression during early mouse
development (Lee et al., 1990; Poirier et al., 1991). lgf-2 has
116
three promoters in the mouse which are differentially utilized
during development (Sasaki et al., 1992) whereas H19 has one
promoter which is a CpG island (Ferguson-Smith et al., 1993)
and 2 downstream enhancers (Yoo-Warren et al., 1988). The
fact that both genes are often simultaneously expressed in
tissues suggests that they might share a common mechanism of
transcriptional regulation. Several clonal derivatives of a cell
line originally derived from a normal mouse embryo were
therefore isolated (Eversole-Cire et al., 1993) and examined for
the coordinate expression of lgf-2 and H19 using a sensitive
and quantitative RT-PCR assay similar to that previously
described for expression analysis of lgf-2 in Chapter 3
(Fig 4-1).
All cells were examined for gene expression in both the
logarithmic and stationary phases of cell growth. Figure 4-1
shows that the Normal Cl 1 and Normal 3-5b clones expressed
the promoter 3 driven transcript of lgf-2 and also HI 9, and
that the levels of both mRNAs were increased upon confluence.
Only lgf-2 promoter 3 transcription was assessed in this
experiment since the levels of promoter 2 and promoter 3
transcripts for several of the normal clones were generally
very similar in a specific sample (see for example Fig. 4-2).
The Normal 2-Sb clone also expressed lgf-2 and HJ9 upon
confluency but at much lower levels whereas the Normal 3-3b
117
Figure 4-1. Expression of lgf-2 and H I 9 in subclones
of normal ceils.
RT-PCR analysis of lgf-2 and H19 expression in the logarithmic
(log) phase and confluency (conf) in several subclones derived
from a normal embryo at el5 (lanes 1 to 4 and 5 to 8,
respectively). Lane 9 is a control in which no cDNA was added
to assay for possible contamination. The levels of GAPDH
mRNA were determined to demonstrate that the different
cDNA pools contained approximately equivalent amounts of
RNA templates. PCR was performed for 28 cycles for the
amplifications using the lgf-2 and H I9 primers and for 21
cycles for GAPDH primers. Parameters for the PCR were chosen
so that the reactions were performed under non-saturating
conditions of the assay (data not shown). RT-PCR products
were analyzed by Southern blot analysis, hybridization to a
digoxigenin-labeled lgf-2 fragment H or a 3'-end labeled
internal HI9 or GAPDH oligomer, and exposure to X-ray film to
visualize the results.
118
Normal Cl 1
Normal 2-5b
Normal 3-3b
Normal 3-5b
Normal Cl 1
Normal 2-5b
Normal 3-3b
Normal 3-5b
no cDNA
clone expressed neither gene. The data are consistent with a
single control mechanism for both loci since both genes were
up- or downregulated in unison.
Coordinate Growth Regulation of Igf~2 and H 1 9
More detailed experiments were next performed to examine
the effects of cell growth phase on the levels of both lgf-2 and
H19 mRNA. Figure 4-2 (left side) shows that the levels of
these fgf-2 transcripts generally increased in the Normal Cl 1
cells upon confluency (compare lanes 1 and 5). Decreases in
lgf-2 transcript levels observed on days 2 and 4 were
probably a result of restimulation of cells by medium changes
performed on days 1 and 3, respectively. Figure 4-2 also
shows that the levels of H19 mRNA in the same cultures
increased as the cells became growth arrested due to
confluency. Although decreases were not observed for H I9
transcript levels on days 2 and 4 as were seen for lgf-2y there
was a generalized increase in transcription as the cells became
quiescent. The effects of confluency on the expression levels of
the two genes were mirrored when subconfluent cells were
forced to become quiescent by serum starvation (Fig. 4-2, right
side). The levels of lgf-2 promoter 2 and 3 and also HI 9
transcripts behaved identically by increasing mRNA levels 48 h
after growth factor depletion and showed a decline as cells
120
Figure 4-2. Coordinate expression of lgf-2 and H19 in
Normal Cl 1 cells.
Expression levels of lgf-2 from Promoter 2 (P2) and from
Promoter 3 (P3) and HI9 in the Normal Cl 1 subclone were
analyzed for response to cell growth using RT-PCR. Lanes 1
through 5 show expression of lgf-2 and H19 in cells during the
various stages of cell growth. Lanes 7 through 10 show
expression levels of lgf-2 and H I9 during serum starvation,
and lanes 11 through 14 show expression when cells were
restimulated by the addition of serum. Lanes 6 and IS are
controls for contamination in which no cDNA was added to the
reaction. GAPDH amplification was performed to demonstrate
equivalent levels of RNA template in each cDNA sample used.
PCR was performed for 26 cycles for lgf-2 and H19 samples
and for 21 cycles for GAPDH samples and results analyzed as in
Fig. 4-1.
121
< 0 M M
■ o - Q
( • > ro
I I | D * y 1 S
| ( | Day 2
| | | Day 3
I • • D * > ' 4 o
I 9 I
no cDNA
0 Hr
4 Hr
8 Hr
48 Hr
4 Hr
8 Hr
24 Hr
48 Hr
no cDNA
I
I
I
0 5 % S e ru m 2 0 % Serum
were stimulated to reinitiate cell division by the addition of
20% fetal calf serum. These experiments again demonstrated
coordinate control of mRNA levels for all three transcripts as a
function of cell growth.
The MatDi7 l-6b clone expressed at least 50 fold less lgf-2
mRNA than the Normal Cl 1 or the revertant MatDi7 1-la cells
(based on differences in levels of lgf-2 expression in mass
cultures of normal and MatDi7 cells at passage 21) so that
visualization of the appropriate PCR band required an
overnight exposure time versus the several hours required for
the other clones. The levels of lgf-2 mRNA from promoter 3 in
the MatDi7 l-6b clone demonstrated the same general pattern
of growth regulation as with the other clones even though the
absolute level of expression was much reduced (Fig. 4-3A).
HJ9 mRNA levels in these cells increased upon confluence
arrest similarly to the Normal Cl 1 cells but appeared to be
relatively unaffected by the removal and re-addition of serum
to the medium.
Since the paternal lgf-2 and the maternal H19 appeared to
be coordinately regulated in their responses to cell growth in
normal cells, it was determined whether MatDi7 cells, which
had spontaneously acquired the ability to express the maternal
lgf-2 allele ([MatDi7 1-la]; Eversole-Cire et al., 1993) behaved
123
Figure 4-3A. H19 expression in MatDI7 I -6b cells
becomes deregulated.
RT-PCR analyses of lgf-2 and H19 expression levels in cells
during the logarithmic and stationary phases of cell growth are
shown in lanes 1 through 6. The levels of expression in
response to the removal and addition of serum to the medium
are shown in lanes 7 through 12. (Note: The level of lgf-2
expression is more than 50-fold lower than in Figs. 4-2 and
4-4A, and bands were only visualized after extended exposure
time.) A control for sample contamination is shown in lane 13.
Levels of GAPDH mRNA were determined to demonstrate
similar levels of RNA template in the various cDNA pools used.
PCR was performed for 27 cycles for lgf-2 and H19 and 21
cycles for GAPDH. Amplified products were analyzed as
described in Fig. 4-1.
124
< o
ro
T J
C J
D ay 1 £
. Day 2
Day 3
D ay 4
| Day 5
| D ay 8
I
I
0 Hr
o
in
*
24 Hr
V
48 Hr
2
3
24 Hr
48 Hr
V
3
120 Hr
c
3
no cDNA
M atDi7 1-6b
Figure 4-3B. [3H]Thymldlne Incorporation In
MatD17 l-6b cells Induced to become quiescent either
by confluence arrest or serum starvation.
Solid bars- TdR incorporation as cells became quiescent in
response to confluency. Stippled bars- TdR incorporation in
cells as they responded to the removal and readdition of serum
to the medium. High serum was added to cultures 24 h prior
to Day 4 sampling. Numbers represent an average from
replicate cultures.
126
80000
60000
40000
20000
■ Di1-6b Confluency
□ D M -6b Serum
5 6
similarly, lgf-2 mRNA levels from promoter 3 increased upon
confluence arrest in this cell line (Fig. 4-4A, left side). The
level of H I 9 mRNA, however, was slightly downregulated in
confluence-arrested MatDi7 1-la cells. The two genes also
showed opposite behavior on serum starvation and
restimulation (Fig. 4-4A, right side). Therefore, in
MatDi7 1-1 a cells, lgf-2 and H19 were not coordinately
regulated in their responses to cell growth as they were in
normal cells.
For expression studies using MatDi7 cells, cell cycle
progression was monitored using TdR incorporation. Figures
4-3B and 4-4B show that the amount of TdR incorporation
decreased significantly as the cells became quiescent in
response to confluency or serum starvation. The readdition of
serum to serum-starved cells resulted in an initial increase in
TdR incorporation as the cells reentered the cell cycle which
then decreased as the cells approached confluency (compare
TdR incorporation for Day 4 samples with Days 5 and 8
samples, Fig. 4-3B and 4-4B). The MatDi7 cells were therefore
responding to the changing culture conditions appropriately.
128
Figure 4-4A. Igf~2 and H19 demonstrate reciprocal
expression patterns in MatDI7 1-ta cells.
RT-PCR analysis of lgf-2 and H19 expression levels in
MatDi7 1-la cells as a function of the growth status of the
cells. Lanes 1 through 6 demonstrate the levels of lgf-2 and
H I 9 transcripts as cells undergo confluence arrest. The effect
of serum depletion on the expression of these genes is shown
in lanes 7 through 9, and the effects of serum addition to
serum-starved MatDi7 1-la cells is shown in lanes 10 through
12. Lane 13 contains a no cDNA control sample. Amplifications
of GAPDH was performed to ensure that the cDNA samples
contained equivalent amounts of RNA template. PCR was
performed for 28 cycles for lgf-2 and H19 and for 21 cycles for
GAPDH amplification. PCR products were analyzed as described
in Fig. 4-1.
129
no cDNA
Figure 4-4B. pHJThymldine Incorporation in
MatDi7 1-la cells induced to become quiescent either
by confluence arrest or serum starvation.
Solid bars- TdR incorporation as cells became quiescent in
response to confluency. Stippled bars- TdR incorporation in
cells as they responded to the removal and readdition of serum
to the medium. High serum was added to cultures 24 h prior
to Day 4 sampling. Numbers represent an average from
replicate cultures.
131
3H -TdR I n c o r p o r a t io n ( c p m )
60000
50000
40000
30000
20000
10000
■ Di 1 -1 a Confluency
□ Oi 1 -1 a Serum
Reactivation of the Maternal lgf-2 Allele is Associated
with Increased Methylation of the 5' Region of lgf-2
and of the H19 Promoter
A region upstream of the lgf-2 promoters and the H19
promoter itself are methylated on the paternal but not the
maternal allele, suggesting that cytosine methylation may have
importance in the molecular control of imprinting (Sasaki et al.,
1992; Ferguson-Smith et al., 1993). The methylation status of
these regions (see Fig. 4-5 for map and summary of results) in
the normal and disomic cell lines were therefore analyzed,
paying special attention to possible methylation changes
associated with reactivation of lgf-2 expression in the
revertant cell line MatDi7 l*la.
A methylation analysis of the region 5' upstream of lgf-2
was performed using DNA extracted from the subclones
Normal Cl 1, MatDi7 1-la, and MatDi7 l-6b digested with
Bam\\\ to generate specific bands and double digested with
either of the methylation-sensitive enzymes H p a \\o xS m a \.
Probes used in the analysis are given in Fig. 4-5 and a typical
Southern blot analysis (hybridized with probe A) is shown in
Fig. 4-6. Hpa II digestion of all three subclones yielded bands
of 3.9, 3.5, and 3.2 kb. However, a comparison of the
133
Figure 4-5. Map of region 5* to lgf-2 and the H 19
promoter and summary of the methylation analysis of
the region 5' to lgf-2 and the H19 p ro m o te r.
A. Map of region 5' to lgf-2 and the HJ9 promoter. Exons and
pseudoexons (\y) are shown as solid filled boxes. The scales for
the two regions are different. The probes used in the analyses
are depicted as stippled boxes. The positions of the restriction
enzyme sites and distances between their recognition sites are
indicated. The positions of the sites for methylation-sensitive
enzymes are given as follows: Hae II (Ha), Hha\ (Hh), Hpa II
(Hp), and Sma I (Sm) (additional Mspl (HpaII) sites are present
upstream of the Hpl site located in the region 5' to lgf-2 based
on results obtained for Msp I digestion). Slashed lines
represent approximately 90 kb of intervening DNA sequence
between lgf-2 and HI 9. B. Summary of the methylation
analysis of the region 5' to lgf-2 and the HI 9 promoter. A
closed circle indicates complete methylation of a CpG
dinucleotide, a partially closed circle indicates partial
methylation, and an open circle indicates no methylation.
Lines without circles represent sites whose methylation status
could not be determined in this analysis. Methylation analysis
of the Sm a I site designated Sm2 was previously described in
Chapter 2 and is included for completion.
134
*
I * ■
- !
f
s
*
J •- *- •-
V i H
©- O "
1/3
i
«r
C T «r
, *F
% =
if
?-
\i f-
o o-
C f c ct
< «
h.
I t
135
Figure 4-6. Increased methylation of the region 5* to
the lgf-2 promoters correlates with reactivation of the
maternal Igf~2 allele.
Methylation status of the 5' region of Igf-2 was determined for
the subclones Normal Cl 1, MatDi7 1-1 a, and MatDi7 l-6b.
Isolated DNAs were digested with Bam HI and either the
methylation-sensitive enzymes 5 m al or Hpa II. Mspl, an
enzyme insensitive to methylation status of the internal
cytosine of the Hpa U sequence (CCGG), was used as a control.
The probe used in the Southern blot analysis was probe A, a
4.8 kb BamHl fragment (Fig. 4-5A). The sizes of fragments
generated are indicated and the results from this and other
analyses are summarized in Fig. 4-5B.
136
137
to
u
a
w a
« «
\
0
s
1 I
g v
* I
• «
62'l
to t* « m im
w .r
BamHl o
Bam m/Hpa I I
BamHVMapi
o
£
£
5
£ % ■ Bam H J
£ Bam m/Sma l
\ BamHI/Hpall
BamhH/Mgpl
£ Bam H t
£ Bam WSma I q
[ Bam Hl/Hpa M
Bam Hl/Mspl
intensities of the individual bands revealed that the DNA
isolated from the revertant MatDi7 1-la was more methylated
since the predominant fragment was the 3.9 kb band (compare
lanes 2, 6 and 10). Therefore, the 5* upstream region was more
methylated at HpaU sites in DNA isolated from MatDi7 1-la
cells than in DNA isolated from MatDi7 1 -6b cells or from
Normal Cl 1 cells. Methylation results were confirmed by a
second analysis in which the DNA was double digested with
£coR I and HpaU or Sm all and then hybridized with probe A1
(data summarized in Fig. 4-5). The results obtained with the
cell lines were compared with those obtained from DNA
extracted from normal or MatDi7 embryos. Subtle differences
in methylation patterns had occurred in culture (not shown),
but none were as notable as the increased methylation of this
region in the revertant MatDi7 1-la cells (summarized in
Fig 4-5).
The methylation status of several sites within the H19
promoter (Fig. 4-5) were next measured in the subclones by
digesting extracted DNAs with either Apal, HindUl, or BamH I
to generate specific fragments which were double digested
with methylation-sensitive enzymes. The results for all of the
Southern blot analyses are summarized in Fig. 4-5. The 1.0 kb
A pal fragment which hybridizes to probe 3 was cleaved nearly
to completion by Hpa II generating bands at 0.78 and 0.22 kb
138
in DNAs isolated from MatDi7 1-la and MatDi7 l-6b indicating
that the HpaU site designated Hp5 was nearly completely
unmethylated in these cells (Fig. 4-7, lanes 7 and 11). The
banding pattern generated was similar to that obtained using
DNAs isolated from MatDi7 embryos (lane 15). This site was
also cleaved in DNA isolated from Normal Cl 1 cells but to a
lesser extent (lane 4). The Hp5 site appeared to have
undergone some de novo methylation in the Normal CL 1 cells
since this site was found to be approximately 50% methylated
in DNA isolated from normal embryos (lane 13). The Smal site
(Sm2) was found to be completely methylated in the three cell
lines since treatment with Sma I failed to cleave the DNAs
(lanes 4, 8, and 12) and also in DNAs isolated from normal and
MatDi7 embryos (lanes 14 and 16). Digestions with Mspl were
also performed as a control for complete digestion since this
enzyme recognizes the same restriction site as HpaU and the
same internal sequence recognized by Sma 1 (CCGG) but is not
sensitive to methylation of the internal cytosine (lanes 2, 6,
and 10).
DNAs isolated from Normal Cl 1, MatDi7 1-la, and
MatDi7 l-6b were digested with HindUX and hybridized to
probe 1. Digestion of Normal Cl 1 DNA with HpaU gave rise to
a predominate band at 2.0 kb indicating that site Hp3 was
fairly unmethylated and that any HpaU sites within this
139
Figure 4-7. Methylation analysis of exon 1 of H19 in
normal and MatD17 cell lines.
Methylation status of the H19 promoter region was determined
for the Normal Cl 1, MatDi? 1-la, and MatDi7 l-6b subclones
and for normal (N) and MatDi7 (Di) embryonic DNA. DNAs
were isolated and digested with A p a I and M s p I, H p a II. or
Sma I. Southern blot analysis was performed using a 0.8 kb
EcoRI to Bam HI fragment (probe 3, Fig. 4-5 A) as a probe.
Results are summarized in Fig. 4-5B.
140
Nonnal Cl 1 MatDt7 1-la M«tf»7 1-«to N Ol
-tii.iii.liilifi
inttlilllliutl
kb
1 8 3 4 8 8 7 8 • 10 11 13 IS 14 18 18
141
fragment were methylated (Fig. 4-8, lane 3). The presence of a
faint band at 1.0 kb suggested that the Hpl site was partially
unmethylated and the residual 4.4 kb band dem onstrated that
all of the H p a II sites within the Hindlll fragment were at least
partially methylated. Digestion of DNAs isolated from
MatDi7 1-la and MatDi7 l-6b with H p a II yielded a 2.0 kb
band indicating that site Hp3 was unmethylated in these cells
(lanes 6 and 9).
Figure 4-9 shows that the Hhal and H ae II sites in the
Bam HI fragment were partially methylated in the Normal Cl 1
cells since the 2.6 kb BamYW fragment was reduced in intensity
and new bands of 1.7 kb appeared on the blot. This result was
similar to that obtained for DNA obtained from uncultured
mouse DNA (not shown) and was consistent with previous
observations that the paternal but not the maternal promoter
is fully methylated at these sites (Ferguson-Smith et al., 1993).
The Sma I site on the maternal chromosome appeared to have
undergone some de novo methylation in culture since the
fragment was only partially cut by this enzyme whereas it was
susceptible to cleavage in DNA obtained from uncultured
embryos (not shown). Methylation analysis of these sites in
the Normal 2-5b, Normal 3-3b, and Normal 3- 5b cells
indicated that most sites were partially methylated (lanes 5
through 13).
142
Figure 4-8. Methylation analysis of the HS9 promoter
in normal and MatD17 cell lines.
Methylation status of the H19 promoter region was analyzed
for the Normal Cl 1, MatDi7 1-1 a. and MatDi7 1 -6b subclones.
DNAs were purified and restricted with Hind III and Mspl or
HpaU. Southern blot analysis was performed using a 1.0 kb
Bam HI to Xbal fragment (probe 1, Fig. 4-5 A) as a probe.
Results are summarized in Fig. 4-5.
143
1 44
c n
a >
(O
0
1
M
U
I
0 2
X T
IT
Hind lll/Msp I
Hind lll/Hpa I I
Hind lll/Msp I
Hind lll/Hpa I I
Hind III
Hind lll/Msp I
Hind lll/Hpa I I
Figure 4-9. Increased methylation of the maternal
H I 9 promoter is associated with reactivation of the
maternal Igf~2 allele.
Methylation status of the HI 9 promoter region was performed
for the Normal Cl 1, Normal 2-5b, Normal 3-3b, Normal 3-5b,
MatDi7 l-6b, and MatDi7 1-la subclones. DNA was extracted
and restricted with Bam HI and one of the following
methylation-sensitive enzymes; Hhal , Hae II. or 5m al.
Southern blot analysis was performed using a 1.5 kb
Xha\IBamH\ fragment (probe 2, Fig. 4-5A) as a radiolabeled
probe and the autoradiogram is shown. Results are
summarized in Fig. 4-5.
145
9P I
— t
'
f
Bam H I
* 0
t
I
Bam Hl/Hha 1
U
«
1
BamHVHaa II
*
•
BamHVSma 1
(I)
•
»
t
Bam Hl/Hha 1
Oi
•
Bam HVHaa II
1 i i
Bam Hl/Sma 1
»
1 •
Bam Hl/Hha 1
« «
Bam HVHaa II
o
1 1 i
BamHVSma 1
-
• I i
Bam HVHha 1
I
«
BamHI/Haa II
w
• 1 •
Bam HVSma 1
* ■
I I
Bam Hl/Hha I
t*
•
i
Bam HVHaa 1 1
o»
I 1
BamHVSma 1
/ / / < k k
o o o
s s s
r r r
1.S
1 .7 ^
to
d »
i
5
•
Bam HI
» ‘t •
BamHVHhat
tf> «
«
Bam HVHaa II
M
o
1
I i
Bam Hl/Sma 1
V
9
I
"J
I
Therefore, the reduced or lack of expression of H19 (see
Fig. 4-1) in these clones was not a result of hypermethylation
of the H I9 promoter. Figure 4-9 also shows that the H h a I,
Hae II, and Sma I sites in the revertant MatDi7 1-la clone had
undergone de novo methylation in culture when compared to
the MatDi7 l-6b clone which was much less active in Igf-2
expression (compare lanes 14, 15, and 16 to lanes 18, 19, and
20). Indeed, the pattern in the revertant MatDi7 1-1 a clone
was quite similar to that in the normal cells and showed
de novo methylation of sites within the CpG island of the H I 9
p ro m o ter.
Results of the methylation analysis summarized in Fig. 4-5
indicate that spontaneous expression of the maternal Igf-2
allele in the MatDi7 1-la cells was accompanied by de novo
methylation of the region 5' to the lgf-2 promoter and within
the H 19 promoter. It is important to note that the level of
igf-2 expression in the MatDi7 1-1 a cells was approximately
30 % of the level in Normal Cl 1 cells (data not shown) which
may account for some of the sites not becoming fully
methylated. Therefore, the acquired changes in methylation of
both the 5' region to Igf-2 and the H I9 promoter may be
significant and may be directly associated with reactivation of
the maternal Igf-2 allele. The maternal HJ9 allele in the
revertant clone has therefore acquired a modification pattern
147
similar to that of the imprinted paternal allele, and the
normally silent maternal Igf-2 allele has acquired a
modification pattern similar to that of the active paternal
allele.
DISCUSSION
The reciprocal nature of imprinting of lgf-2 and H19 has
always been intriguing (Surani, 1993). With the exception of
postnatal choroid plexus and leptomeninges, where only Igf-2
is expressed (Bartolomei et al., 1993; DeChiara et al., 1991),
there is a close coordination between their temporal and
spatial patterns of expression (Lee et al., 1990; Poirier et al.,
1991). This has lead to the suggestion that their expression
may be functionally and mechanistically related (Sasaki et al.,
1992; Surani et al., 1993; Bartolomei et al., 1993). The
molecular mechanism for the apparent coordinate expression
of Igf-2 and H I9 from separate parental chromosomes was
investigated in somatic cell cultures derived from normal
embryos or their littermates with maternal uniparental disomy
for distal chromosome 7. The somatic cell cultures provide a
novel in vitro mode) system that allows for the examination of
dynamic aspects of regulation of Igf-2 and H I 9, a major
advantage over in vivo and tumor studies which generally
allow analysis under static conditions.
148
In the studies using somatic cell cultures derived from a
normal embryo, both tgf-2 and HI 9 were coordinately
upregulated in cells forced to become quiescent either by
confluence arrest or serum starvation. These results are
consistent with those obtained from previous experiments in
our laboratory and others which showed an increased level of
Igf-2 and H19 mRNA in quiescent cells (Davis et al., 1987;
Pachnis et al., 1988; Eversole-Cire et. al., 1993). Induction of
the quiescent serum-starved cells to proliferate by the
readdition of serum resulted in the coordinate downregulation
of both genes. A class of genes called growth-arrest-specific
(ga.?), isolated from growth-arrested mouse fibroblasts,
demonstrate similar behaviors to Igf-2 and H I9 in response to
cell growth (Schneider et al., 1988). However, the kinetics of
mRNA accumulation and disappearance vary among the 6
members of this class. The very similar patterns of expression
of Igf-2 and H19 upon the induction of cell growth and growth
arrest may therefore result from a specific regulatory
mechanism shared by these two genes and not from a general
phenomenon affecting the expression of many different genes.
Clonally derived cell lines from normal embryos are useful
to study the expression of these genes, since the regulation of
both Igf-2 and H I9 expression appears to be dependent on cell
149
growth status. Cells derived directly from tissues in vivo may
present a problem since tumors could contain a heterogeneous
mix of cells which may be in different stages of proliferation.
The issue of whether or not these genes compete for a
regulatory element which affects the relative levels of
expression of Igf-2 and H19 from the different parental
chromosomes may therefore be better addressed using a clonal
population of tissue culture cells in which growth status can
carefully be controlled.
Subclones of normal cells which had reduced or absence of
expression of both Igf-2 and H19 were also isolated. The
concomitant loss of expression of both Igf-2 and H19 may
result from the inactivation or loss of a trans-acting regulatory
factor(s) required for expression of the Igf-2IHJ9 locus.
Perhaps it is this putative factor which is itself
growth-regulated and as a result, Igf-2 and H I9 exhibit similar
growth-regulation (see Fig. 4-10).
In the MatDi7 cell lines, the maternal Igf-2 allele which was
expressed at least 50 fold less than in normal cells was also
found to be upregulated upon growth arrest induced by
confluence or serum deprivation. The expression of the
maternal HI9 allele, however, was upregulated only upon
confluence arrest. The failure to observe the same response of
150
Figure 4-10. Proposed model for mechanism of Igf-2
and HI 9 genomic imprints.
The model is modified from that originally proposed by Sasaki
et al., 1992 and Bartolomei et al.( 1993. Igf-2 and H I9 are
depicted as open rectangles and the two H19 enhancer
sequences are shown as stippled polygons. Closed circles
indicate sites of methylation and open circles indicate sites
without methylation. Triangles represent the putative
growth-regulated transcription factor(s) which activate both
genes so that cells that do not express these factors express
neither gene. Arrows signify transcription with thickness of
arrow representing relative levels of transcription. Curved
arrows indicate interaction between the enhancer/factor
complex with either the Igf-2 or H 19 promoters. The model
predicts that the lgf-2IHI9 regulatory factor(s) becomes
upregulated in response to growth arrest and interacts with a
regulatory element required for transcription of the Igf-21HI9
locus. The regulatory element is depicted in the model as the
H I 9 enhancer. The factor/enhancer complex may interact with
the maternal HI 9 promoter which is unmethylated and
transcriptionally competent or the paternal Igf-2 promoters (as
illustrated for the Normal Cl 1 and MatDi7 1 -6b subclones).
Upon de novo methylation of the maternal H I9 alleles and the
151
region 5' to Igf-2, the Igf-2 and H I9 promoters now compete
for the complex resulting in reciprocal expression patterns for
these genes. The proximity of the H19 promoter to the
enhancers may cause preferential interaction of the
factor/enhancer complex to this promoter.
152
Normal
MatDi7
Mat D1 7
Growth
Arrest
153
H I9 expression to serum deprivation in the MatDi7 cells may
result from the higher expression levels of H I9 since these
cells contain two copies of the maternal allele. In the revertant
MatDi7 cells the expression pattern of the H19 allele was found
to be nearly a reciprocal expression pattern to that of Igf-2.
Perhaps the increased expression of the maternal Igf-2
allele(s) in the revertant MatDi7 cells interferes with the
expression of HI 9 leading to the observed reciprocity. Without
knowing whether one or both maternal Igf-2 alleles have
acquired the ability to express Igf-2 in the MatDi7 cells, the
interpretation of these observations is difficult. Further
clarification awaits the analysis of expression of the Igf-2/H 19
locus from a single chromosome.
Recent data support DNA methylation as a potential
mechanism for initiating or reinforcing imprinting of the
Igf-2/HI9 locus. Parental-specific differences in the
methylation of the region upstream of Igf-2 and of the H I 9
promoter have been identified (Sasaki et al., 1992;
Ferguson-Smith et al., 1993; Bartolomei et al., 1993;
Brandeis et al., 1993). It has also been shown that lgf-2 and
H I9 are no longer expressed in a parental-specific manner in
mice which are deficient in DNA methyltransferase activity (Li
et al., 1993). In this regard, the observation that reactivation
of the maternal Igf-2 allele in somatic cell culture was
154
associated with de novo methylation of the region 5' to Igf-2
and of the H19 promoter is particularly interesting.
The result of this methylation was that the maternal allele
in the revertant MatDi7 1-la cells had adopted the methylation
pattern characteristic of the paternal allele (summarized in
Figs. 4-5 and 4-10). The mechanism of this de novo
methylation is not clear, but may be related to the well-known
capacity of CpG islands to become methylated during cellular
immortalization (Jones et al., 1990; Antequera et al., 1990). If
this is correct, then the MatDi7 cell system may be particularly
helpful in determining how imprints are applied or reinforced
at the molecular level. Several recent studies have shown ’loss
of imprinting" in Wilms' tumors resulting in overexpression of
Igf-2 (Rainier et al., 1993; Ogawa et al., 1993). Since many
cancers show aberrant methylation of CpG islands (Baylin et al.,
1991; Jones and Buckley, 1990) it is interesting that recent
studies have now shown an association between "loss of
imprinting" of Igf-2 and increased methylation of the promoter
of the maternal HI9 in tumors (Steenman et al., 1994; Moulton
et al., 1994). The cell cultures may be behaving in a similar
fashion to the tumors with the added advantage that they
allow for a temporal and dynamic approach to issues of
coordinate gene control.
155
Parental-specific expression of HI 9 may be controlled in
part by increased methylation and heterochromatinization of
the repressed paternal allele (Ferguson-Smith et al., 1993;
Bartolomei et al., 1993). However, the failure to identify sites
within the promoter and coding region of H I9 which are
differentially methylated in the gametes and during early
embryogenesis suggests that the increase in methylation and
condensation of the paternal gene may be a secondary control
mechanism and may not be the primary imprinting event
(Ferguson-Smith et al., 1993; Brandeis et al., 1993). Repression
of the maternal Igf-2 allele does not appear to involve
increased methylation and heterochromatinization of the
promoters since both the paternal and the maternal alleles are
relatively unmethylated and equally accessible to nucleases in
this region (Sasaki et al., 1992). However, there is one site
located in the region 5' to Igf-2 which exhibits allele-specific
methylation and which may serve as a primary imprinting
signal (Brandeis et al., 1993). These results now show that
increased methylation of this site (Hp3) as well as other nearby
sites is correlated with increased expression of the Igf-2 locus.
Perhaps, as previously suggested (Surani, 1993), increased
methylation of this region prevents binding of a repressor
molecule thereby allowing transcription of the maternally
derived Igf-2.
156
Transcriptional interference is a mechanism which has been
shown to regulate the expression of several eucaryotic genes
(Proudfoot, 1986; Corbin and Maniatis, 1989). The mechanism
of transcriptional interference is not known, however, studies
performed in vitro have shown that the movement of the RNA
polymerase molecule through an active transcription complex
may disrupt it. Since Igf-2 and HJ9 are closely linked and are
in the same transcriptional orientation, perhaps the reciprocal
imprinting of Igf-2 and H I 9 results from transcription of the
paternal Igf-2 allele which prevents transcription of the
paternal H19. Since the maternal Igf-2 allele is repressed it
would not interfere with transcription of the maternal HI9. It
is also possible that additional gene(s) may be present in the
intergenic region between Igf-2 and HI 9 whose transcription
may affect the expression of the Igf-2!H19 locus.
Regulation of gene expression has also been shown to be
affected by the presence of antisense transcripts. Although, it
was previously shown that antisense Igf-2 transcripts were not
present in the 3' region of the gene (Eversole-Cire et al., 1993)
it has recently been shown that antisense transcripts are
present in the intergenic region between Igf-2 and the insulin
gene (Rivkin et al., 1993). Antisense transcription in the 5'
region of Igf-2 may interfere with expression of the maternal
Igf-2 allele by either physical hindrance of multiple
157
transcription complexes on both DNA strands, by altering the
stability of the Igf-2 transcripts, or by preventing their
translation.
Sasaki et al. (1992) and Bartolomei et al. (1993) have also
proposed a promoter/enhancer competition model to explain
the reciprocal imprinting of Igf-2 and H I 9 , the essential
features of which are given in Fig. 4-10. The model, based on
developmental regulation of B-globin in chickens (Foley and
Engel, 1992), proposes that the enhancers located 3' of H19
(Yoo-Warren et al., 1988) can interact with either the H I9 or
Igf-2 promoters in cis to initiate transcription, thus ensuring
that only one gene is transcribed from a given parental
chromosome at a particular time. Methylation of the H I 9
promoter on the paternal chromosome may prevent its
interaction with the enhancer thus ensuring transcription of
the upstream Igf-2 promoters. The data presented in this
study provides evidence in support of the promoter/enhancer
competition model while also providing novel information on
the regulation of Igf-2 and HI9 expression possibly by the
same transcription factor(s).
158
CHAPTER 5
Activation of an Imprinted Human H 1 9 in
Human-Mouse Somatic Cell Hybrids
INTRODUCTION
Genomic imprinting is thought to play a role in the genesis
of several tumor syndromes. The Beckwith-Wiedemann
syndrome, for example, which is characterized by the
overgrowth of many tissues and may predispose afflicted
individuals to embryonal tumors including Wilms' tumor of the
kidney, has been linked to a region of chromosome lip (Waziri
et al., 1983; Turleau et al., 1984; Koufos et al., 1989; Ping
etal.,1989) which harbors two genomically imprinted genes,
Igf-2 and HI9 (Glaser et al., 1989; Ohlsson et al., 1993;
Giannoukakis et al., 1993; Zhang et al., 1993). A loss of
heterozygosity (LOH) of maternal 11 p i 5 (Mannens et al., 1988)
and uniparental disomy of paternal llp lS (Henry et al., 1991)
have been observed in Wilms’ tumors. Recent studies also
report that a majority of Wilms' tumors not exhibiting these
genetic alterations were found to have lost or "relaxed" the
Igf-2 and/or H19 imprints (Rainier et al., 1993; Ogawa et al.,
1993). Together these findings suggest that Igf-2 and H19 are
involved in the genesis of this disease. If the imprints of the
two genes are mechanistically linked and their expression
159
coordinately regulated then any of the above mentioned
genetic alterations may perturb the expression of the
Igf-2iHi9 locus and lead to the numerous manifestations of
this syndrome.
The human igf-2 and H19 are imprinted similarly to the
mouse homologs with expression of the paternal igf-2 and the
maternal H19 alleles (Ohlsson et al., 1993; Giannoukakis et al.,
1993; Zhang et al., 1993; DeChiara et al., 1991; Bartolomei et al.,
1991). The synteny observed between the human and mouse
chromosomal regions harboring the igf-2/H19 locus (Glaser
etal., 1989) and the conservation of the imprinting event
suggests that perhaps other features such as the transcriptional
regulation of the genes have also been conserved through
evolution.
Previous results have led to the hypothesis that the mouse
igf-2 and H I9 compete for transcriptional regulatory elements
and/or factor(s) and therefore cannot be simultaneously
transcribed from a single chromosome (Sasaki et al., 1992;
Bartolomei et al., 1993; Eversole-Cire et al., submitted). The
analysis of the coordinate regulation of igf-2 and H i 9,
however, is difficult in cells containing more than one copy of
the Igf-2fH19 locus. Therefore, human-mouse somatic cell
hybrids containing a single copy of human chromosome 11
160
were used to study the regulation of the human lgf-2 and H 1 9
in a mouse genetic background.
MATERIALS AND METHODS
Cell Culture
Cells of the human rhabdomyosarcoma RD line (McAllister
et al., 1969), at passage 53, were cultured in Dulbecco's
Modified Eagle Medium (DMEM) containing 10%
heat-inactivated fetal calf serum (HIFCS; Tissue Culture
Biologicals), penicillin (100 Ll/ml), and streptomycin (100
ug/ml; GIBCO). Cells from the C3H10T1/2 Cl 8 ([10T1/2],
Reznikoff et al., 1973A) line, at passage 16, were grown in
Eagle's basal medium supplemented as described above.
Microcell hybrid cells (Thayer and Weintraub, 1990) were
grown in medium containing 15% HIFCS for RNA isolation and
in medium containing 10% HIFCS for DNA isolation. Media was
supplemented with 250 ug/ml Geneticin (G418 sulfate; GIBCO).
BUdR Treatment
Microcell hybrids were seeded at 0.5x106/100-mm dish for
RNA isolation and 200 cells/60-mm dish for cytotoxicity
studies. Cells were treated with lxlO '6 M or 1x10-5 M BUdR
(Sigma) approximately 24 h after seeding. A stock solution of
BUdR was made in PBS and filter sterilized and PBS only was
161
added to untreated control dishes. The cells were washed
twice with unsupplemented medium to remove the drug 24 h
after addition of the analog and fresh medium added. Medium
was changed three days thereafter for dishes to be used for
RNA isolation and approximately 8 days after removal of the
drug for dishes to be used for cytotoxicity studies. Colonies
were stained and counted on the twelfth day after drug
removal. BUdR treatments were performed in a darkened
room since the drug is light sensitive. RNA was isolated 6 days
after removal of the drug.
Isolation of Nucleic Acids
DNA was isolated from cells in the late log phase of cell
growth and resuspended in TB (10 mM Tris-HCI [pH8.0],
0.1 mM EDTA [pH 8.0]). Total cellular RNA was isolated from
cultures in the logarithmic or stationary phases by the method
of Chomczynski and Sacchi (1987).
Polymerase Chain Reaction (PCR)
An aliquot of DNA (100 ng) was mixed with the following
reagents in a total volume of 50 ul for the PCR: lx PCR buffer
(50mM KC1, 10 mM Tris-Cl [pH 8.3], 1.5 mM MgCl2, 0.01%
[wt/vol] gelatin [Boehringer Mannheim], each dNTP at 200 uM,
the 5' and 3' primers at 1 uM. Samples were heated to 94°C
for 10 min prior to the addition of Tag polymerase
162
(2U; Boehringer Mannheim). An overlay of 50 ul of sterile
mineral oil was added to each sample, and the samples were
then subjected to 30 cycles of PCR. For the Igf-2 GT
polymorphism, each PCR cycle consisted of incubation periods
of 1 min at 94°C, 30 sec at 55°C, and 1 min at 72°C. PCR
products were analyzed by electrophoresing an aliquot on a 1%
agarose gel in lx TBE (0.089 M Tris-borate, 0.003 M EDTA).
Southern blot analysis was then performed using a
digoxigenin-labeled GT repeat oligomer as a probe in
accordance with manufacturer's recommendations.
Reverse Transcription (RT)-Polymerase Chain Reaction
(PCR)
RT-PCR was performed as previously described with the
following conditions for the various reactions (Chapters 3 and
4; Eversole-Cire et al., 1993).
Each PCR cycle using the mouse Igf-2 (Mu Igf-2) and mouse
HI9 (Mu HI9) primers consisted of incubation periods of
1 min at 94°C, 30 sec at 65°C, and 1 min at 72°C. PCR was
performed for 28 cycles using a volume of cDNA corresponding
to 100 ng of RNA template.
Each PCR cycle using the mouse MyoD (Mu MyoD) primers
consisted of incubation periods of 1 min at 94°C, 1 min at 60°C,
163
and 1 min at 73°C. PCR was performed for 29 cycles using a
volume of cDNA corresponding to 100 ng of RNA template.
Each PCR cycle using the human Igf-2 (Hu Igf-2) primers
consisted of incubation periods of 1 min at 94°C, 30 sec at 60 °C 1
and 1 min at 72°C. PCR was performed for 26 cycles using a
volume of cDNA corresponding to 100 ng of RNA template.
Each PCR cycle for the human HI9 (Hu HI9) consisted of
incubation periods of 30 sec at 94°C, 30 sec at 65°C, and 1 min
at 73°C. PCR was performed for 35 cycles using a volume of
cDNA corresponding to 250 ng of RNA template.
Each PCR cycle for the GAPDH (mouse) consisted of
incubation periods of 1 min at 94°C, 1 min at 60°C, and 1 min
at 73°C. PCR was performed for 21 cycles using a volume of
cDNA corresponding to 100 ng of RNA template.
164
O lig o m ers
Oligomers for amplification of both the Mu and Hu Igf-2 and
H19 as well as for the Mu MyoD were designed to span an
intron so that amplification of cDNA was not contaminated with
DNA sequences.
Igf-2 GTl(S') 5'CTC ATA CTT TAT GCA TCC CCG3'
Igf-2 GT2 (3 ) 5’ GCC TGA TCC ATA CAG ATA TCG3’
Igf-2 GT (internal oligo) 5’GTG TGT GTG TGT GTG TGT GTG3'
Mu Igf-2 OP1 (5' oligo) 5'AGC GGC CTC CTT ACC CAA CTT CAG3'
Mu Igf-2 OP8 (3‘ oligo) 5'AAG GCC TGC TGA AGT AGA AGC CG3’
Mu Igf-2 OPS (int. oligo) 5'ATC CCA GTG GGG AAG TCG ATG
TTG3’
Mu HI9 OP1 (5* oligo) S'TCA TAG CAC CCA CCC ACC CC3'
Mu HI9 OP2 (3' oligo) 5’ GGG AAG GCG AGG CCT CAA GC3’
Mu HI9 OP3 (int. oligo) S’ GGG CCT TTG AAT CCG GGG AC3’
Myo PQ3 (5’ oligo) S AGA CCA CCA ACG CTG ATC GCC GCA A3’
Myo PQ2 (3’ oligo) 5GTA GTA GGC GGT GTC GTA GCC ATT C3’
Myo PQ1 (int. oligo) S’AAT GAG GCC TTC GAG ACG CTC AAG C3’
Hu Igf-2 1 (S’ oligo) S’ TG TCG ACC CCT CCG ACC GTG3’
Hu Igf-2 2 (3’ oligo) S’GGG CTG CAG ACT TGC GGC AG3’
Hu Igf-2 3 (internal oligo) S’ACC TGG AAG CAG TCC ACC CA3'
Hu H19 1 (5’ oligo) S’ GAC TCA GGA ATC GGC TCT GG3*
Hu H19 2 (3’ oligo) S’ GCC AAG GTG GCT CAC ACT CAT
Hu H19 3 (internal oligo) S’ CGC TGG AGG AGC TCA GCT CT3’
GAPDHX (S' oligo) S’ CAG CCT CGT CCC GTA GAC AAA ATG G3’
GAPDH2 (3’ oligo) 5TTC TGG GTG GCA GTG ATG GCA TGG A3
G A P D m (internal oligo) S’ CGG TGC TGA GTA TGT CGT GG3’
165
RESULTS
Somatic Cell Hybrids Containing Single Copies of the
Human Igf-2/H 19 Locus
It has been hypothesized that Igf-2 and H I 9 may be
regulated by the same transcription factor(s) and possibly by a
shared regulatory element (Sasaki et al., 1992; Bartolomei
el al., 1993; Eversole-Cire et al., submitted). Transcription of
Igf-2 and H I9 from different parental chromosomes may
result from competition between Igf-2 and H I 9 for a shared
regulatory domain. The expression of the Igf-2IH19 locus
must be analyzed from a single chromosome to determine if
enhancer competition results in only one of the genes n o rm a lly
being transcribed from a single parental chromosome.
Therefore, human-mouse somatic cell hybrids containing a
single copy of human chromosome 11 which harbors the
Igf-2/HJ9 locus were used to study the transcription of Igf-2
and H19 from the same DNA double strand.
Somatic cell hybrids containing single human chromosomes
were derived by Lugo et al. (1987). The retroviral vector,
ZlP-neoSV(X) 1, was used to integrate the neomycin resistant
gene, a dominant selectable marker, into individual human
chromosomes in human fibroblasts. Diploid human fibroblasts
166
were originally derived from a sample of foreskin and cultured
to passage 10 at which time the cells were transduced. The
hybrid cell line HDm-18, containing a single copy of human
chromosome 11, was generated by fusing human skin
fibroblast microcells containing human chromosomes with an
integrated neo-containing retroviral vector with Swiss 3T6
mouse cells (Fig. 5-1). Retention of human chromosomes in
the HDm-18 hybrid was accomplished by selection in the
antibiotic G418.
10T1/2 microcell hybrids ( 10T1 /2( 11 n) clones 7, 2, 5, and
6) were formed by fusing HDm-18 microcells with 10 T 1/2 cells
by Thayer and Weintraub (1990). Since the human
chromosome 11 contained in the 10Tl/2( 11 n) hybrids passed
through a Swiss 3T6 mouse cell background, it was important
to determine if passage of the human igf-2 and H I9 through a
different genetic background affected the expression of the
igf-2 and H I9 in the 10T1/2 microcells. Therefore, a 10T1/2
hybrid, 10HSm-2, which also contains chromosomes 4 and 8, in
which human chromosomes were directly transferred from
primary human fibroblasts was also included in this study.
The 10T1/2 microcell hybrids have previously been shown to
form muscle when cultured in differentiation medium (Thayer
and Weintraub, 1990). Since regulation of igf-2 and H I 9
expression may be affected in myogenic cells
167
Figure 5-1. Diagram of human-mouse somatic cell
hybrid generation.
Human fibroblasts were infected with the vector
ZlP-neoSV(X) 1 and selected with G418 to obtain cells
containing human chromosomes with an integrated neo
(indicated by darkening of the open figure 8). Microcells
formed from human fibroblasts containing human
chromosomes with an integrated neo were fused with Swiss
3T6 mouse cells to generate the HDm-18 hybrid which
contained a single copy of chromosome 11 (Lugo et al., 1987).
HDm-18 hybrid microcells were fused with C3H10T1/2 cells to
generate the 10T 1/2(1 In) clones 2, 5, 6, and 7 (Thayer and
Weintraub, 1990). The 10HSm-2 hybrid was formed by fusing
microcells from primary human fibroblasts containing
chromosomes with an integrated neo with C3H10T1/2 cells. In
addition to containing a single human chromosome 11, the
10HSm-2 hybrid also contained human chromosomes 4 and 8.
All downward arrows signify selection with G418.
168
Human Fibroblast
ZlP-neoSV(X) 1
Select with 0416
C3H10T1/2
Human chromosome
Swiss 3T6 Mouse Cells
with integrated neo
10HSm-2 HDm-18
10T1/2(11 n)
C3H10T1/2
clone 2 clone 6 clone 5 clone 7
169
(Tollefson et al., 1989; Pachnis et al., 1988), the degree of
myogenesis and the levels of expression of the mouse M yoD
were also assessed in the microcell hybrids to determine if the
myogenic capability of the hybrids altered the normal
expression pattern of either Igf-2 and/or H I9 (Pachnis et al.,
1988; Eversole-Cire et al., 1993).
Retention of Human Igf-2 In Microcell Hybrids
The site of integration of the retrovirus was determined to
be at 11 q 1.4-2.2 (Lugo et al,, 1987). Since the Igf-2/HI9
locus is located on the p arm of chromosome 11, it was first
necessary to demonstrate that this region was not lost during
cell culturing, since fragmentation of chromosomes can occur in
microcell hybrids when under selection (Leach et al., 1989).
Retention of the human genomic lgf-2 sequences in the
microcell hybrids is shown in Fig. 5-2. The mouse embryonic
cell line 10T1/2 and the human rhabdomyosarcoma cell line RD
were used to demonstrate amplification of the mouse (Mu) and
human (Hu) Igf-2 and HI9 sequences, respectively (see lanes 1
and 2). Similarity between the human and mouse sequences in
the amplified region allowed for the simultaneous amplification
of both the human and mouse sequences in the hybrids. All of
the hybrid lines retained the introduced Hu Igf-2 as judged by
the amplification of the human sequence in these samples (see
170
Figure 5-2. Human lgf-2 genomic sequence Is retained
in human-mouse somatic cell hybrids.
PCR performed to demonstrate the presence of the Hu lgf-2
sequence using primers which flank the GT polymorphism
located in the 3' region of the gene. Conditions for the PCR
were such that both the human and mouse sequences were
amplified although the mouse sequences were less efficiently
amplified. Lane 1 shows the size of the amplified product from
the Mu lgf-2 sequence in 10T1/2 cells and lane 2 shows the
size of the amplified product from the Hu lgf-2 sequence in RD
cells. Lanes 3 through 8 demonstrate the presence of both the
Hu lgf-2 and Mu Igf-2 in the somatic cell hybrids. Lane 9
contained a sample in which no DNA was added to the PCR to
assay for possible contamination. PCR products were analyzed
by Southern blot analysis, hybridized with a
digoxigenin-labeled GT repeat oligomer, and exposed to X-ray
film.
171
10T1/2
RD
HDm-18
10HSm-2
10T1/2(11n)7
10T1/2(11n)2
10T1/2(11n)5
10T1/2(11n)6
no DNA
lanes 3 through 8). Although the intensities of the amplified
products may suggest that the gene dosage of the Hu Igf-2 was
greater than that of the Mu Igf-2t the difference in intensities
was related to the fact that the PCR primers were designed
complementary to the human Igf-2 sequences and not the
mouse Igf-2 sequences.
Expression of the Mouse Igf~2 and H I9 in Microcell
H yb rid s
The expression of the Mu Igf-2 and Mu HI9 were first
analyzed using an RT-PCR assay to determine whether the
microcell hybrids expressed the mouse genes appropriately
and were potentially permissive for expression of the human
homologs. (The quantitative RT-PCR assay used was previously
described in Chapters 3 and 4). The expression analysis was
performed on cells in both the logarithmic and stationary
phases of cell growth since it was previously shown that the
mouse Igf-2 and H19 mRNA levels were upregulated in
quiescent cells (Chapters 3 and 4; Davis et al., 1987; Pachnis
et al., 1988; Eversole-Cire et al., 1993).
Expression of Mu Igf-2 and Mu HI9 in the mouse 10T1/2
cell line and the human RD cell line was analyzed to
demonstrate specific amplification of the mouse sequences
173
Figure 5-3. Expression of the mouse !gf~ 2,H 19 , and
M yoD sequences In human-mouse somatic cell hybrids.
RT-PCR analysis of expression levels of the mouse Igf-2, H19 ,
and MyoD in control cell lines and in hybrid cell lines. Lanes 1
and 2 are controls which contained cDNA prepared from
confluent 10T1/2 and RD cells, respectively, to demonstrate
that the primers were specific for the mouse sequences.
Expression of the mouse genes in the microcell hybrid cells
during the logarithmic (log) phase and upon confluency (conf)
is shown in lanes 3 through 8 and 9 through 14, respectively.
Lane 15 contained a negative control in which no cDNA was
added to the reaction to assay for possible contamination. The
number of PCR cycles performed for each primer set was
chosen so that the reactions were carried out under
non-saturating conditions (as determined from cycle curves
generated using cDNA prepared from mouse fibroblast cell
lines). PCR was performed for 28 cycles for amplification of
Mu Igf-2 and Mu HI9 and 29 cycles for MyoD using a volume
of cDNA which corresponded to 100 ng of template RNA. PCR
products were analyzed by Southern blot analysis, hybridized
with digoxigenin-labeled internal oligomers, and exposed to
X-ray film.
174
175
Mu Igf2 ■ *
Mu H19 ■ *
MyoD1 - ►
C \J
1 2 3
HDm-18
Log Conf
8 9 10 11 1 2 1 3 14 1 5
using this assay (Fig. 5-3, lanes 1 and 2, respectively). The
presence of amplified products in only the mouse 10T1/2
sample (lane 1) demonstrates specificity of the assay for mouse
sequences. The relative levels of Mu lgf-2 and Mu H I 9
expression were similar based upon: 1- the same cDNA pool
was used to amplify both genes, 2- the same amount of
starting template and the same number of PCR cycles were
used to amplify both genes, 3- similar levels of ethidium
bromide staining of the amplified products for both genes in a
given sample.
Expression levels of Mu Igf-2 in the donor cell line, HDm-18,
and in the primary transfer hybrid cell line, 10HSm-2, were
upregulated in confluent cells (lanes 3, 9 and 4, 10,
respectively). Interestingly, the expression of Mu H I 9,
however, was only growth-regulated in the 10HSm-2 hybrid.
Of the 10T 1/2(1 In) hybrids, only clone 2 demonstrated an
increase in the level of Mu lgf-2 and Mu H I9 expression in
quiescent cells (lanes 6 and 12). In clones 7, 5, and 6, the
expression levels of Mu lgf-2 as well as Mu H19 were elevated
in proliferating cells and did not undergo further increases in
confluent cells (compare lanes 5, 7, 8 and 11, 13, 14).
The expression of the Mu MyoD was also analyzed to
determine if the degree of muscle formation in the microcell
176
hybrids affected the regulation of Mu Igf-2 and/or Mu H 19.
Mu MyoD was not expressed in the nonmyogenic 10T1/2 cell
line (lane 1). Although the RD cell line expresses the Hu MyoD
(Tapscott et al., 1993) an amplified product was not present for
the RD sample demonstrating that the assay was specific for
amplification of the Mu MyoD. Of the hybrids which
demonstrated growth regulation of both the Mu Igf-2 and
Mu HI9, Mu MyoD was expressed at low levels in growing
10HSm-2 hybrid cells yet high levels in confluent cells (lanes 4
and 10) and was not expressed in the 1 0 T l/2 (lln ) clone 2
hybrid (lanes 6 and 12). 1 0 T l/2 (lln ) clones 7, 5 and 6,
exhibited high levels of Mu MyoD during proliferation and
upon confluency (lanes 5, 7, 8, and 11, 13, 14). The HDm-18
hybrid expressed low levels of MyoD (lanes 3 and 9).
The elevated level of Mu MyoD mRNA during cell growth is
indicative of the presence of proliferating myoblasts (Davis
et al., 1987). Since the Mu H I9 is actively expressed during
the proliferation of myoblasts and does not undergo a further
increase upon the formation of myotubes (Pachnis et al., 1988),
it is possible that the myogenic capability of the 1 0 T l/2 (ltn )
clones affected the normal growth regulation of the Mu H I 9 . It
is not yet clear what caused the lack of growth regulation of
Mu Igf-2 in these 10Tl/2( 1 In) hybrids but if Igf-2 and H I 9
are coordinately regulated then perhaps the elevated level of
177
Mu HI 9 expression during cellular proliferation may lead to
the increased level of Mu Igf-2 expression that was also
observed during cell growth.
The Level of Human igf-2 Expression is Greater than
that of Human H19 in Somatic Cell Hybrids
Expression of the Hu Igf-2 and Hu H I9 was next analyzed in
the microcell hybrids (Fig. 5-4). It is important to note that the
levels of expression of the Hu Igf-2 and Hu H I9 were such that
to detect Hu H I9 sequences 2.5 times the initial input of
template used for amplification of the Hu Igf-2 were used and
an additional 9 cycles were performed for the PCR. Initial PCR
conditions were established using cDNA prepared from mRNA
isolated from normal human urothelium. Amplification of the
Hu Igf-2 and H19 performed using the same amount of starting
template and cycle numbers yielded more amplified product
for lgf-2 based on ethidium bromide staining of the PCR
products (data not shown).
Amplification was performed on samples from the mouse
10T1/2 and human RD cell lines to demonstrate specific
amplification of the human sequences (lanes 1 and 2). A PCR
product was only observed for the RD sample indicating that
1 78
Figure 5-4. Expression of human Igf~2 and H19 in
human-mouse somatic cell hybrids.
RT-PCR analysis of human lgf-2 and H19 expression levels in
control cell lines and in microcell hybrids. Lanes 1 and 2 are
controls which contained cDNA prepared from confluent
10T1/2 and RD cells, respectively, to demonstrate that the
primers used resulted in specific amplification of the human
lgf-2 sequence. Expression of the genes in the hybrid cell lines
during the logarithmic (log) phase of cell growth and upon
confluency (conf) is shown in lanes 3 through 8 and 9 through
14, respectively. Lane 15 is a control in which no cDNA was
added to the PCR to assay for possible contamination of the
reactions. PCR was performed for 26 cycles for amplification of
lgf-2 and for 35 cycles for amplification of HI 9. Initial input
of RNA template for amplification of H19 was 2.5 times that
used for amplification of lgf-2 sequences (i.e. cDNA volume for
PCR corresponded to 100 ng of template RNA for Igf-2 and
250 ng for H I 9). Levels of mouse GAPDH expression were
determined to demonstrate approximately equivalent levels of
RNA template in the various cDNA pools used. Products were
analyzed by Southern analysis, hybridized with
digoxigenin-labeled internal oligomers specific for the Igf-2,
H I 9, and GAPDH sequences, and exposed to X-ray film.
179
180
Hu Igf2
Hu H19
Mu GAPDH
3 4
10HSm-2
Log Conf
N ( \ J W
c " c ~ c" c“
N N Ifl m
? £ £ £
C M £
w _ co i C C C- C- <
C M CM C M C M t— C CM CM CM CM Z
O O O O ^ O O O O O o
6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5
only the Hu lgf-2 and Hu HJ9 were amplified. The hybrid cell
line, HDm-18 which demonstrated growth regulation of the
Mu Igf-2 and 10HSm-2 which demonstrated growth regulation
of both the Mu lgf-2 and Mu H I 9, also expressed the Hu Igf-2
in a growth dependent manner (lanes 3, 4 and 9, 10). The
10HSm-2 hybrid also expressed very low levels of the Hu H I 9
in the log phase but not upon confluency (compare lanes 4 and
10). Interestingly, the levels of expression of the Hu Igf-2 and
Hu H19 in the primary transfer hybrid cell line, 10HSm-2,
were higher than in the HDm-18 hybrid. Therefore, the Swiss
3T6 cells do not appear to be as permissive for expression of
the human genes as the 10T1/2 cells.
The 10 T 1 /2( 11 n) clone 2 expressed low levels of Hu Igf-2
which were increased upon growth arrest (lanes 6 and 12) as
was the Mu Igf-2 in this hybrid. The 1 0 T l/2 (lln ) clones 7
and 6 expressed similar levels of the Hu Igf-2 in both
proliferating and stationary cells (lanes 5, 11 and 8, 14,
respectively) demonstrating that expression of the Hu Igf-2
was not responsive to growth arrest in these two clones. Only
the 10T 1/2(1 In) clone 7 expressed Hu H19 which was not
growth-regulated (lanes 5 and 11). A lack of growth regulation
was also observed for the Mu Igf-2 and Mu H I9 in clones 7
and 6. Surprisingly, 1 0 T l/2 (lln ) clone S did not express a
detectable level of either the Hu Igf-2 mRNA or the Hu H I 9
181
mRNA (lanes 7 and 13) even though both the Mu Igf-2 and
Mu HI 9 were expressed in these cells.
Both the human and mouse Igf-2 and H I9 are reciprocally
imprinted such that Igf-2 is expressed from the paternal
chromosome and H19 from the maternal chromosome.
Therefore, the high level of expression of the Hu Igf-2
compared to that of Hu HI9 in the microcell hybrids was
suggestive that the chromosome captured in the hybrids was of
paternal origin. It has been reported that expression of the
imprinted H19 is generally not detected (Bartolomei et al.,
1993) whereas expression of the imprinted Igf-2 can be
detected using a sensitive RT-PCR assay (Chapter 3; Sasaki et
al., 1992; Eversole-Cire et al., 1993). Therefore, based on
differences between the overall level of Hu Igf-2 and Hu HI 9
expression in the mouse microcell hybrids, the Hu H I9 imprint
may have been maintained in the mouse genetic background.
Activation of an Imprinted Human H19 Gene upon
BUdR Treatment
The Mu Igf-2 can be activated in somatic cell culture by
treatment with the thymidine analog, BUdR (Eversole-Cire
etal., 1993). BUdR, which has been shown to alter the
expression of several genes in cell culture (Tapscott
182
et al., 1989), is thought to act by inducing changes in
chromatin structure after incorporation into DNA. The mouse
and human Igf-2 and H19 are imprinted in a similar manner
indicating that the imprinting event for these genes has been
conserved (Ohlsson et al., 1993; Giannoukakis et a)., 1993;
Zhang et al., 1993; DeChiara et al., 1991; Bartolomei et al.,
1991). It has also been suggested that Igf-2 and H I9 may be
imprinted by a common mechanism (Sasaki et al., 1992;
Bartolomei et al., 1993; Eversole-Cire et al., submitted). If
Igf-2 and H I9 share an imprinting mechanism which represses
the paternal HI9 while allowing expression of the paternal
Igf-2, then perhaps reactivation of the imprinted paternal H I 9
would result in the inactivation of the paternal Igf-2. The
microcell hybrids were, therefore, treated with BUdR in an
attempt to activate the imprinted human H I9 and to determine
if activation of the normally repressed H19 would affect the
expression of Igf-2 from the same parental chromosome.
The effect of BUdR on the expression levels of Hu Igf-2 and
Hu HI9 was assessed using RT-PCR (Fig. 5-5) and quantitated
by densitometric analysis (Table 5-1) to determine if the
analog would induce expression of the Hu HI 9 and repress
expression of the Hu Igf-2 in the microcell hybrids. The
specificity of the assay for amplification of only human
sequences was demonstrated (lanes 1 and 2). An increase in
183
Figure 5-5. Expression of human Igf~2 is
downregulated while expression of human HS9 is
upregulated in somatic cell hybrids upon BUdR
treatment.
RT-PCR analysis of human Igf~2 and H19 expression levels in
control cells and in hybrid cells. Lanes 1 and 2 contained
controls to demonstrate specificity of PCR primers. Expression
levels of lgf-2 and H19 in hybrid cells treated with 10*6 M and
10*5 M BUdR are shown in lanes 4, 7, 10, 13, 16, and 19 and
lanes 5, 8, 11, 14, 17, and 20, respectively. Expression levels in
untreated cells are shown in lanes 3, 6, 9, 12, 15, and 18.
Lane 21 is a no cDNA negative control. RT-PCR was also
performed with mouse G A PD //-specific primers to demonstrate
similar levels of the RNA template in the cDNA pools used. PCR
products were analyzed by Southern blot analysis, hybridized
with digoxigenin-labeled internal oligomers, and exposed to
X-ray film.
1 84
381
BUdR (M)
10T1/2
no cDNA
TABLE 5-1. Effect of BUdR treatment on Igf-2 and
H I9 expression in microcell hybrids
H ybrid BUdR
conc(M)
Igf-2
change in
exp."
H 19
change in
exp."
GAPDH
change in
exp."
HDm-18 0 1.00 1.00 1.00
I O 6 4.38 4.17 1.12
10-5
0.98 2.80 1.13
lOHSm 0 1.00 1.00 1.00
io-6 0.87 5.55 0.98
10-5
0.78 16.26 1.00
1O T1 / 2 (1 1 n)7 0 1.00 1.00 1.00
IO*6 1.52 3.65 0.87
IO '5
1.01 9.37 0.88
10T 1 / 2 (1 1 n)2 0 1.00 1.00 1.00
io-6 0.59 5.90 0.92
IO*5
0.26 9.47 0.92
1 OT 1 / 2 (11 n)5 0 1.00
io-6 1.10
10-5
1.02
1 0 T l/2 ( 1 1 n)6 0 1.00 1.00 1.00
10-6
0.92 12.93 1.00
10*5
0.46 23.61 1.09
a Levels of gene expression were determined by densitometric
analysis of Southern blot hybridizations of the RT-PCR products
with internal oligomers specific to each sequence. Expression
levels were calculated as fold changes in expression in treated
samples compared to untreated samples. Results are from one
experim ent.
186
the expression of Hu Igf-2 and Hu H19 was observed in the
HDm-18 hybrid at the lower dose of BUdR which decreased at
the higher dose of the analog (lanes 3 through 5). Treatment of
the microcell hybrid 10HSm-2 with BUdR resulted in a
dose-dependent 16-fold increase in the expression of the
Hu HI9 (lanes 6 through 8). BUdR treatment of the
10T l/2(lln) clones 7, 2, and 6 (lanes 9 through 14 and 18
through 20) also resulted in a dose-dependent increase,
ranging from approximately 9- to 23-fold, in Hu H 19
expression. A slight increase in Hu H19 mRNA levels was also
observed in clone 5 at the higher dose of the analog (lanes 15
through 17). The dose-dependent increase in expression of
Hu HI 9 was accompanied by a dose-dependent decrease in
expression of the Hu Igf-2, ranging from approximately
1.3-fold to 4-fold, in the hybrids 10HSm-2 and 10T1 /2( 11n)
clones 2 and 6 (lanes 6 through 8, 12 through 14, and 18
through 20, respectively). A decrease in Hu Igf-2 expression
was not observed for 10T l/2(lln) clone 7 (lanes 9 through
11). Expression of the Hu Igf-2 was not detected in 10T1/2
clone 5 (lanes 15 through 17).
The cytotoxic effects of BUdR were analyzed to determine
whether the changes in gene expression resulted from
cytotoxicity (Fig. 5-6). The lower dose of BUdR (IO*6 M) did
not appear to have significant effects on the survival of any of
187
Figure 5-6. Cytotoxicity study for BUdR treatment of
human-mouse somatic cell hybrids.
Solid bars represents the average number of surviving colonies
in control dishes in which PBS was added. Stippled bars
represent the average number of colonies in cultures treated
with 10 6M BUdR. Hatched bars represent the average number
of colonies in cultures treated with 10'5M BUdR. Number of
colonies are an average of surviving colonies from duplicate
dishes.
188
Number o f Colonies
50
■ PBS
□ 10-6 M BUdR
□ 10-5 M BUdR
HDm-18 10HSm-2 10T1 /2( 1 1 n}7 10T1 /2( 1 1 n)2 10T l 11 n)5 1 0T1 /2( 1 1 n)6
00
C O
Somatic Cell Hybrids
the hybrid celts whereas the higher dose of the analog (10's M )
affected the survival of the cells to varying degrees. Although
two of the cell lines which demonstrated increases in response
to treatment with BUdR, 1 OT 1 /2( tin) clones 7 and 2, showed
approximately 33% and 44% survival at the higher dose of
BUdR, respectively, the two hybrids which showed the largest
increase in Hu H I9 expression, lOHSm and 1 0 T t/2 (lln ) clone
6, exhibited 73% and 60% cell survival, respectively.
Therefore, the cytotoxic effects of BUdR were probably not
responsible for the increase in Hu HI9 ex p ressio n .
BUdR treatment has also been shown to interfere with
myogenesis in cell culture (Tapscott et al., 1989) resulting from
inhibition of expression of MyoD by the nucleotide analog.
Therefore, the degree of muscle production in the hybrid cells
treated with BUdR was also visually assessed (data not shown).
All of the clones with the exception of the HDm-18 hybrid
formed varying amounts of muscle which in all cases was
inhibited in a dose-dependent manner in the BUdR-treated
cultures, presumably as a consequence of decreased M yoD
expression. Since myogenesis in 1 0 T l/2 (lln ) clone 7 was
inhibited by BUdR exposure but the expression of Hu Igf-2 did
not decrease, it is likely that the decrease in expression of
Igf-2 in the other muscle-producing hybrids was not a
response to the decreased myogenesis in the BUdR-treated
190
cultures. The decreased muscle production in the
BUdR-treated hybrids should not lead to the increased
expression of Hu H I9 if the human HI 9 is regulated as the
mouse HJ9 in myogenic cell lines (Pachnis et al., 1988). It is
most likely that treatment with BUdR directly affected the
expression of the imprinted HI 9. The decrease in expression of
Hu Igf-2 may be a consequence of the increased expression of
Hu H19 from the same parental chromosome.
D ISCU SSIO N
Igf-2 and H I 9 have been shown to be physically linked in
both the mouse and human genomes (Zemel et al.( 1992).
Recently it has also been demonstrated that Igf-2 and H I9 are
imprinted in the same directions in mouse and human (Ohlsson
et al., 1993; Giannoukakis et al., 1993; Zhang et al., 1993;
DeChiara et al., 1991; Bartolomei et al., 1991). Therefore, it has
been proposed that these two genes lie within an imprinted
domain which allows expression of either the paternal Igf-2 or
the maternal H I9 allele (Zemel et al., 1992). It has been
hypothesized that the reciprocal imprinting of Igf-2 and H I 9 is
a consequence of competition between Igf-2 and H I 9 for a
common regulatory region (Sasaki et al., 1992; Bartolomei et
al., 1993; Eversole-Cire et al., submitted). If this competition
model is correct, the expression of Igf-2 and H I9 should be
191
exclusive from the different parental alleles. Human-mouse
somatic cell hybrids containing single copies of the human
Igf-2fHI9 locus were analyzed to test this theory.
Expression analysis of the microcell hybrids revealed that
the human Igf-2 and H 1 9 were expressed in a mouse genetic
background suggesting that the transcriptional regulation of
these genes has been conserved between these two species.
The considerably higher level of expression of the Hu Igf-2
compared to that of the Hu H 19 suggested that the human
chromosome present in the microcell hybrids was of paternal
origin since Igf-2 is primarily expressed from the paternal
allele while H 19 is repressed. If the human HI 9 imprint has
been maintained in the mouse genetic background this would
indicate that the mouse cells contain all the necessary
information to maintain and propagate the human H I9 imprint.
In most cases, the expression patterns of the mouse Igf-2
and H19 were very similar to one another and also similar to
the human Igf-2 expression pattern in the microcell hybrids.
This result indicates that the human Igf-2 is subject to the
same transcriptional regulation as the mouse Igf-2 and HI9.
Expression of the human genes in a mouse genetic background
indicates that in addition to homology between the mouse and
human Igf-2 and H I9 sequences (Brannan et al., 1990; Rotwein
192
and Hall, 1990), other features of these genes such as
transcriptional regulation have been preserved between these
two species. Perhaps the molecular imprinting mechanism has
also been conserved between mouse and man.
The 1 OT 1 /2( 11 n) clone 5 was found to express both the
mouse igf-2 and H I9 but failed to express the human genes.
Therefore, the lack of expression of the human Igf-2 in this
microcell hybrid was not a result of the cells being
nonpermissive for expression of the human homolog. The
cause of the failure of the 10T 1/2(1 In) clone 5 to express the
human Igf-2 is not yet known. This hybrid may prove to be
quite useful and informative if the failure to express the
human gene results from modification of the imprinting
domain. A comparison of a putative imprinting region between
10Tl/2( 11 n) clone 5 and one of the 10T1/2 clones which
expressed the human Igf-2 and/or H I 9 may yield information
regarding the molecular nature of the imprint.
The analysis of the allele-specific expression of Igf-2 and
HJ9 is complicated in cells containing more than one copy of
the gene. Human-mouse somatic cell hybrids containing single
copies of the human Igf-2 (HI 9 locus provide an excellent
system in which to analyze the expression of Igf-2 and H I 9
from a single parental chromosome. This study reports on the
193
increased expression of the imprinted human HI9 upon
treatment with BUdR. Previously, it has been shown that a
mouse imprinted Igf-2 could be activated in somatic cell
cultures upon BUdR treatment (Chapter 3; Eversole-Cire et al.,
1993). Activation of both an imprinted Igf-2 as well as an
imprinted H I9 by the same chemical treatment may indicate
that the imprints for these two genes are similar. It is possible
that the imprints for Igf-2 and H19 are the same epigenetic
modification which when present on the maternal allele
represses expression of Igf-2 and when present on the
paternal allele represses expression of HI9.
The dose-dependent increase in expression of the human
HI9 in the microcell hybrids upon treatment with BUdR was
associated with a corresponding dose-dependent decrease in
expression of the human Igf-2. Although the decrease in
expression of the human Igf-2 is less pronounced than the
increase in human H I9 expression it does not necessarily
indicate that there is not a reciprocal relationship between the
expression levels of Igf-2 and HI 9. The less dramatic decrease
in Igf-2 expression may result from cells within the population
of BUdR-treated cells which did not respond to the drug. The
nonresponsive cells may be expressing Igf-2 at high levels
which may mask the decrease in Igf-2 expression in responsive
cells. Since the microcell hybrids only contain one copy of the
194
human igf-2/Hi9 locus it is likely that the induced expression
of H i 9 was interfering with the expression of igf-2. These
results support the notion that igf-2 and H19 compete for
transcriptional regulatory elements which allows expression of
only one of these genes normally being transcribed from a
single parental chromosome.
195
CHAPTER 6
Summary and Conclusions
Studying the molecular mechanism of genomic imprinting is
complicated by the presence of both an active and an inactive
allele within a single cell. Therefore, a comparative analysis of
the differential epigenetic modifications of an imprinted gene is
difficult. To circumvent this dilemma, a genetic approach was
utilized which used mice containing a duplication of the distal
region of the maternal chromosome 7, which harbors the
reciprocally imprinted genes, Igf-2 and H I 9, without the
paternal complement ([MatDi7]; Cattanach and Kirk, 1985;
Searle and Beechey, 1990). Fibroblast cell cultures were
derived from MatDi7 embryos and their normal littermates to
facilitate a study of the molecular mechanism of genomic
im printing.
Igf-2 expression was found to be growth-regulated in cells
derived from a normal embryo with mRNA levels increasing
substantially in the stationary phase of cell growth (Chapter 2).
Igf-2* however, was generally not expressed in cells derived
from a MatDi7 embryo, even after immortalization in culture,
indicating that the maternal imprint may have been
maintained in these cells. Therefore, the maternal Igf-2 allele
196
was kept silent by an as yet unidentified mechanism and the
paternal Igf-2 allele was either active or inactive, depending on
the growth status of the cell.
Methylation analysis of Igf-2 showed that the gene was
unmethylated in both normal and MatDi7 cells, even at high
passage, indicating that extensive de novo methylation had not
occurred during the process of cell culturing. Although Igf-2
was considerably unmethylated in the MatDi7 cell line, the
gene was found to be extensively methylated in an
oncogenically transformed cell line. Hence, the lack of
methylation of Igf-2 in the MatDi7 cells did not appear to
result from an inherent inability of the gene to become
methylated. However, extensive methylation of the CpG island
within this gene did not appear to be the mechanism used to
repress the maternal allele of Igf-2 in culture. Since the
maternal Igf-2 allele in the MatDi7 cells had not undergone any
apparent changes during the in vitro process of cell culturing
this newly derived cell line provided an excellent in vitro
system in which the molecular mechanism of imprinting could
be studied.
Mechanisms which control the expression of genes have
been studied for many years using somatic cell cultures.
Induced expression of the normally silent maternal Igf-2 with
197
an agent whose mechanism of action is known may provide
information as to the nature of the mechanism responsible for
repression of the maternal Igf-2. Therefore, treatment with
agents known to alter gene expression patterns in cell culture
were performed on MatDi7 fibroblast cell cultures in an
attempt to induce expression of the repressed maternally
derived Igf-2 (Chapter 3). Expression of the maternal Igf-2
was not induced when the cells were treated with
1 -B-D-arabinofuranosylcytosine, 2-deoxyuridine, sodium
butyrate, or calcium ionophore or when subjected to thermal
shock. Since none of these agents induced expression of the
maternal Igf-2 allele in the MatDi7 cells, it is unlikely that
mechanisms related to these perturbations were responsible
for the repression of the maternal Igf-2.
Expression of the maternal Igf-2 allele was, however,
induced in a dose-dependent manner by treatment of both
low- and high-passage MatDi7 cells with 5-aza-2'-deoxyuridine
or 5-bromodeoxyuridine, agents known to alter methylation
patterns and/or chromatin conformation. Expression was also
induced in subclones derived from single immortalized MatDi7
cells demonstrating that the increase in expression did not
result from a selection of pre-existing cells which expressed
Igf-2 at high levels within the mass culture. It was significant
that BUdR treatment also increased Igf-2 expression in MatDi7
198
cells immediately after explant indicating that results obtained
with immortalized cells might have relevance to in vivo
processes. Therefore, the mechanism of the Igf-2 imprint may
involve subtle changes in the methylation profile or chromatin
conformation of the gene since induction of the repressed allele
was only observed with drugs known to alter these two
epigenetic modifications.
The expression of the paternal Igf-2 was shown to be
growth-regulated and, therefore, reversal of the repressed
state of the maternal Igf-2 should result in the maternal allele
being subject to the same growth regulations as the paternal
allele. Coordinate derepression of Igf-2 promoters 2 and 3 was
observed in both chemically-induced and spontaneous
revertants. The acquired expression of Igf-2 was
growth-regulated indicating that the reversion event had
removed the imprint imposed on the maternal Igf-2. The
relationship between the spontaneous activation of the
maternal Igf-2 in vitro and reversible imprinting in the germ
line, as well as the lack of such imprints in some somatic
tissues, such as the leptomeninges and the choroid plexus, is
currently not known.
The reciprocal imprinting of Igf-2 and H I9 yet similar
spatio-temporal expression patterns during mouse embryonic
199
development has led to the hypothesis that Igf-2 and H I9 are
coordinately regulated. The molecular basis for the apparent
coordinate regulation of Igf-2 and H I9 from different parental
chromosomes was investigated in the normal and MatDi7
somatic cell cultures (Chapter 4). In cells derived from a
normal embryo, the two genes appeared to respond similarly to
the same regulatory factor(s) since similar decreases in the
overall expression levels of Igf-2 and H I9 were observed in
various subclones and both genes were upregulated when cells
became quiescent in response to confluence arrest or serum
deprivation. In a subclone of MatDi7 cells (MatDi7 1-la) which
spontaneously acquired the ability to express the normally
repressed maternally derived Igf-2, the Igf-2 allele
demonstrated the same growth-regulated expression pattern,
whereas the H I9 allele demonstrated a reciprocal expression
pattern to that observed for Igf-2.
Perhaps, the increased expression of the maternal Igf-2
allele(s) in the MatDi7 1-la subclone interfered with the
expression of the maternal HI 9 allele by competing for a
shared regulatory element. Competition for a common element
may have lead to the reciprocal expression patterns observed
for Igf-2 and H19 in the revertant cells. However, since it is
not yet known whether one or both of the maternal Igf-2
alleles reverted, further clarification awaits the analysis of
200
expression of Igf-2 and HI9 in cells containing a single copy of
the Igf-2/H19 locus.
Methylation analysis of the maternal Igf-2 and H19 in the
MatDi7 1-la cells showed de novo methylation of sites
upstream of Igf-2 and also within the H 1 9 promoter, regions
previously shown to be methylated only on the paternally
derived genes. In effect, the maternal allele had adopted a
methylation pattern characteristic of the paternal allele. Since
the acquired changes in methylation were confined to regions
thought to have potential importance in the regulation of Igf-2
and H I 9, the changes may be directly associated with
reactivation of the maternal Igf-2.
Previous results have led to the hypothesis that the mouse
Igf-2 and H I 9 compete for transcriptional regulatory elements
and/or factor(s) and therefore cannot be transcribed
simultaneously from a single chromosome. The regulation of
the human Igf-2 and H I 9 was studied using human-mouse
somatic cell hybrids containing a single copy of the human
Ig f-2 /H I9 locus to analyze the expression of Igf-2 and H I 9
from a single parental chromosome (Chapter 5). Analysis of
the relative expression levels of the human Igf-2 and H 19
suggested that the human chromosome present in the somatic
cell hybrids was of paternal origin. If the imprint imposed on
201
(he human paternal H I 9 was maintained in the mouse genetic
background this would indicate that mouse cells contain all the
necessary information to maintain and propagate the H 1 9
human imprint. Perhaps the imprinting mechanism has been
conserved between these mammalian species.
Activation of a mouse imprinted lgf-2 in somatic cell
cultures upon BUdR treatment was demonstrated in Chapter 3.
Interestingly, treatment of the human-mouse microcell hybrids
with BUdR resulted in a dose-dependent increase in expression
of the human imprinted H19 (Chapter 5). Activation of both an
imprinted mouse Igf-2 as well as a human H19 by the same
chemical treatment suggests that the imprints imposed on
Igf-2 and H19 may be similar. Perhaps the genomic imprint
for Igf-2 and H I9 are the same epigenetic modification which
when present on the maternal chromosome suppresses Igf-2
expression and when present on the paternal chromosome
suppresses H J9 expression.
The BUdR-induced dose-dependent increase in H I 9
expression in the human-mouse somatic cell hybrids was
accompanied by a dose-dependent decrease in Igf-2
expression. The induced expression of the human H I9 may
interfere with expression of Igf-2 resulting in the observed
decrease. These results support the hypothesis that Igf-2 an d
202
H I9 compete for transcriptional regulatory element(s) which
allows only Igf-2 or H 19 to be transcribed from a single
parental chromosome at a given time.
The somatic cell cultures, both the MatDi7 cell lines and also
the human-mouse microcell hybrids, provided novel in vitro
model systems to study the nature of parental imprinting. The
use of in vitro systems allowed an examination of the dynamic
aspects of the regulation of imprinted genes as well as the
influence of epigenetic modifications on their expression.
Therefore, the in vitro cell cultures are a unique experimental
system which may complement in vivo studies on the
molecular mechanism of genomic imprinting.
203
LIST OF REFERENCES
Adams, S. O., S. P. Nissley, S. Handwerger, and M. M. Rechler.
1983. Developmental patterns of insulin-like growth factor-I
and -II synthesis and regulation in rat fibroblasts. Nature, 302,
1 5 0 -1 5 3 .
Antequera, F., J. Boyes, and A. Bird. 1990. High levels of de
novo methylation and altered chromatin structure at CpG
islands in cell lines. Cell, 62, 503-514.
Barlow, D. P., R. Stoger, B. G. Herrmann, K. Saito, and N.
Schweifer. 1991. The mouse insulin-like growth factor type-2
receptor is imprinted and closely linked to the Tme locus.
Nature, 349, 84-87.
Bartolomei, M. S., S. Zemel, and S. M. Tilghman. 1991. Parental
imprinting of the mouse H19 gene. Nature, 351, 153-155.
Bartolomei, M. S., A. L. Webber, M. E. Brunkow, and S. M.
Tilghman. 1993. Epigenetic mechanisms underlying the
imprinting of the mouse H I9 gene. Genes Dev., 7, 1663-1673.
Barton, S. C., M. A. H. Surani, and M. L. Norris. 1984. Role of
paternal and maternal genomes in mouse development.
Nature, 31 1. 374-376.
Baylin, S. B., M. Makos, J. Wu, R.-W. C. Yen, A. de Bustros,
P. Vertino, and B. D. Nelkin. 1991, Abnormal patterns of DNA
methylation in human neoplasia: Potential consequences for
tumor progression. Cancer Cells, 3, 383-390.
Becker, P. B., S. Ruppert, and G. Schiitz. 1987. Genomic
footprinting reveals cell type-specific DNA binding of
ubiquitous factors. Cell, 51, 435-443.
204
Benedict, W. F., and P. A. Jones. 1979. Mutagenic, clastogenic
and oncogenic effects of 1-B-D-arabinofuranosylcytosine.
Mutation Research, 65, 1-20.
Bird, A. P. 1986. CpG-rich islands and the function of DNA
methylation. Nature, 321, 209-213.
Bird, A. P. 1987. CpG islands as gene markers in the
vertebrate nucleus. TIG, 3, 342-347.
Bischoff, R., and H. Holtzer. 1970. Inhibition of myoblast fusion
after one round of DNA synthesis in 5-bromodeoxyuridine.
J. Cell Biol., 44, 134-150.
Brandeis, M.t T. Kafri, M. Ariel, J. R. Chaillet, J. McCarrey, A.
Razin, and H. Cedar. 1993. The ontogeny of allele-specific
methylation associated with imprinted genes in the mouse.
EMBO J., 12, 3669-3677.
Brannan, C. I., E. C. Dees, R. S. Ingram, and S. M. Tilghman.
1990. The product of the HI 9 gene may function as an RNA.
Mol. Cell. Biol., 10, 28-36.
Brunkow, M. E., and S. M. Tilghman. 1991. Ectopic expression
of the H19 gene in mice causes prenatal lethality. Genes and
Dev., 5. 1092-1101.
Cattanach, B. M., and C. V. Beechey. 1990. Autosomal and
X-chromosome imprinting. Development Suppl., 63-72.
Cattanach, B. M., and M. Kirk. 1985. Differential activity of
maternally and paternally derived chromosome regions in
mice. Nature, 315, 496-498.
Cattanach, B. M., J. A. Barr, E. P. Evans, M. Burtenshaw, C. V.
Beechey, S. E. Leff, C. I. Brannan, N. G. Copeland, N. A. Jenkins,
and J. Jones. 1992. A candidate mouse model for Prader-Willi
syndrome which shows an absence of Snrpn expression.
Nature Genetics, 2, 270-274.
205
Cedar, H. 1988. DNA methylation and gene activity. Cell, 53,
3-4.
Chomczynski, P., and N. Sacchi. 1987. Single-step method of
RNA isolation by acid guanidinium thiocyanate-phenol-
chloroform extraction. Anal. Biochem., 162, 156-159.
Church, G. M., and W. Gilbert. 1984. Genomic sequencing. Proc.
Natl. Acad. Sci. USA, 81, 1991-1995.
Corbin, V., and T. Maniatis. 1989. Role of transcriptional
interference in the Drosophila melanogaster Adh promoter
switch. Nature, 337, 279-282.
Craig, J. M., and W. A. Bickmore. 1994. The distribution of CpG
islands in mammalian chromosomes. Nature Genetics, 7,
3 7 6 -3 8 2 .
Creusot, F., G. Acs, and J. K. Christman. 1982. Inhibition of DNA
methyltransferase and induction of Friend erythroleukemia cell
differentiation by 5-azacytidine and 5-aza-2'-deoxycytidine.
J. Biol. Chem., 257, 2041-2048.
Davis, R. L., H. Weintraub, and A. B. Lassar. 1987. Expression
of a single transfected cDNA converts fibroblasts to myoblasts.
Cell, 51, 987-1000.
DeChiara, T. M., A. Efstratiadis, and E. J. Robertson. 1990. A
growth-deficiency phenotype in heterozygous mice carrying an
insulin-like growth factor II gene disrupted by targeting.
Nature, 345, 78-80.
DeChiara, T. M., E. J. Robertson, and A. Efstratiadis. 1991.
Parental imprinting of the mouse insulin-like growth factor II
gene. Cell, 64, 849-859.
Doerfler, W. 1983. DNA methylation and gene activity. Ann.
Rev. Biochem., 52, 93-124.
206
Drummond, I. A. S., A. S. Lee, E. Resendez, Jr., and R. A.
Sieinhardt. 1987. Depletion of intracellular calcium stores by
calcium ionophore A23187 induces the genes for
glucose-regulated proteins in hamster fibroblasts. J. Biol.
Chem., 262, 12801-12805.
Eversole-Cire, P., A. C. Ferguson-Smith, H. Sasaki, K. D. Brown,
B. M. Cattanach, F. A. Gonzales, M. A. Surani, and P. A. Jones.
1993. Activation of an imprinted Igf-2 gene in mouse somatic
cell cultures. Mol. Cell. Biol., 13, 4928-4938.
Feinberg, A. P., and B. Vogelstein. 1984. A technique for
radiolabeling DNA restriction endonuclease fragments to high
specific activity. Anal. Biochem., 137, 266-267.
Ferguson-Smith, A. C., B. M. Cattanach, S. C. Barton, C. V.
Beechey, and M. A. Surani. 1991. Embryological and molecular
investigations of parental imprinting on mouse chromosome 7.
Nature. 351, 667-670.
Ferguson-Smith, A. C., H. Sasaki, B. M. Cattanach, and
M. A. Surani. 1993. Parental-origin-specific epigenetic
modification of the mouse H I9 gene. Nature, 362, 751-755.
Foley, K. P., and J. D. Engel. 1992. Individual stage selector
element mutations lead to reciprocal changes in B- vs. e -globin
gene transcription: genetic confirmation of promoter
competition during globin gene switching. Genes and Dev., 6,
7 3 0 -7 4 4 .
Gardiner-Garden, M., and M. Frommer. 1987. CpG islands in
vertebrate genomes. J. Mol. Biol., 196, 261-282.
Giannoukakis, N., C. Deal, J. Paquette, C. G. Goodyer, and
C. Polychronakos. 1993. Parental genomic imprinting of the
human 1GF2 gene. Nature Genetics, 4, 98-101.
207
Glaser, T., D. Housman, W. H. Lewis, D. Gerhard, and C. Jones.
1989, A fine-structure deletion map of human chromosome
lip : analysis of J1 series hybrids. Som. Cell. Mol. Gen., 15,
4 7 7 -5 0 1 .
Glenn, C. C., K. A. Porter, M. T. C. Jong, R. D. Nicholls, and
D. J. Driscoll. 1993. Functional imprinting and epigenetic
modification of the human SNRPN gene. Hum. Mol. Genet., 2,
2 0 0 1 -2 0 0 5 .
Groudine, M., and K. F. Conkin. 1985. Chromatin structure and
de novo methylation of sperm DNA: Implications for activation
of the paternal genome. Science, 228, 1061-1068.
Hao, Y., T. Crenshaw, T. Moulton, E. Newcomb, and B. Tycko.
1993. Tumour-suppressor activity of H19 RNA. Nature, 365,
7 6 4 -7 6 7 .
Hayashizaki, Y., H. Shibata, S. Hirotsune, H. Sugino, Y. Okazaki,
N. Sasaki, K. Hirose, H. Imoto, H. Okuizumi, M. Muramatsu,
H. Komatsubara, T. Shiroishi, K. Moriwaki, M. Katsuki,
N. Hatano, H. Sasaki, T. Ueda, N. Mise, N. Takagi, C. Plass, and
V. M. Chapman. 1994. Identification of an imprinted U2af
binding protein related sequence on mouse chromosome 11
using the RLGS method. Nature Genetics, 6, 33-39.
Henry, I., C. Bonaiti-Pellil, V. Chehensse, C. Beldjord,
C. Schwartz, G. Utermann, and C. Junien. 1991. Uniparental
paternal disomy in a genetic cancer-predisposing syndrome.
Nature, 351, 665-667.
Hildebrandt, M., and W. Nellen. 1992. Differential antisense
transcription from the Dictyostelium EB4 gene locus:
implications on antisense-mediated regulation of mRNA
stability. Cell, 69, 197-204.
Holliday, R. 1989. A different kind of inheritance. Sci. Amer .
June, 60-73.
208
Horikoshi, T., K. D. Danenberg, T. H. W. Stadlbauer,
M. Volkenandt, L. C. C. Shea, K. Aigner, B. Gustavsson.
L. Leichman, R. Frbsing, M. Ray, N. W. Gibson, C. P. Spears, and
P. V. Danenberg. 1992. Quantitation of thymidylate synthase,
dihydrofolate reductase, and DT-diaphorase gene expression in
human tumors using the polymerase chain reaction. Cancer
Res., 52, 108-116.
Howlett, S. K., and W. Reik. 1991. Methylation levels of
maternal and paternal genomes during preim plantation
development. Development, 113, 119-127.
Ikejiri, K., M. Furuichi, T. Ueno, T. Matsuguchi, K. Takahashi,
H. Endo, and M. Yamamoto. 1991. The presence and active
transcription of three independent leader exons in the mouse
insulin-like growth factor II gene. Biochim. Biophys. Acta.,
1089, 77-82.
Jones, P. A. 1985. Effects of 5-azacytidine and its
2-deoxyderivative on cell differentiation and DNA methylation.
Pharmac. Ther., 28, 17-27.
Jones, P. A., and J. D. Buckley. 1990. The role of DNA
methylation in cancer. Adv. Cancer Res., 54, 1-23.
Jones, P. A., M. J. Wolkowicz, W. M. Rideout III, F. A. Gonzales,
C. M. Marziasz, G. A. Coetzee, and S. J. Tapscott. 1990. De novo
methylation of the MyoD I CpG island during the establishment
of immortal cell lines. Proc. Natl, Acad. Sci. USA, 87,
6 1 1 7 -6 1 2 1 .
Kafri, T., M. Ariel, M. Brandeis, R. Shemer, L. Urven, J. McCarrey,
H. Cedar, and A. Razin. 1992. Developmental pattern of
gene-specific DNA methylation in the mouse embryo and germ
line. Genes and Dev., 6, 705-714.
Kafri, T., X. Gao, and A. Razin. 1993. Mechanistic aspects of
genome-wide dem ethylation in the preim plantation mouse
embryo. Proc. Natl. Acad. Sci. USA, 90, 10558-10562.
209
Kalscheuer, V. M., E. C. Mariman, M. T. Schepens, H. Rehder, and
H. -H. Ropers. 1993. The insulin-like growth factor type-2
receptor gene is imprinted in the mouse but not in humans.
Nature Genetics, 5, 74-78.
Kass, S. U.. J. P. Goddard, and R. L. P. Adams. 1993. Inactive
chromatin spreads form a focus of methylation. Mol. Cel). Biol..
13, 7372-7379.
Keith, D. H , J. Singer-Sam, and A. D. Riggs. 1986. Active X
chromosome DNA is unmethylated at eight CCGG sites clustered
in a guanine-plus-cytosine-rich island at the 5' end of the gene
for phosphoglycerate kinase. Mol. Cell. Biol., 6, 4122-4125.
Keshet, I., J. Lieman-Hurwitz, and H. Cedar. 1986. DNA
methylation affects the formation of active chromatin. Cell, 44,
5 3 5 -5 4 3 .
Kim, K. S., and A. S. Lee. 1986. The effect of extracellular Ca2 +
and temperature on the induction of the heat-shock and
glucose-regulated proteins in hamster fibroblasts. Biochem.
Biophys. Res. Commun., 140, 881-887.
Koufos, A., P. Grundy, K. Morgan, K. A. Aleck, T. Hadro,
B. C. Lampkin, A. Kalbakji, and W. K. Cavenee. 1989. Familial
W iedemann-Beckwith syndrome and a second Wilms tumor
locus both map to 11 p i 5.5. Am. J. Hum. Genet., 44, 711-719.
Kruh, J. 1982. Effects of sodium butyrate, a new
pharmacological agent, on cells in culture. Mol. Cell. Biochem.,
42, 65-82.
Leach, R. J., M. J. Thayer, A. J. Schafer, and R. E. K. Fournier.
1989. Physical mapping of human chromosome 17 using
fragment-containing microcell hybrids. Genomics, 5, 167-176.
210
Lee, J. E., J. Pintar, and A. Efstratiadis. 1990. Pattern of the
insulin-like growth factor II gene expression during early
mouse embryogenesis. Development, 110, 151-159.
Leff, S. E., C. I. Brannan, M. L. Reed, T. Ozcelik, U. Francke,
N. G. Copeland, and N. A. Jenkins. 1992. Maternal imprinting
of the mouse Snrpn gene and conserved linkage homology with
the human Prader-Willi syndrome region. Nature Genetics, 2,
2 5 9 -2 6 4 .
Li, E., C. Beard, and R. Jaenisch. 1993. Role for DNA
methylation in genomic imprinting. Nature, 366, 362-365.
Lugo, T. G., B. Handelin, A. M. Killary, D. E. Housman, and
R. E. K. Fournier. 1987. Isolation of microcell hybrid clones
containing retroviral vector insertions into specific human
chromosomes. Mol. Cell. Biol., 7, 2814-2820.
Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular
cloning: a laboratory manual. Cold Spring Harbor Laboratory,
New York.
Mannens, M., R. M. Slater, C. Heyting, J. Bliek, J. de Kraker,
N. Coad, P. de Pagter-Holthuizen, and P. L. Pearson. 1988.
Molecular nature of genetic changes resulting in loss of
heterozygosity of chromosome 11 in Wilms' tumours. Hum.
Genet., 81, 41-48.
Matsumoto, K., C. Gaetano, W. H. Daughaday, and C. J. Thiele.
1992. Retinoic acid regulates insulin-like growth factor II
expression in a neuroblastoma cell line. Endocrinology, 130,
3 6 6 9 -3 6 7 6 .
McAllister, R. M., J, Melnyk, J. Z. Finklestein, E. C. Adams, Jr., and
M. B. Gardner. 1969. Cultivation in vitro of cells derived from
a human rhabdomyosarcoma. Cancer, 24, 520-526.
211
McGrath, J., and D. Solter. 1984. Completion of mouse
embryogenesis requires both the maternal and paternal
genomes. Cell, 37, 179-183.
Meinsma, D., W. Scheper, P. E. Holthuizen, J. L. Van den Brande,
and J. S. Sussenbach. 1992. Site-specific cleavage of IGF-I1
mRNAs requires sequence elements from two distinct regions
of the IGF-II gene. Nucleic Acids Res., 20, 5003-5009.
Miura, Y., and F. H. Wilt. 1971. The effects of
5-bromodeoxyuridine on yolk sac erythropoiesis in the chick
embryo. J. Cell Biol., 48, 523-532.
Monk, M., and M. Grant. 1990. Preferential X-chromosome
inactivation, DNA methylation and imprinting. Development
Supplement, 55-62.
Monk, M., M. Boubelik, and S. Lehnert. 1987. Temporal and
regional changes in DNA methylation in the embryonic,
extraembryonic and germ cell lineages during mouse embryo
development. Development, 99, 371-382.
Moulton, T., T. Crenshaw, Y. Hao, J. Moosikasuwan, N. Lin,
F. Dembitzer, T. Hensle, L. Weiss, L. McMorrow, T. Loew,
W. Kraus, W. Gerald, and B. Tycko. 1994. Epigenetic lesions at
the H19 locus in Wilms' tumour patients. Nature Genetics, 7,
4 4 0 -4 4 7 .
Mtlller-Taubenberger, A., J. Hagmann, A. Noegel, and G. Gerisch.
1988. Ubiquitin gene expression in Dictyostelium is induced by
heat and cold shock, cadmium, and inhibitors of protein
synthesis. J. Cell Sci., 90, 51-58.
Ogawa, O., M. R. Eccles, J. Szeto, L. A. McNoe, K. Yun, M. A. Maw,
P. J. Smith, and A. E. Reeve. 1993. Relaxation of insulin-like
growth factor II gene imprinting implicated in Wilms tumor.
Nature, 362, 749-751.
212
Ohlsson, R„ A. Nystriim, S. Pfeifer-Ohlsson, V, Tbhdnen,
F. Hedborg, P. Schofield, F. Flam, and T. J. Eckstrom. 1993.
IGF2 is parentally imprinted during human embryogenesis and
in the Beckwith-Wiedemann syndrome. Nature Genetics, 4,
94-97.
Pachnis, V., C. I. Brannan, and S. M. Tilghman. 1988. The
structure and expression of a novel gene activated in early
mouse embryogenesis. EMBO J., 7, 673-681.
Pfeifer, G. P., S. D. Steigerwald, R. S. Hansen, S. M. Gartler, and
A. D. Riggs. 1990. Polymerase chain reaction-aided genomic
sequencing of an X chromosome-linked CpG island: Methylation
patterns suggest clonal inheritance, CpG site autonomy, and an
explanation of activity state stability. Proc. Natl. Acad. Sci. USA,
87, 8252-8256.
Ping, A. J., A. E. Reeve, D. J. Law, M. R. Young, M. Boehnke, and
A. P. Feinberg 1989. Genetic linkage of Beckwith-Wiedemann
syndrome to llp l5 . Am. J. Hum. Genet., 44, 720-723.
Poirier, F., C.-T. J. Chan, P. M. Timmons, E. J. Robertson,
M. J. Evans, and P. W. J. Rigby. 1991. The murine HI9 gene
is activated during embryonic stem cell differentiation in vitro
and at the time of implantation in the developing embryo.
Development, 113, 1105-1114.
Proudfoot, N. J. 1986. Transcriptional interference and
termination between duplicated a-globin gene constructs
suggests a novel mechanism for gene regulation. Nature, 322,
5 6 2 -5 6 5 .
Rainier, S., L. A. Johnson, C. J. Dobry, A. J. Ping, P. E. Grundy, and
A. P. Feinberg. 1993. Relaxation of imprinted genes in human
cancer. Nature, 362, 747-749.
213
Rappolee, D. A., K. S. Sturm, O. Behrendtsen, G. A. Schultz,
R. A. Pedersen, and Z. Werb. 1992, Insulin-like growth factor
II acts through an endogenous growth pathway regulated by
imprinting in early mouse embryos. Genes Dev., 6, 939-952.
Reed, M. L., and S. E. Leff. 1994. Maternal imprinting of
human SNRPN, a gene deleted in Prader-Willi syndrome.
Nature Genetics, 6, 163-167.
Reik, W„ A. Collick, M. L. Norris, S. C. Barton, and M. A. Surani.
1987. Genomic imprinting determines methylation of parental
alleles in transgenic mice. Nature, 328, 248-251.
Reznikoff, C. A., D. W. Brankow, and C. Heidelberger. 1973A.
Establishment and characterization of a cloned line of C3H
mouse embryo cells sensitive to postconfluence inhibition of
division. Cancer Res., 33, 3231-3238.
Reznikoff, C. A., J. S. Bertram, D. W. Brankow, and
C. Heidelberger. 1973B. Quantitative and qualitative studies
of chemical transformation of cloned C3H mouse embryo cells
sensitive to postconfluence inhibition of cell division. Cancer
Res., 33, 3239-3249.
Riggs, A. D., and P. A. Jones. 1983. 5-Methylcytosine, gene
regulation, and cancer. Advances in Cancer Research, 40, 1-30.
Rivkin, M., K. M. Rosen, and L. Villa-Komaroff. 1993.
Identification of an antisense transcript from the IGF-II locus
in mouse. Mol. Rep. Dev., 35, 394-397.
Rotwein, P., and L. J. Hall. 1990. Evolution of insulin-like
growth factor II: Characterization of the mouse IGF-II gene
and identification of two pseudo-exons. DNA and Cell Biol., 9,
7 2 5 -7 3 5 .
214
Rutter, W. J., R. L. Pictet, P. W. Morris. 1973. Toward molecular
mechanisms of developmental processes. Ann. Rev. Biochem.,
42. 601-646.
Sanford, J. P., H. J. Clark, V. M. Chapman, and J. Rossant. 1987.
Differences in DNA methylation during oogenesis and
spermatogenesis and their persistence during early
embryogenesis in the mouse. Genes Dev., 1, 1039-1046.
Sapienza, C., A. C. Peterson, J. Rossant, and R. Balling. 1987.
Degree of methylation of transgenes is dependent on gamete of
origin. Nature, 328, 251-254.
Sasaki, H., P. A. Jones, J. R. Chaillet, A. C. Ferguson-Smith,
S. C. Barton, W. Reik, and M. A. Surani. 1992. Parental
imprinting: potentially active chromatin of the repressed
maternal allele of the mouse insulin-like growth factor II (/#/2)
gene. Genes Dev., 6, 1843-1856.
Schneider, C., R. M. King, and L. Philipson. 1988. Genes
specifically expressed at growth arrest of mammalian cells.
Cell. 54, 787-793.
Schubert, D., and F. Jacob. 1970. 5-Bromodeoxyuridine-
induced differentiation of a neuroblastoma. Proc. Natl. Acad.
Sci. USA, 67, 247-254.
Searle, A. G., and C. V. Beechey. 1990. Genome imprinting
phenomena on mouse chromosome 7. Genet. Res., 56, 237-244.
Steenman, M. J. C., S. Rainier, C. J. Dobry, P. Grundy, I. L. Horon,
and A. P. Feinberg. 1994. Loss of imprinting of IGF2 is linked
to reduced expression and abnormal methylation of H19 in
Wilms’ tumour. Nature Genetics, 7, 433-439.
215
Stellwagen, R. H., and G. M. Tomkins. 1971 A. Differential effect
of 5-bromodeoxyuridine on the concentrations of specific
enzymes in hepatoma cells in culture. Proc. Natl. Acad. Sci.
USA, 68. 1147-1150.
Stellwagen, R. H., and G. M. Tomkins. 197IB, Preferential
inhibition by 5-bromodeoxyuridine of the synthesis of tyrosine
aminotransferase in hepatoma cell cultures. J. Mol. Biol., 56,
16 7 -1 8 2 .
Stempien, M. M., N. M. Fong, L. B. Rail, and G. I. Bell. 1986.
Sequence of a placental cDNA encoding the mouse insulin-like
growth factor II precursor. DNA, 5, 357-361.
Stfjger, R., P. Kubicka, C. -G. Liu, T. Kafri, A. Razin, H. Cedar, and
D. P. Barlow. 1993. Maternal-specific methylation of the
imprinted mouse Igf2r locus identifies the expressed locus as
carrying the imprinting signal. Cell, 73, 61-71.
Stryer, L., 1988. Biochemistry, W. H. Freeman and Company/
New York, 823-827.
Surani. M. A. H., S. C. Barton, and M. L. Norris. 1984.
Development of reconstituted mouse eggs suggests imprinting
of the genome during gametogenesis. Nature, 308, 548-550.
Surani, M. A., N, D. Allen, S. C. Barton, R. Fundele, S. K. Howlett,
M. L. Norris, and W. Reik. 1990. Developmental consequences
of imprinting of parental chromosomes by DNA methylation.
Phil. Trans. R. Soc. Lond. B, 326, 313-327.
Surani, M. A. 1993. Silence of the genes. Nature, 366,
3 0 2 -3 0 3 .
Surani, M. A., H. Sasaki, A. C. Ferguson-Smith, N. D. Allen,
S. C. Barton, P. A. Jones, and W. Reik. 1993. The inheritance
of germline-specific epigenetic modifications during
development. Phil. Trans. R. Soc. Lond. B., 339, 165-172.
216
Swain, J. L., T. A. Stewart, and P. Leder. 1987. Parental legacy
determines methylation and expression of an autosomal
transgene: a molecular mechanism for parental imprinting.
Cell, 50, 719-727.
Takagi, N., and M. Sasaki. 1975. Preferential inactivation of
the paternally derived X chromosome in the extraembryonic
membranes of the mouse. Nature, 256, 640-642.
Tapscott, S. J., A. B. Lassar, R. L. Davis, and H. Weintraub. 1989.
5-Bromo-2'-deoxyuridine blocks myogenesis by extinguishing
expression of MyoDl. Science, 245, 532-536.
Tapscott, S. J., M. J. Thayer, and H. Weintraub. 1993. Deficiency
in rhabdomyosarcomas of a factor required for MyoD activity
and myogenesis. Science, 259, 1450-1453.
Taylor, E. R., E. A. P. Seleiro, and P. M. Brickell. 1991.
Identification of antisense transcripts of the chicken
insulin-like growth factor-II gene. J. Mol. Endocrinology, 7,
145-154.
Taylor, S. M., and P. A. Jones. 1982. Mechanism of action of
eukaryotic DNA methyltransferase. J. Mol. Biol., 162, 679-692.
Thayer, M. J., and H. Weintraub. 1990. Activation and
repression of myogenesis in somatic cell hybrids: Evidence for
frcmj-negative regulation of MyoD in primary fibroblasts. Cell,
63, 23-32.
Tollefsen, S. E., J. L. Sadow, and P. Rotwein. 1989. Coordinate
expression of insulin-like growth factor II and its receptor
during muscle differentiation. Proc. Natl. Acad. Sci. USA, 86,
1 543-1547.
217
Toniolo, D., G. Martini, B. R. Migeon, and R. Dono. 1988.
Expression of the G6PD locus on the human X chromosome is
associated with demethylation of three CpG islands within 100
kb of DNA. EMBO J., 7, 401-406.
Turleau, C., J. de Grouchy, F. Chavin-Colin, H. Martelli, M. Voyer,
and R. Charlas. 1984. Trisomy 11 p i 5 and
Beckwith-Wiedemann syndrome. A report of two cases. Hum.
Genet., 67. 219-221.
Tykocinski, M. L., and E. E. Max. 1984. CG dinucleotide clusters
in MHC genes and in 5' demethylated genes. Nuc. Acids Res.,
12, 4385-4396.
Ueda, T., K. Yamazaki, R. Suzuki, H. Fujimoto, H. Sasaki,
Y. Sakaki, and T. Higashinakagawa. 1992. Parental
methylation patterns of a transgenic locus in adult somatic
tissues are imprinted during gametogenesis. Development,
116, 831-839.
Waziri, M., S. R. Patil, J. W. Hanson, and J. A. Bartley. 1983.
Abnormality of chromosome 11 in patients with features of
Beckwith-Wiedemann syndrome. Pediatrics J., 102, 873-876.
Weintraub, H. 1985. Assembly and propagation of repressed
and derepressed chromosomal states. Cell, 42, 705-711.
Weintraub, H., G. L. Campbell, and H. Holtzer. 1972.
Identification of a developmental program using
bromodeoxyuridine. J. Mol. Biol., 70, 337-350.
Wolf, S. F., and B. R. Migeon. 1985. Clusters of CpG
dinucleotides implicated by nuclease hypersensitivity as
control elements of housekeeping genes. Nature, 314, 467-469.
218
Yen, P. H., P. Patel, A. C. Chinault, T. Mohandas, and L. J. Shapiro.
1984. Differential methylation of hypoxanthine
phosphoribosyltransferase genes on active and inactive human
X chromosomes. Proc. Natl. Acad. Sci. USA, 81, 1759-1763.
Yoo-Warren, H., V. Pachnis, R. S. Ingram, and S. M, Tilghman.
1988. Two regulatory domains flank the mouse H19 gene.
Mol. Cell. Biol., 8, 4707-4715.
Zemel, S., M. S. Bartolomei, and S. M. Tilghman. 1992. Physical
linkage of two mammalian imprinted genes, HI 9 and
insulin-like growth factor 2. Nature Genetics, 2, 61-65.
Zhang, Y., T. Shields, T. Crenshaw, Y. Hao, T. Moulton, and
B. Tycko. 1993. Imprinting of human H19: allele-specific CpG
methylation, loss of the active allele in Wilms tumor, and
potential for somatic allele switching. Am. J. Hum. Genet., 53,
113-124.
219
APPENDIX
A b b rev iatio n s
5-aza-CdR...................................................... 5 -aza-2 '-d eo x y cy tid in e
5-aza-C R ......................................................5 -aza-cy tid in e
b p ...................................................................base pair(s)
BUdR..............................................................5 -b ro m o d eo x y u rid in e
cDNA..............................................................complementary DNA
co n f................................................................confluent
C......................................................................cytosine
DMEM............................................................. Dulbecco’s Modified Eagle's
Medium
DNA................................................................deoxyribonucleic acid
dNTP...............................................................deoxy nucleoside
triphosphate
DTT.................................................................dithiothreitol
EDTA..............................................................eth y len ed i am ine
tetraacetic acid
G...................................................................... guanosine
G 418 ..............................................................Geneticin
GAPDH..........................................................g ly c erald e h y d e-3 -
phosphate dehydrogenase
h ...................................................................... hour(s)
HIFCS.............................................................heat-inactivated fetal calf
serum
I g f - 2 ..............................................................insulin-like growth factor 2
k b ................................................................... kilobase pair(s)
K Q .................................................................. potassium chloride
log....................................................................logarithm ic
M......................................................................m olar
M atD i7.......................................................... maternally disomic,
paternally deficient for
distal chromosome 7
M EM .............................................................. Eagle’s Minimum Essential
Medium
MgCl2............................................................ magnesium chloride
min...................................................................m inute(s)
m l ....................................................................m illiliter
220
mM..................................................................m illim olar
mm.................................................................. m illim eter
MMLV RT ............................................Moloney murine leukemia
virus reverse transcriptase
ng.....................................................................nanogram
°C................................ .....................................degree centigrade
P 2 ...................................................................lgf-2 promoter 2
P 3 ...................................................................lgf-2 promoter 3
PBS.................................................................phosphate-buffered saline
RNA................................................................ribonucleic acid
RNase A..........................................................ribonuclease A
RNasin............................................................. ribonuclease inhibitor
RT-PCR.......................................................... reverse transcription-
polymerase chain reaction
SDS.................................................................sodium dodecyl sulfate
s e c * ........................................................ second(s)
TBE................................................................. 0.089 M TRIS-borate,
0.003 M EDTA
TCA............................................. ....................trichloroacetic acid
TE.................................................................... 10 mM TRIS-HC1,
1 mM EDTA (pH 8.0)
U...................................................................... unit(s)
ug.....................................................................m icrogram
ul...................................................................... m icroliter
u M .................................................................. m icrom olar
u m .................................................................. m icro m eter
U V .................................................................. ultraviolet
221

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Dna Methylation Patterns In Human Tumor Suppressor Genes

PDF

An in vivo study of G protein coupled receptor mediated signaling

PDF

A coactivator complex among GRIP1, CARM1, and TIF1alpha contributes to gene activation directed by androgen receptor

PDF

A model for the mechanism of agonism and antagonism in steroid receptors

PDF

An E2F-independent mechanism in transcriptional control of the hamster histone H3.2 gene expression

PDF

Studies Of Topoisomerase Ii As The Target Of Anti-Cancer Agents

PDF

Transcriptional induction of the 78 kiloDalton glucose regulated protein (grp78) gene through depletion of intracellular calcium ion from endoplasmic reticulum

PDF

Different roles of p160 coregulators in myogenesis

PDF

Analysis of the function and regulation of the 94kDa glycoprotein of the glucose-regulated protein family.

PDF

Characterization of target cell entry by murine leukemia viruses

PDF

Investigation of the role of epigenetic modification of DNA and chromatin in aberrant gene silencing in cancer cells

PDF

Anomalies at NGX6 locus: Potential involvement in feline lymphomas

PDF

Cyclic amp binding and protein kinase activities in hepatoma culture cells

PDF

Alternative splicing of amyloid precursor protein RNA in response to in vitro and in vivo neuronal insults, and in cell lines transfected with a mini-gene construct

PDF

Molecular And Regulatory Properties Of Multiple Cyclic Nucleotide Phosphodiesterases

PDF

Development and secretions of salivary glands using mouse models

PDF

5-aza-2-deoxycytidine resistance and gene structure in mouse cells stripped of DNA methylation

PDF

Characterizing the function of murine epididymal secretory protein 1 (ME1) in hematopoietic stem cells

PDF

Coactivator synergy for nuclear receptors: Regulatory roles of arginine methyltransferase activity in transcription

PDF

Biochemical factors determining tumor response to 5,10-dideazatetrahydrofolate, a folate antimetabolite inhibitory to de novo purine biosynthesis

Asset Metadata

Creator Eversole-Cire, Pamela Louise (author)

Core Title Molecular mechanism of genomic imprinting in somatic cell cultures

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Biochemistry and Molecular Biology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag biology, cell,biology, genetics,biology, molecular,OAI-PMH Harvest

Language English

Advisor Jones, Paul A. (committee chair), Stallcup, Michael R. (committee member), Stellwagen, Robert H. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c20-570052

Unique identifier UC11226396

Identifier 9600972.pdf (filename),usctheses-c20-570052 (legacy record id)

Legacy Identifier 9600972.pdf

Dmrecord 570052

Document Type Dissertation

Rights Eversole-Cire, Pamela Louise

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA