UNIVERSITY OF CINCINNATI
Date:______
I, ______, hereby submit this work as part of the requirements for the degree of: in:
It is entitled:
This work and its defense approved by:
Chair: ______
The Visualization, Quantification and Modeling of
Genomic Instability in the Mouse and in Cultured Cells
A dissertation submitted to the graduate school of the
University of Cincinnati
In partial fulfillment of the
requirements for the degree of
Doctorate of Philosophy (PhD)
In the department of Molecular Genetics, Biochemistry and
Microbiology of the College of Medicine
2006
by
Jon Scott Larson
B.S., University of Toledo, 1999
Committee: Dr. James Stringer (chair) Dr. Tom Doetschman Dr. Joanna Groden Dr. Carolyn Price Dr. Peter Stambrook Abstract
Multicellular organisms are mosaic in nature because of genetic
alterations that occur in somatic cells. There are many factors that can contribute
to the formation of such alterations including aberrant DNA repair, environmental
insults, epigenetic modification, errors in DNA replication and errors in
chromosome duplication/segregation. To further the study of the distributions,
frequencies and rates at which some alterations can occur, mouse reporter
models were implemented.
The Tg(βA-G11PLAP) transgenic mutation reporter mouse harbors an
allele (G11 PLAP ) that is rendered incapable of producing its functional enzyme
because of a reading frame shift caused by an insertion of 11 G:C basepairs.
Spontaneous deletion of one G:C basepair from this mononucleotide repeat
restores gene function, and cells with PLAP activity can be detected
histochemically. G11 PLAP mice enable mutant cells to be visualized in situ and
were used to study variation during early development, in the germline, under
oxidative stress and in solid tumors.
To study LOH in diverse cell types in the body another reporter model was implemented. Mice that carry two different fluorescent protein genes as alleles of a locus were generated to address this issue because LOH would change a cell’s phenotype from bichrome to monochrome. As a step in assessing the utility of this approach, we derived MEF and ES cell lines from mice that carried two different fluorescent protein genes as alleles at the chromosome 6 locus,
ROSA26. FACS showed that the vast majority of cells in each line expressed the
1 two marker proteins at similar levels, but populations exhibited extrinsic and intrinsic noise with respect to expression. In addition, cells with a monochrome phenotype were frequent (10-4). In ES cells, all monochrome events were
accompanied by allele loss. Mitotic recombination appeared to be the major
cause, although UPD also appeared to have contributed to LOH. These cells
provided a novel assay for studying genetic/karyotypic stability of cultured ES
cells. Results obtained from studies with these cells support the need for caution
regarding the use of cultured stem cells in therapy.
2 Acknowledgements
There have been a few people who have impacted my life in such a way
that it forever changed who I am and helped to shape me into the person that I
am today. I owe credit to these invaluable influences for the encouragement,
direction and determination given to me for everything leading up to the
preparation of this document: dissertation for the Ph.D. degree granted by the
Department of Molecular Genetics, Biochemistry and Microbiology at the
University of Cincinnati, College of Medicine.
Thank you to my Ph.D. advisor/ mentor, Dr. Jim Stringer. I sometimes
wonder how I was so fortunate to get a project in his lab. He has provided more guidance and support towards me than I thought I would need, giving me the much need reality checks that kept me on track. I have learned much from him. I would also like to thank the Stringer family; Saudra, Hillary and Alice. They have been like family to me, forgiving my occasional blunders (sorry Mr. Wiggles et. al.). They will always be in my heart and mind.
Jim and the past/present members of the lab (Saundra Stringer Ph.D.,
Scott Keely Ph.D., Megan Hersh Ph.D., Jared Fischer and Carolyn Tindal) have
all shared both the joys and frustrations of my successes, as well as the utter
failures. I thank you for many hours of intellectually stimulating, entertaining and
sometimes absolutely absurd conversions, and interlaboratory activities. I will
sincerely miss working with all of you.
I would also like to thank my graduate committee members, Tom
Doetschman Ph.D., Joanna Groden Ph.D., Carolyn Price Ph.D. and Peter
3 Stambrook Ph.D. I couldn’t imagine having a more enthusiastic and supportive
group of mentors, all renowned in their areas of specialties. They were
instrumental in training me, always pushing me to dive deeper, think clearly, and
work harder.
“Thank you” to the members of the department who have extended to me
both their assistance and moral support, throughout my tenure as a graduate
student. To Rachel Sellmeyer, Dorie Lane, Peggy Casselman , Vicki Morris,
Felicia Romaine, Moying Yin, Tina Grisham, Issac Houston, Brock Schwitzer,
Justin Huddleson, Kelly Flory, Elizabeth Loreux and Jorge Muniz, my most sincere gratitude. You all have become dear friends whom I will miss.
Lastly, and most importantly, I would like to thank the love of my life Iva
Dostanic (Larson). We met and shared many great times while in ‘Molgen’. She
has been my muse and support through the most challenging, emotional and
exciting times. Thanks Bebe!
4
5 Table of Contents
List of Abbreviations p8
Chapter 1 – Introduction
Abstract p12
Background p13
Chapter 2 – Modeling Variation in Tumors in vivo
Abstract p26
Introduction p27
Results p30
Discussion p42
Materials and Methods p45
Chapter 3 - Increased Mutation in Mice Genetically Predisposed to
Oxidative Damage in the Brain
Abstract p50
Introduction p51
Results p53
Discussion p59
Materials and Methods p60
Chapter 4 - Impact of Mismatch Repair Deficiency on Genomic
Stability in the Maternal Germline and during Early Embryonic
6 Development
Abstract p63
Introduction p64
Results p68
Discussion p74
Materials and Methods p80
Chapter 5 - Expression and Loss of Alleles in Cultured Mouse
Embryonic Fibroblasts and Stem Cells Carrying Allelic
Fluorescent Protein Genes
Abstract p85
Introduction p86
Results and Discussion p89
Conclusion p103
Materials and Methods p108
Chapter 6 – Dissertation Summary p119
References p123
7 Abbreviations
Aif apoptosis inducing factor
APAF-1 apoptotic protease activating factor 1
APC adenomatous polyposis coli
APE1 Apurinic endonuclease 1
APRT adenine phosphoribosyltransferase
BAX BCL2 associated X
BCIP 5-bromo-4-chloro-3-indolphosphate
BCL2 B-cell lymphoma 2
Bl6 black six mouse strain
Bp basepair(s)
CFP cyan fluorescent protein
Chk1 checkpoint kinase1 c-myb Cellular DNA binding proteins encoded by the myb
gene
DMEM Dulbecco’s modified eagle media
DNA deoxyribonucleic acid
DNA-PKcs Protein Kinase, DNA activated, catalytic subunit dPBS Dulbecco’s phosphate buffered saline
DSB double strand break
E. coli Escherichia coli
E2F-4 transcription factor that control expression of a variety
of genes involved in cell cycle regulation
8 EMS Ethanomethylsulfate
ES cell embryonic stem cell
FACS fluorescent activated cell sorting
FISH fluorescent in situ hybridization
Flash FLice-ASsociated Huge protein
FVB/N friend virus B/NIH mouse strain
G11 PLAP reporter transgene that contains a mononucleotide
run of eleven GC base pairs in the human PLAP
gene hMSH3 human MutS homolog 3 hMSH6 human MutS homolog 6
HNPCC Hereditary nonpolyposis colorectal cancer
Hq Harlequin
ICE Family of aspartate-specific cysteine proteases,
caspases
IGF2R insulin-like growth factor II receptor
IR ionizing radiation
LIF Leukemia inhibitory factor
LOH loss of heterozygosity
MAPK mitogen activated protein kinase
MEF mouse embryonic fibroblast
MgCl2 magnesium chloride
MLSN1 Melastatin 1
9 MMR mismatch repair
MMTV mouse mammary tumor virus
MR mitotic recombination
MSI microsatellite instability
Myc family of retrovirus-associated DNA sequences (myc)
originally isolated from an avian myelocytomatosis
virus
NADH reduced nicotinamide adenine dinucleotide (NAD+)
Neu neu Proto-Oncogene Protein, A cell surface protein-
tyrosine kinase receptor that is found to be
overexpressed in a significant number of
adenocarcinomas
PBS phosphate buffered saline
PCR polymerase chain reaction
Pdcd8 programmed cell death 8
PLAP placental alkaline Phosphatase
PMS2 post meiotic segregation 2
PyMT polyoma virus middle T-antigen
RAD50 human homologue of a yeast gene, required for
spontaneous and induced mitotic recombination,
meiotic recombination and mating-type switching.
RAS family of viral oncogenes
Sky spectral karyotyping
10 SV40 simian virus 40
Taq Thermos aquaticus
TCF family of DNA-binding proteins that are primarily
expressed in T-LYMPHOCYTES
Tg(βA-G11PLAP) transgenic mouse with a G11 PLAP allele driven by a
human beta actin promoter
TGFbRII transforming growth factor beta receptor II
UPD uniparental disomy
XPG xeroderma pigmentosum
YFP yellow fluorescent protein
11 Chapter 1
Introduction
Abstract
Multicellular organisms are mosaic in nature because of genetic alterations that occur in somatic cells. There are many factors that can contribute to the formation of such alterations including aberrant DNA repair, aberrant cell cycle, environmental insults (e.g. oxidative stress, radiation), epigenetic modification (altering expression), errors in DNA replication and errors in chromosome duplication/segregation. To further the study of the distributions, frequencies and rates at which some alterations can occur, mouse reporter models were implemented. Transgenic mice (G11 PLAP) that allow mutant cells to be visualized in situ were used to study variation during early development, in the germline, under oxidative stress and in solid tumors. Also a dual fluorescent reporter system was developed to further studies of LOH in the mouse and in cultured cells.
Background
Our bodies are comprised of billions of cells and are genetically mosaic.
This mosaicism is caused by normal differentiation and random mutations.
Random mutations occur at low frequencies in normal somatic cells and sometime manifest themselves in the form cancer. For the most part, cancer is an evolutionary process in somatic cells. Therefore cancers can be defined by the epigenetic and genetic alterations that underlie these processes [1-9].
Cancer cells often acquire a malignant phenotype due to the accumulation of
12 chromosomal rearrangements or mutations. Such events accumulate over time
and can result in an altered or loss of function in genes that are essential for the
proper maintenance of cellular processes such as cell growth, regulation of cell
cycle and apoptosis [10]. Additionally, many genes have been characterized to
fit in into one of two groups; those that promote cell growth (proto-oncogenes)
and those that suppress it (tumor suppressor genes or anti-oncogenes). A lone
mutational event does not result in malignant phenotype, but cumulative events
in a combination of tumor suppressor genes and/or proto-oncogenes can cause
malignancy. A common cause for the accumulation of mutations is the inability to
maintain genomic stability as seen in several forms of heritable cancers that can
be attributed to the loss-of-function of DNA repair elements. Therefore, loss of
DNA repair mediated control of genomic stability results in a predisposition for
cancer. Furthermore, genomic instability caused by improper DNA repair is of
major significance in understanding how cells become malignant. This review will
focus on the importance of model systems for understanding the mechanisms of
mutation that can result in the alteration of function of critical genes, e.g. tumor-
suppressor genes.
Malignant phenotypes that result from alterations of tumor-suppressor
genes are recessive. Therefore at least two steps are required (in diploid cells)
before the phenotype can manifest, as seen in the disease retinoblastoma.
Retinoblastoma is the prototype of heritable cancers. Throughout most of the
world it is a fairly uncommon disease with an occurrence of approximately 5x10-5
[11]. In about 60% of these cases the disease is not inherited [11]. These
13 individuals will develop the tumor in only one eye. The remaining 40% (2x10-5) can be attributed to a germline mutation in the RB1 gene [11] that is on chromosome 13. (RB1, through interaction with transcription factor E2F, is essential in the regulation of the cell cycle.) The individuals who inherit the germline mutation have no family history of retinoblastoma, indicating that mutation occurred in parental germ cells. The mutant allele is passed on, in
Mendelian fashion, to approximately 50% of their offspring. Individuals who have the germline mutation develop multiple tumors during early childhood, usually in both of their eyes [11]. This demonstrates how the familial type of the disease acts in an autosomal dominant fashion. Tumors with germline mutations are nearly distributed in Poisson fashion with a mean of 3 tumors, between both eyes
[11].
An approximate frequency of transforming events can be calculated considering that embryonic retinoblasts, the target cells, are approximately in the order of 107 cells [12]. Therefore the estimated rate of a tumor forming event
occurring can be approximated to be 3x10-7 per cell per generation [12]. Applying
this frequency towards the two cell divisions necessary to form a non-hereditary
tumor, as well as accounting for the multiplication of “once-hit” retinoblasts, shows that the expected incidence would be approximately 3x10-5 [12]. This
corresponds with what has been observed in non-heredity retinoblastoma.
Therefore, both types of the disease can be described in terms of spontaneous
somatic and germline mutation rates [11]. These observations form the basis of
the so-called “two-hit” hypothesis. The phenomena of secondary mutation, the
14 transforming “hit” by which the remaining functional gene (e.g. RB1) is lost, has come to be referred to as “loss of heterozygosity” (LOH). This process is seen in many cancers including hereditary non-polyposis colorectal cancer (HNPCC), familial adenomatous polyposis (FAP), neck, squamous cell, esophageal, and some breast cancers [13-18]; further discussed herein.
Microsatellite instability
Microsatellite instability is another hallmark genetic event in the
transformation process. For example, in HNPCC mutations in genes functioning
in DNA mismatch repair (MMR) are inherited in pedigrees highly susceptible to
colon cancer [4, 19-22]. Efforts have been made to detect MMR-deficient
fractions in human populations, from biological, medical, and social points of
view. Among the MMR genes, the hMSH2 or the hMLH1 gene is most frequently
mutated in HNPCC patients [22].However, in many types of sporadic cancers
hMSH2 or hMLH1 mutation is rare [22]. Since the MMR [22] system repairs
mono- and dinucleotide repeats looping out of the DNA strand, which are caused
by slippage of replication polymerases, alteration of mono- and dinucleotide
repeats in microsatellite DNA sequences, termed microsatellite instability (MSI)
or replication error (RER), has been widely used as a marker to detect MMR-
deficiency. This type of repeat alteration was found in 90% of HNPCC patients
[22] and in 10 ± 30% in sporadic cancer patients.
In the genome, nucleotide repeats are present in a surprisingly high
number of coding regions given their inherent instability. More than 20 human
genes have been reported to have a mononucleotide repeat (of at least 8
15 basepairs) in a coding region and many of these have been observed to suffer
mutation [5-17]. Surprisingly, the list of genes carrying a coding-mononucleotide
repeat includes genes that perform functions critical for preventing mutation.
Four of the genes encoding components of the human DNA mismatch repair
(MMR) system (Msh3, Msh6, Pms2, and Mlh3) contain mononucleotide microsatellites in their coding sequences [17]. Many more human genes are at risk of suffering mutation due to mononucleotide repeat instability. Searches of
the ~33,000 human coding sequences have identified mononucleotide repeats with 9 or more basepairs in 365 entries [16-18].
To facilitate studying these genetic alterations, methods such as the Big
Blue mouse are used to furnish information about mutations in organs of mice.
However this model does not provide any information about the location or type of the individual cells that carry the mutation. In models that allow for in situ
detection of mutant cells, such as the classical coat-spot test and the Dbl-1 test ,
only limited tissues can be studied, and the nature of the mutations are usually
undefined. In addition, available methods for monitoring mutagenic activity do not
lend themselves to ready comparison with carcinogenic activity in the same
histological context. These deficiencies were surmounted by developing
transgenic mice that will allow cells that contain specific mutations to be identified
in situ [23, 24]. Such animals allow rapid detection of the effects of mutagens by
identification of the nature of the mutation induced, and allowing comparisons of
the mutational load of a tissue and the probability of tumor formation in the same
tissue. The same mice can be used to monitor individual cells of an animal for
16 changes brought about by other factors, including mistakes made during
transcription and translation and changes caused by viral infections.
Table 1. Reporter sequence of human placental alkaline phosphatase. Adapted from [24].
In many of the studies described herein, mice expressing the human placental alkaline phosphatase (PLAP) gene were used to detect mutant cells in situ in the tissue of a mouse [25-27]. PLAP mice are transgenic for the G11 allele of a human Placental Alkaline Phosphatase (PLAP) gene driven by a human beta-actin promoter. The G11 allele of the PLAP gene does not produce enzyme due to a frameshift induced by a mononucleotide repeat containing 11 G:C basepairs. Loss of one G:C basepair restores enzyme production (Table 1).
Loss of heterozygosity
Such tract length variation in nucleotide repeats can result in destabilizing events that result in loss of heterozygosity and loss of gene expression at critical loci. By definition, heterozygosity exists when each allele (for any diploid locus) possesses differences in the DNA sequence. As is the case with retinoblastoma
17 and other forms of heritable human cancers, a germline mutation in the familial allele of a tumor-suppressor gene possesses alterations that abolish the production of functional protein from that allele. On a functional basis, the consequences of such inherited heterozygosity manifest themselves when the remaining functional allele is lost at either locus. This is best understood for tumor-suppressor gene loci; RB1, WT1 and TP53 where LOH has been shown to be hallmark in retinoblastoma, Wilms tumor, and Li-Fraumeni syndrome, respectively. In addition, LOH has also been implicated in familial autosomal dominant diseases such as polycystic kidney disease [28], hereditary breast/ovarian cancer (BRCA1, BRCA2 genes)[13, 29] and familial adenomatous polyposis (APC gene)[3]. It has also been suggested that LOH may serve as a source of somatic cell variants that could have selective growth advantages during the course of development [30]. In heritable cancers the first “hit” is acquired as a germline mutation and the second “hit” refers to any events that result in inactivation of the remaining functional allele in the somatic cell [11].
Studying early events that result in LOH will have a significant impact in the understanding of the conversion of normal to malignant cell types by understanding the expression of recessive phenotypes such as impairment/loss of DNA repair mechanisms, cell cycle deregulation, and changes in cell to cell communications [17]. However, it is important to note that although LOH occurs frequently in many cancers, it is difficult to discern whether it accounts for malignant phenotypes or is a result of the increased genomic instability associated with tumors.
18 Normal cellular processes (e.g. DNA repair) have the potential to make
errors that result in the loss of a gene or its product. LOH events can result from
intragenic events that are usually locus restricted such as point mutation, gene conversion, or interstitial deletion. Epigenetic inactivation of an allele can also result in functional LOH [31]. Alternatively, LOH may also arise from the loss of multiple alleles by mitotic recombination (MR) or chromosome loss/re-duplication
(nondisjunction). Of these mechanisms, the most commonly observed in existing models for LOH are chromosomal nondisjunction and mitotic recombination. The former is caused by the mis-segregation of chromosomes during mitosis. If a metaphase cell, with an abnormal mitotic spindle, proceeds through mitosis then the resulting daughter cells can inherit the improper number of chromosomes, as illustrated in Figure 1. The daughter cell that suffers a loss of a chromosome (1n) would usually be expected to die. The other daughter, with an extra chromosome
(3n), can divide and produce normal daughter cells (2n) one of which experiences LOH.
19
Figure 1. Illustrated schematic of chromosome missegregation. Adapted from [32]
Mitotic recombination, usually a DNA repair mechanism in response to double strand breaks (DSB), can also result in LOH. For example, if homologous (non- sister) chromatids, one with a normal allele and the other with a mutant allele, recombine, followed by normal X-segregation, then a daughter cell with LOH is produced (Figure 2). It is important to appreciate that LOH occurred over the entire region beyond the crossover. This feature, LOH at multiple linked loci, allows detection of MR by microsatellite analysis. Both mechanisms produced the LOH at the locus of interest, e.g. a tumor-suppressor gene.
20
Figure 2. Illustrated schematic of homologous mitotic recombination. Adapted from [32]
One model for the detection and study of LOH utilizes mice that contain an interrupted adenine phosphoribosyltransferase gene (aprt). In humans, autosomal recessive Aprt deficiency is a rare genetic disease that frequently results in 2, 8- dihydroxyadenine nephrolithiasis, a form of kidney stones [33].
The Aprt knockout mice have an equivalent phenotype [33]. The Aprt gene
21 resides ~10cM from the telomere on mouse chromosome 8, chromosome 16 in
human [34]. The function of Aprt in mouse and human is well characterized and
is similar in both organisms. Enzyme is expressed in all tissue types and it is
known that both somatic and germline mutations (in most codons) can inactivate
the gene [33]. It has also been shown that the loss of cellular Aprt appears to
have no effect on cell viability [35] Loss of Aprt expression produces a selectable
phenotype due to the inability of the cell to incorporate the toxic adenine
analogue 2,6 diaminopurine (DAP) [35].
Studies in human T-lymphocytes from four individuals who are
heterozygous for Aprt, due to germline mutation in one allele, revealed that the
frequency of DAP resistant (DAPr) cells ranged from 2x10-5 to 15x10-5. Further
analysis showed that 76% of DAPr clones had lost polymorphic markers flanking
the functional Aprt allele, revealing that these clones had experienced LOH at the
Aprt locus. Furthermore, FISH analyses showed that 9 of 10 clones had two copies of Aprt and normal diploid karyotype. These results are all consistent with
MR as the mechanism for LOH [36]. Nonetheless, while these results clearly showed that MR is common in human cells, they left open the possibility that MR is elevated only in T-lymphocytes and not other cell types. T-cells undergo programmed recombination events in order to assemble a functional TCR gene.
A mouse model for the study of Aprt LOH was made by targeting and
thereby functionally inactivating the gene via insertion of the bacterial Neo
cassette [37]. In animals bred heterozygous (Aprt+/neo), skin cells were harvested and immediately placed under selection with DAP. On day 12 clones were
22 analyzed for LOH. DAPr colonies occurred at a median frequency of12x10-5. Of the DAPr colonies, 80% exhibited LOH by MR and the remaining 20% retained the Aprt allele that lacked the Neo cassette. These were shown to harbor point mutations [38].
The Aprt model has been used to detect LOH in several cell types, in response to many factors (e.g. genetic backgrounds and DNA damage). The following experiments are examples of the usefulness of the Aprt model system:
Liang et al. (1995) demonstrated that genetic stability of mouse splenic T- lymphocytes is reduced in animals which lack p53, in response to ionizing radiation. Whole animals were irradiated (4Gy). Two months after treatment T- cells were isolated and placed under DAP selection. Animals that lacked functional p53 showed an eight fold increase in the proportion of DAPr colonies.
Genetic analysis revealed that MR and interstitial deletion were elevated 7-fold and 33-fold, respectively, in p53 null cells. Intragenic events and epigenetic inactivation occurred at similar frequencies among all p53 genotypes. These data suggest that p53 may play a role in homologous recombination and non- homologous end joining, mechanisms of DNA repair, in response to radiation- induced genetic instability (1). Winjnhoven et al. (2003) used the Aprt model to detect LOH, in T-cells, in response to treatments with different classes of chemical carcinogens that are understood to cause DNA lesions, e.g. adducts, methylation, or inter-strand crosslinks. Patterns of LOH varied among hybrid strains analyzed, and among tested chemicals, suggesting strain specific effects.
However, MR was demonstrated to be the primary mechanism for LOH in
23 response to treatments with dimethylbenz[a]anthracene (DMBA) and
methylnitrosurea (MNU) [39]. The Aprt model has also been used to look at
mechanistic differences of LOH in mouse embryonic fibroblasts (mefs) versus
embryonic stem (ES) cells (Cervantes et al. 2002). In mefs, MR accounted for
80% of all DAPr LOH events, as mentioned. However, in ES cells, chromosomal
loss/ reduplication, resulting in uniparental disomy, accounted for half of all DAPr
colonies. In addition Aprt deficient ES cells accumulate with time. This
spontaneous LOH is worrisome due to the potential increase of risk in tumor
formation after stem cell therapy [40].
The studies/experiments described above help demonstrate the potential
usefulness for this model in understanding LOH and its involvement in carcinogenesis. Although, it is important to note that although mutant cells are clonally expanded (for genetic analysis) in vitro, it is likely that they have lost Aprt enzymatic activity in vivo due to both the long half-life of Aprt enzyme and their ability to survive in high concentrations of DAP medium.
There is another mouse model for the study of somatic autosomal LOH
which uses the thymidine kinase (Tk) gene as a marker. Tk is located on a
different chromosome and is closer to the telomere than Aprt, residing on mouse
chromosome 11 ~6cM from the telomere. However, the Tk model is similar to the
Aprt model in that loss of expression results in a selectable phenotype when
cultured in medium with the toxic nucleotide analogue 5-bromodeoxyuridine.
Mutant frequencies in studies with Tk+/neo T-cells are comparable to those in the
24 Aprt model and implicate MR as the primary mechanism for LOH [41]. Tk mutants have not been observed in other cell types [34].
Available mouse models for studying the LOH phenomena have already proved their utility in understanding how LOH is involved in the deregulation of cellular processes and cancer progression. However, they also possess considerable limitations. The models are considered in vivo, but events that result in loss of Aprt activity would have to have occurred prior to cells undergoing selection in culture. The Aprt system has been used in cells from skin, ear, spleen and kidney. That is, only cells of tissues that proliferate in culture can be studied. To date, attempts to apply this assay to other tissues have been unsuccessful.
To study LOH in diverse cell types in the body another reporter model was implemented which distinguishes events by a detectable cellular phenotype.
Mice that carry two different fluorescent protein genes as alleles of a locus were generated to address this issue because LOH would change a cell’s phenotype from bichrome to monochrome. Characterization and studies with this model are discussed herein. The goal of these, and future, studies is to further the understanding the somatic variation among different cell types and their distributions within specific tissues of the whole animal, ultimately unveiling the corresponding mechanisms that result in genetic perturbations that occur during development and cancer progression.
25 Chapter 2
Modeling variation in tumors in vivo
Reference
The majority of this chapter is reproduced from the article, “Modeling variation in tumors in vivo” published in the Proceedings of the National Academy if Sciences
(PNAS). February 15, 2005. Volume 102, Issue 7 pp 2408-13. Therefore, there is some redundancy among chapter introductions.
Purpose of study
It is seen that tumors posses increased frequencies of genetic aberrations, including mutations. However, there is little known about how mutation events contribute to tumor growth (by cell proliferation). By identifying mutant cells within the tissue architecture of solid tumors, it would be possible to model the forces
(e.g. hyperproliferation, mutator phenotype) that contribute to the expansion of the solid tumor. The primary focus of this study was to use the PLAP model for in situ detection of mutant cells to model the spatial distribution of mutant cells within a solid tumor.
Abstract
Transgenic mice that allow mutant cells to be visualized in situ were used to study variation in tumors. These mice carry the G11 PLAP transgene, which is a mutant allele rendered incapable of producing its enzyme product by a frameshift caused by insertion of a tract of G:C basepairs in the plap coding region.
Spontaneous deletion of one G:C basepair from this tract restores gene function, and cells with PLAP activity can be detected histochemically. To study tumors,
26 the G11 PLAP transgene was introduced into the polyoma virus middle T antigen mammary tumor model. Tumors in these mice exhibited up to 300 times more
PLAP+ cells than normal tissues. PLAP+ cells were located throughout each tumor. Many of the PLAP+ cells were singlets, but clusters also were common with one cluster containing more than 30,000 cells. Comparison of these data to simulations produced by computer models suggested that multiple factors were involved in generating mutant cells in tumors. While genetic instability appeared to have occurred in most tumors, large clusters were much more common than expected based on instability alone.
Introduction
Cells in tumors tend to have multiple mutations [42-44]. The presence of multiple changes may be due to genomic instability, where each cell in the tumor exhibits an increased rate of mutation [45]. Alternatively, proliferation could be involved. For example, were a cell to acquire a mutation that causes it and its progeny to proliferate more, this would increase the chance of a second mutation in one of the progeny. Repeated cycles of this process would allow multiple mutations to accumulate in a single genome, even though the rate of mutation per cell per generation is normal [42].
The contributions of genomic instability and hyperproliferation to mutation in cells of a tumor cannot be resolved simply by counting mutant cells because both processes can produce large numbers of these. However, the two processes would be expected to generate tumors with different phenotypes with respect to the positions of mutant cells. If genomic instability predominates, then
27 many independent mutant cells would arise in the tumor. Most of these mutants would tend to reside at locations separate from other mutant cells. By contrast, if
hyperproliferation predominates, then mutant cells would tend to be in a few
large clusters.
The positions of mutant cells can be studied using a transgenic mouse
(G11 PLAP) that allows mutant cells to be visualized in situ in tissue sections [23,
24, 46]. G11 PLAP mice carry a mutant allele of a human Placental Alkaline
Phosphatase (PLAP) transgene (G11). This mutant allele does not produce
enzyme activity because of the presence of a tract of 11 G:C basepairs that shifts ribosomes into the wrong translational reading frame. Studies on cultured cells
carrying this allele showed that deletion of one G:C basepair from this tract
restores enzyme activity and produces a cell that stains histochemically [23, 24,
46].
The G11 tract in the PLAP transgene mimics mononucleotide repeats in
the genome, which, considering their propensity to suffer mutation via deletion or
insertion of basepairs during DNA replication, are surprisingly common in coding
regions [47]. More than 20 human genes have been reported to have a
mononucleotide repeat (of at least 8 basepairs) in a coding region, and many of
these have been observed to suffer mutation. The list of genes carrying a coding-
mononucleotide repeat includes genes that perform functions critical for
preventing cancer, such as DNA repair (RAD50, DNA-PKcs, hMSH3, hMSH6,
XPG), and regulation of cell proliferation, apoptosis, and tumor growth
(TGFbetaRII, IGF2R, CHK-1, BAX, ICE, caspase-5, FLASH, Apaf- 1, E2F- 4,
28 TCF-4, Apc c-myb, MYCL, CtIP, MLSN1) [21, 48-57]. Many more human genes
are at risk of suffering mutation due to mononucleotide repeat instability. A
search of ~33,000 human coding sequences identified mononucleotide repeats
with 9 or more basepairs in 365 entries [58]. Another study found 336 messenger
RNAs containing either G7 or C7, and 4382 containing either A7 or T7 [59].
Previous studies on the G11 PLAP mouse showed that PLAP+ cells were easily
detectable in cells in the four organs examined in detail (brain, heart, kidney,
liver) [60]. The average frequency of PLAP+ cells in normal tissues was
approximately 1.5x10-5, far lower than can be detected by DNA amplification, but
consistent with results of selection experiments, which have shown that
mononucleotide tracts act as hot spots for frameshift mutations in vivo in lymphocytes [61].
It is well known that cells defective for mismatch repair (MMR) exhibit a
High- Microsatellite-Instability (MSI-H) phenotype, whereby changes in simple
sequence tracts occur so frequently that populations of cells tend to be
heterogeneous with respect to the length of any given tract [21]. Cells proficient
in repair do not exhibit MSI-H, but evidence is accumulating suggesting that low-
level instability can occur [21]. A so-called Low-Microsatellite-Instability phenotype (MSI-L) has been reported in cells from tissues that are either inflamed or hyperplastic, and in carcinogen induced rat mammary tumors [51, 62,
63]. In addition, changes were seen in somatic cell microsatellites of offspring of individuals exposed to radiation from the Chernobyl disaster [64]. The PLAP+ cells present in MMR proficient mice may be formed by the same process that
29 causes MSI-L, and the G11 allele may respond to the same kinds of environmental insults.
To study mutation in tumors in situ, the G11 PLAP gene was introduced into a mammary tumor mouse model [65]. These mice carry a transgene that causes expression of the polyoma virus middle T antigen in mammary cells.
Middle T-antigen is located at the cell membrane where it binds and activates kinases that ultimately activate the Ras signal transduction pathway [66, 67]. The middle T-antigen is not in the nucleus, and thus not directly involved in DNA replication or repair. Nevertheless, it was reasonable to anticipate that cells in this tumor model might exhibit reduced genetic stability because middle T- antigen activates the Ras pathway, and such activation has been linked to genetic instability [68-70].
Tumors in these mice contained up to 300 times more PLAP+ cells than normal tissues. PLAP+ cells were located throughout each tumor and many were situated as singlets, suggesting that genetic instability occurred. However,
comparison of the tumor data to simulations produced by computer models
suggested that additional factors were involved in generating mutant cells in
these tumors.
Results
Tumors exhibited an elevated number of mutant cells
To assess the instability of the PLAP-negative phenotype in situ, mice
were sacrificed and tumors and major organs were removed and frozen.
Sections were cut from frozen tissues and PLAP+ cells were detected by
30 histochemical staining. PLAP+ cells were scored in 18 tumors from 10 G11 mice.
Normal breast, brain, heart, kidney and liver were also analyzed. MMTV PyVT mice that lacked the PLAP gene provided 11 control tumors. No PLAP+ cells were observed in such mice.
PLAP+ cells were common in sections cut from tumors. Figure 1 shows some examples of the different section phenotypes observed. Tumors exhibited many more PLAP+ cells than normal tissues (Figure 2A). The arithmetic mean frequency in tumors was more than 100 fold greater than that in heart, which was the normal tissue that exhibited the most PLAP+ cells (Figure 2A). The frequency varied greatly among the tumors, and one tumor (714) exhibited a very large number of PLAP+ cells (108,513), which inflated the mean. However, 16 of the other 17 tumors exhibited at least 10 times more PLAP+ cells than normal breast tissue. The mean frequency in these 17 tumors was 2036, which is 85 fold greater than that in normal breast. The median frequencies were 1378 and
24 in tumors and normal breast tissues, respectively. Normal breast tissue from mice carrying the PyMT transgene did not exhibit more PLAP+ cells than breast tissue from mice that lacked the PyMT transgene. The viral onco-protein was shown to express in most, if not all mammary cells of MMTV PyVT mice by immunohistochemical analysis of frozen sections from mammary glands (data not shown). The MMTV promoter is estrogen dependent, and is not very active in the other tissues analyzed (brain, heart, kidney and liver). Therefore, it was expected that the mean frequencies of PLAP+ cells in tissues of MMTV PyVT mice would not differ from those previously observed in mice lacking the PyMT
31 transgene. Comparison of the MMTV PyVT to those previously obtained showed
this to be the case [60]. The frequency of PLAP+ cells in tumors was not
correlated with frequency in the normal tissues of the same animal. None of the
animals studied exhibited a generalized high level of mutation. Therefore, the
phenotypes of the tumors were a feature of tumor cells.
Figure 1. Representative images of mammary tumor sections stained for PLAP activity. Cells with PLAP activity are dark. (A): Tumor from a mouse that did not carry the PLAP gene. Line indicates 66 microns. (B): Arrow indicates a single PLAP+ cell. Line indicates 265 microns. (C): Arrow indicates a group of PLAP+ cells. Line indicates 1300 microns. (D): A very large group of PLAP+ cells. Line indicates 1300 microns.
The increased number of PLAP+ cells in tumors was not due to more
extensive expression of the PLAP gene in tumors. Expression was assessed by
32 examining tumors and tissues from mice that carried a revertant allele of the G11
PLAP transgene. In studies to be described elsewhere, mice carrying a reverted
G11 PLAP gene were obtained by passing this gene through female mice that
lacked a functional copy of the mismatch repair gene Pms2. Some of the
offspring from these females were found to express PLAP in tail tissue. These
putative germline revertant mice were crossed to wild type mice to verify that the
PLAP+ phenotype was transmitted as a dominant trait. Approximately half of the
cells in tumors and tissues from mice carrying the germline revertant allele were
positive for PLAP activity. Therefore, the PLAP gene was transcribed and
translated in the same fraction of cells in tumors and normal tissues.
Prevalence of single mutant cells suggests genetic instability. PLAP+ cells were situated both alone and in clusters. Single PLAP+ cells were of particular
interest because their frequency should not be affected by proliferation of PLAP+ cells. Hence, the frequency of single PLAP+ cells should be indicative of the rate of mutation in a given tissue. Compared to normal tissues, tumors tended to exhibit many more single PLAP+ cells. Figure 2B shows that tissues exhibited
approximately 4 single PLAP+ cells per million cells (median 3.5). This value was
similar to that reported previously [60]. By contrast, the median frequency of
single PLAP+ cells in tumors was close to 50 per million. In addition there were
more single PLAP+ cells in all but 3 of 13 tumors. In large populations, mutation
rates can be approximated from the frequency of mutants [71]. Therefore, the
rate of mutation in tumors appeared to be at least 10 fold higher than in normal
cells.
33
Figure 2. Frequencies of PLAP+ cells in tumors and normal tissues. (A): Total PLAP+ cells. (B): Single PLAP+ cells. Tissue data were from tissues listed in panel A.
Clusters of mutant cells.
If mutants arise at a constant rate at random times, and proliferate normally, then the sizes and frequencies of clusters are predictable. Under these conditions, the points on graphs that plot the product of cluster size and frequency versus cluster size will fall on a horizontal line. Figure 3 shows that normal tissues conformed to this expectation. The line produced from tissue data was similar to that produced by a computer simulation that employed random mutation at a rate of 1 x 10-6 events per cell.
When considered together, tumors exhibited cluster phenotypes that also tended to conform to expectations based on stochastic mutation followed by uniform proliferation. Figure 3 shows that the median values obtained from tumors
34 produced a line that was quite similar to a computer simulation that employed
random mutation at a rate of 1 x 10-4 events per cell. The median- value line
ends at 32 cells because less than half of the tumors exhibited larger clusters.
However larger clusters were observed in some tumors (see Figure 1 and
below).
While in the aggregate, tumors tended to produce clusters in numbers
consistent with a simple stochastic model, individual tumors exhibited different
cluster phenotypes, many of which were dramatically different from the
predictions of simple models. Figure 4 shows the relationship between cluster
size, and the product of cluster size and frequency for each of the tumors. In
order to facilitate plotting all of the data on the same scale, cell-number values
were normalized by dividing them by the number of single PLAP+ cells observed
in a given tumor. When all of the data were plotted, Figure 4a, three tumors
(125b, 123b, and 714) stood out. Tumor 714 had a very large region of
contiguous PLAP+ cells (see Figure 1.). All sections from this tumor had a very large patch of PLAP+ cells, which contained approximately 30,000 cells.
Similarly, most sections from tumor 123b contained a patch with approximately
1000 PLAP+ cells. Tumor 125b had a different phenotype. One section from
tumor 125b exhibited 37 small clusters, each containing approximately 30 cells.
The shapes of the lines plotted for tumors 714, 123b and 125b indicated that they
all had an overabundance of large clusters compared to what would be expected
from the number of single PLAP+ cells. The sizes of the overabundant clusters
varied in the three tumors. Tumor 125b was at one extreme, with an apparent
35 excess of clusters larger than 16 cells. Tumor 714 was at the other extreme, with
a very large cluster containing tens of thousands of cells. Tumor 123b exhibited
an intermediate phenotype, with a large number of clusters larger than 128 cells.
It seemed possible that these large clusters formed via hyperproliferation of
PLAP+ cells.
Figure 3. Median frequencies of clusters of PLAP+ cells in tissues and tumors conformed to expectation based on random mutation followed by normal proliferation in an exponential growth model. The values on the Y axis are the product of cluster size and cluster frequency. Dotted lines labeled sim10exp-4 and sim10exp-6 show median values obtained from three simulations performed with the probability of mutation set at 1 x 10-4 and 1 x 10-6, respectively.
To examine this possibility, a computer model that caused mutant cells to proliferate twice as fast as normal cells was developed. While this model
36 produced results that resembled the data from tumor 125b, the model data did
not fit data from tumors 123b and 714 (Figure 4a). These results suggest that
PLAP activity did not directly cause hyperproliferation. Instead, it appears that
the excess of larger clusters was formed by hyperproliferation of a subset of the
PLAP+ cells in the tumor.
The large numbers of PLAP+ clusters in tumors 714, 123b and 125b
imposed a graphical scale that obscured details in the data from the other
tumors. Therefore, a second graph, Figure 4b, which did not include data from
these three tumors, was constructed. Figure 4b shows that tumors varied
greatly with respect to PLAP+ cell cluster frequencies. To assess the significance
of this variation, cluster counts from each tumor were subjected to probability
analysis. The frequency of single PLAP+ cells in tumors was used to calculate
the probabilities that larger clusters would form at the frequencies observed,
assuming random mutation and exponential proliferation of all cells. For
example, if the frequency of single PLAP+ cells in a tumor were 100, then random
mutation and exponential proliferation would be expected to produce approximately 50 2-cell clusters, 25 4-cell clusters, etc. Using these expected
values, the probability of the observed frequency of each cluster (p) can be
estimated via Fisher’s exact test. Figure 5 shows some of the results of the
calculations performed on 11 tumors. The two panels in Figure 5 display data
from tumors selected and grouped according to cluster phenotypes. Figure 5a
shows data from the four tumors that exhibited cluster frequencies that were not
extremely improbable. The p values for these tumors tended to exceed 0.05,
37 although several values were as low as 0.01. Figure 5b shows the opposite
extreme in tumor phenotypes. These three tumors deviated markedly from
expectations. Many of the p values were in the range of 10-6, and some were
many orders of magnitude lower. The other four tumors subjected to probability
analysis were more similar to those in Figure 5b than to those in Figure 5a.
Therefore, formation of clusters of PLAP+ cells in most tumors was not governed by normal proliferation of randomly mutated cells. Instead, other factors, which presumably varied among tumors and within them, produced the highly idiosyncratic phenotypes observed.
Figure 4A
38
Figure 4B
Figure 4. Variation in frequencies of PLAP+ cell-clusters in tumors. For each tumor, the relative number of clusters of a given size was calculated by dividing the number of clusters with more than one PLAP+ cell by the number of single PLAP+ cells. The values on the Y axis are the products of these numbers and cluster sizes. A) Data from all tumors. The simulation allowed mutant cells to double twice as fast as normal cells. .B) Data from all tumors except 123b, 125b and 714.
39
Figure 5A
Figure 5B
40 Figure 5. Probabilities of observed cluster frequencies in tumors. For each tumor, the expected numbers of clusters of various sizes were calculated based on the frequency of single cells and assuming random mutation and exponential growth of all cells. The probabilities of the observed cluster frequencies are plotted on the Y axis. A) Data from the four tumors that deviated least from expectations. B) Data from the three tumors that deviated most from expectations.
Discussion
In principal, genetic instability and hyperproliferation can contribute to the number
of mutant cells in tumors. Both mechanisms would appear to be involved in
generating the PLAP+ cells observed in the MMTV PyMT tumor model.
A role for genetic instability is suggested by the high numbers of single PLAP+
cells, which would be expected if there were an increased rate of mutation per
cell per generation. Alternatively, some or all of the isolated PLAP+ cells may
have been produced by cell migration. However, migration does not explain the
tendency of median tumor cluster counts to conform to expectations based on
random mutation and exponential expansion of both mutant and wildtype cells.
In addition, studies on genetic heterogeneity of human tumors suggest that cell
migration is limited [72]. If genetic instability did occur, it probably was not
directly due to polyoma virus middle T-antigen. The viral protein is not in the
nucleus, and thus not directly involved in DNA replication or repair. Middle T-
antigen is located at the cell membrane where it binds and activates kinases that
ultimately activate the Ras signal transduction pathway [66, 67]. The connection
of PyMT to Ras signaling may be pertinent because previous studies have
shown that expression of activated H-Ras in transformed cultured cells can
41 cause chromosome aberrations [68-70, 73, 74]. H-Ras-mediated genome destabilization has been shown to proceed through the MAP kinase pathway [70,
73]. The mechanism involved at the DNA level is not known, but chromosome aberrations suggest increased DNA breakage. Thus, Ras-pathway dysfunction is a candidate for causing the observed instability in tumor cells expressing polyoma middle T-antigen (PyMT). However, it is important to note that normal mammary tissue expressed PyMT but did not exhibit the PLAP instability phenotype. Hence, it seems that activation of the Ras pathway is not sufficient to cause instability. Alternatively, this pathway may not be activated by the PyMT protein in morphologically normal mammary cells. This possibility seems remote given the effects that PyMT is known to exert directly on the tyrosine kinases that activate this pathway [66]. If the Ras pathway is activated in normal mouse mammary cells, then this change is insufficient to transform them because the number of tumors is very small compared to the number of mammary cells. Lack of transformation of mouse mammary cells by activation of the Ras pathway alone is in agreement with findings obtained in experiments on a human breast epithelial cell line transfected with activated H-ras [75]. Insufficiency of activated
Ras for in vivo tumorigenesis fits with the well established observation that formation of an activated ras gene is not sufficient to confer full transformation on cultured cells [76, 77]. Full transformation usually requires at least one additional change, and such changes often impair functions that normally maintain genome stability. In human mammary tumors, a destabilizing mutation often occurs in
BRCA 1 or BRCA 2 [78-81].
42 The mutations that collaborate with PyMT to induce the tumors studied here are not known, but previous expression profiling studies suggest that repair functions in most of the cells in each tumor are normal. PyMT-induced murine mammary tumor cells exhibited a distinct expression profile that resembled those of tumors induced by either Ras or Neu [82]. This phenotype differed from those of tumors induced by SV40 large T-antigen, which directly influences p53 function. Neither PyMT nor SV40 T-antigen caused changes in the expression of
DNA repair genes. Although the expression profile did not detect changes in repair functions, it is possible that activation of the Ras pathway by PyMT caused increased chromosome damage, which in turn caused some cells to lose a function, such as mismatch repair, which is needed to maintain the PLAP G11 allele. Cells that are rendered less able to repair can mutate the G11 allele at a very high rate. Lack of mismatch repair has been shown to increase the rate of
G11 mutation at least a thousand fold [60]. It is relevant to note that mutations in mismatch repair genes have been seen in cultured transformed human breast epithelial cells [83]. Physiological stress in the tumor may also be pertinent. A study of inflamed colon epithelium showed that this tissue exhibited microsatellite instability and overexpressed two base excision-repair enzymes, AAG, the major
3-methyladenine DNA glycosylase, and APE1, the major apurinic site endonuclease. These enzymes appear to have the ability to cause microsatellite instability because instability was induced when they were overexpressed in yeast and cultured human cells [84]
43 Whereas genetic instability is a plausible explanation for the numerous
single PLAP+ cells in tumors, the frequent occurrence of clusters of PLAP+ cells
suggests that abnormal proliferation was also a major factor. However, simple
hyperproliferation of PLAP+ cells does not explain the tumor phenotypes,
because this phenomenon would produce more clusters of every size. In fact,
cluster counts in tumors were extremely variable. In addition, studies on mouse
3T3 cells expressing PLAP showed that PLAP had no effect on proliferation in culture (J. R. Stringer, unpublished observation). Furthermore, expression of ectopic alkaline phosphatase did not cause human cells to become oncogenic in nude mice [85].
Multiple factors appear to have been involved in generating the
phenotypes of most the tumors studied. One possibility is that different tumors,
and, or, different segments of a given tumor mass contained cells that behaved
differently, either with respect to mutation, proliferation, or both. Phenotypic and
genotypic heterogeneity is common in human tumors [20, 86-92]. The evolution
of heterogeneous tumors can be simulated by mathematical models that include
selection in addition to mutation and proliferation (or lack thereof) [93]. It seems
probable that all three forces shaped the tumors in the MMTV PyMT model, and
that the PLAP+ phenotype served as a neutral indicator of tumor heterogeneity.
Materials and Methods
Mice
Transgenic FVB/N mice carrying the G11 PLAP frameshift reporter gene
were generated as described in [23, 24]. Transgenic mice that develop
44 mammary tumors with high frequency (strain FVB/N-TgN(MMTV PyMT)634Mul)
were acquired from The Jackson Laboratory (Bar Harbor, Maine). Heterozygous
MMTV PyMT mice were mated with G11 PLAP mice from line 100A [23, 24, 60].
The PLAP gene was detected by amplification of tail DNA as previously
described [60]. The MMTV PyMT transgene was detected in tail DNA using the
PCR protocol provided by The Jackson Laboratory (http://jaxmice.jax.org). Mice
were observed until tumors became apparent, which typically happened by the
age of 6 months regardless of sex. Tumors were taken from 20 mice, 4 males
and 16 females. Ten of these mice (3 males, 7 females) carried the G11 PLAP
transgene. When multiple tumors occurred in a mouse, they were labeled with
the mouse-identifier number followed by a letter, such 125a, 125b, etc. Tumors in
one location were often comprised of more than one mass. It was not possible to
determine if such tumors were from a single source or not. Tumors varied in size,
shape and solidity. The largest was approximately 6 cm3 in volume. Most
however were between 1 and 3 cm3. There was no relationship between tumor
size, shape or solidity and PLAP phenotype.
Preparation and staining of tissues.
Organs and tumors were removed and either frozen in OCT embedding
compound on dry ice or snap frozen in liquid N2-cooled isopentane. Cryosections
were cut 10μm thick, mounted on slides, and fixed in 2% formaldehyde/ 0.2%
glutaraldehyde in phosphate buffer (0.15 M NaCl/ 2.7 mM KCl/ 1.47 mM KH2PO4/
4.86 mM Na2HPO4, pH 7.4) (PB) for 10 minutes at room temperature.
Endogenous phosphatases in fixed tissues were inactivated by incubation of
45 mounted sections at 65oC for one hour in PB. To detect PLAP activity, sections
were incubated in 1 mg/ml 5-Bromo-4-Chloro-3-Indolyl Phosphate and Nitro-BT
(BCIP/NBT) in 0.1M Tris, pH 10 for 30 - 90 minutes at 37oC. Sections were then
washed in PB and stained for 5 minutes in Nuclear Fast Red. Stained sections
were dehydrated by sequential washes with aqueous in ethanol starting at 10%
ethanol and ending with 90%. Dehydrated slides were dried in air, and sections
were placed under cover slips using permount.
Enumeration of PLAP+ cells.
Images were captured with a Spot Jr digital camera (Diagnostic
Instruments). The number of cells examined per tissue sample was estimated as
follows. The cells in a field of view (FOV) at 400X magnification were counted,
and the area occupied by these cells was determined. The area of each section
was measured from a digital image at 20X magnification. Corrections were made
to account for holes in sections. The number of cells in the section was computed
from the ratio of section and 400X FOV areas. The total number of tumor cells examined exceeded 1 x 108. The total number of PLAP+ cells observed in tumors
exceeded 1 x 105. The number of PLAP+ cells in each tumor section varied, but
at least 60 PLAP+ cells were counted in each tumor. This quota was easy to
reach because most tumors had more than 100 PLAP+ cells per section, and on
the order of 1000 PLAP+ cells were scored in most tumors. Normal tissues
exhibited far fewer PLAP+ cells. Counting was pursued until at least 20 PLAP+
cells had been seen in each tissue sample examined.
46 Scoring individual and clusters of PLAP+ cells.
A PLAP+ cell was scored as a singlet if there were no other PLAP+ cells
within an area defined by 10 cell diameters, and preceding and succeeding serial
sections lacked a PLAP+ cell in the location of interest. Clusters of PLAP+ cells
were scored on individual sections, which defined them in two dimensions (X and
Y). The Z dimension was examined by viewing serial sections, which showed
that, in general, clusters extended into the Z dimension, and that the larger the
cluster, the more sections it occupied. These data confirmed expectations that
clusters would occupy three dimensional space. However, data from the X-Y
dimension were sufficient for determining the relationship between cluster
frequency and size, and ignoring the Z dimension simplified the analysis.
Therefore, only data from individual sections (X-Y data) were used to define
these relationships.
Computer models
The genesis and proliferation of mutant cells in a tumor was simulated by
an algorithm that did the following: Each simulation started with a single PLAP-
negative cell, which produced two cells, which produced four cells etc, until 1
million cells were produced. At each doubling, the number of mutant cells
produced was determined by multiplying the number of cells present by the probability of mutation. Simulations were performed with the probability of mutation set at either 1x10-4, 1x10-5, or 1x10-6, and 100 simulations were
performed for each probability. Spatial associations of PLAP+ cells originating
from simulated spontaneous random mutations were represented in one-
47 dimensional space. Clusters of PLAP+ cells were defined as adjacent PLAP+ cells in the string of simulated cells. To assess the effects of hyperproliferation of mutant cells, the algorithm was modified to cause each mutant cell to double twice for each doubling of non-mutant cells.
Probability analysis
If mutations are generated during DNA replication and the probability of mutation is p events per cell per replicative cycle, then there will be p(N/2) single- cell mutants present in a population of N cells because these mutants were generated during the previous round of cell division. If the population increases uniformly by doubling each cell in it at each generation, then the number of clusters containing 2 cells will be p(N/4) because these pairs of cells came from single cells in a previous generation when the population size was N/2. This relationship extends to clusters of all sizes and the number of clusters containing
2K cells will be p(N/2k+1). The proportions calculated for different K’s were
compared to observed proportions using the Fisher’s exact test for the following
2x2 table.
# of single mutants (N/2) - (# of single mutants)
# of clusters containing 2K (N/2k+1) - (#of clusters containing 2K
mutants mutants)
48 Chapter 3
Increased mutation in mice genetically predisposed
to oxidative damage in the brain
Reference
This chapter is reproduced from the article, “Increased mutation in mice
genetically predisposed to oxidative damage in the brain” published in Mutation
Research: Fundamental mechanisms if mutagenesis. November 22, 2004.
Volume 556, Issue 1-2 pp 127-34. Therefore, there is some redundancy among chapter introductions.
Purpose of study
Harlequin mice present a neurodegenerative phenotype. Studies suggest that the
primary cause is oxidative stress due to inactivation of the Aif gene. Since
oxidation of DNA is mutagenic it seemed plausible that mutation may have a role
in the neurodegenerative phenotype in these mice. More specifically, we sought
out to identify the cell type and region where mutation occurs within the brains,
compared to wildtype and other tissues, undergoing oxidative stress.
Abstract
Harlequin (Hq) mice develop ataxia due to an X-linked recessive mutation
in the gene encoding Apoptosis-Inducing Factor (Aif). Brain cells in Hq mice contain the modified base 8-hydroxydeoxyguanosine (8-OHdG), suggesting that the defect in Aif causes increased DNA oxidation in these cells. Because oxidative damage is mutagenic, Hq mice might suffer increased mutation in the
49 brain. To examine this possibility, mutation in the brain was assessed using the
Tg(βA-G11PLAP) mouse model, which allows mutant cells to be visualized in
tissue sections in situ. Hq/Y mice exhibited more and larger patches of PLAP
positive tissue in the brain. PLAP+ cells were observed in all areas of the brain.
No increase in the number of PLAP+ cells was seen in three other tissues,
suggesting that the effect of Aif deficiency on mutation was specific to brain.
Introduction
Harlequin (Hq) mice carry an X-linked recessive mutation in the Pdcd8
(programmed cell death 8) gene, which encodes a protein known as Apoptosis-
Inducing Factor (Aif) [94-96]. While obviously named for its role in activating apoptosis [96], in the brain, Aif protein appears to act to prevent apoptosis by protecting against oxidative damage [95]. Such a role is presumably mediated by the NADH oxidase activity of Aif. Aif contributes a major fraction of the NADH oxidase activity released from the mitochondrial intermembrane space [97]. The
Hq allele of the Pdcd8 gene contains an intronic insertion of retrovirus proviral
DNA, which reduces production of Aif to a level less than 20% of normal [95]. As
Hq mice age, the amount of the modified base 8-hydroxydeoxyguanosine (8-
OHdG) increases in a variety of cell types in the brain, including cerebellar
granule cells, which are less abundant in older, ataxic Hq mice [95]. Oxidative
damage due to Aif deficiency has been implicated in loss of cerebellar granule
cells [95]. Explanted cerebellar granule cells from Hq mice are more susceptible
to peroxide-induced apoptosis, but this sensitivity was reduced when the cells
were transfected with a plasmid encoding Aif [95].
50 Oxidative damage is mutagenic [98-103]. Therefore, Hq mice might be
expected to suffer increased mutation in the brain. Oxidation-induced mutation
should be detectable using the Tg(βA-G11PLAP) mouse model, which allows mutant cells to be visualized in tissue sections in situ [23, 46, 60]. Guanosines are targets of oxidative damage and this damage causes loss of nucleotides from tracts of GC base pairs [23, 45, 46]. Tg(βA-G11PLAP) mice carry a mutant allele of a human Placental Alkaline Phosphatase (PLAP) transgene (Tg). This allele does not produce enzyme activity because of the presence of a tract of 11 G:C base pairs (G11) that shifts translation out of the correct reading frame. Deletion of one G:C base pair from this tract restores enzyme activity, which can be detected in situ by histochemical techniques.
The βA-G11PLAP allele mimics mononucleotide repeats in the genome,
which, given their inherent instability, are present in a surprisingly high number of coding regions. More than 20 human genes have a coding mononucleotide repeat at least 8 base pairs long, and many of these repeats have been observed
to suffer mutation [21, 22, 48-55, 57, 104]. The list of genes carrying a coding
mononucleotide repeat includes genes involved in DNA repair, regulation of cell
proliferation, apoptosis, and tumor growth. Many more human genes are at risk of suffering mutation due to mononucleotide repeat instability. A search of
~33,000 human coding sequences identified mononucleotide repeats with 9 or
more base pairs in 365 entries [58].
To study the effect of the Hq allele on mutation, the βA-G11PLAP allele
was introduced into Hq/Y mice. These mice exhibited significantly more PLAP+
51 tissue in all areas of the brain. In contrast, hearts, kidneys and livers did not
exhibit more PLAP+ cells, suggesting that increased mutation was specific to
brain. PLAP+ cells occurred in all regions of the brain in both wild type and Hq/y
mice, but PLAP+ areas were more numerous and larger in Hq mice.
Results
More and larger areas of tissue in Aif-deficient brains showed evidence of
mutation.
The effect of Aif deficiency was assessed by comparing the numbers of
PLAP+ areas per brain. Each stained area was scored as one event. Events
were chosen as the endpoint for this study because the anatomy of brain cells
can complicate attempts to count individual stained cells. PLAP is attached to
the outside of the plasma membrane and is found on the membranes of neuronal
axons and dendrites that extend far from the neuron cell body [105]. This property can make it difficult to locate the nucleus of the cell that expresses
PLAP.
The mean frequency of PLAP+ events was three fold higher in Hq/y mice. (Figure
1, Table 1). The difference between the mean frequencies of PLAP+ events in
the brain of wild type and mutant mice was statistically significant (p<0.01).
PLAP+ cells were not observed in the brains of five control mice lacking the βA-
G11PLAP gene.
52
[26]
To compare the sizes of PLAP+ areas, the numbers of nuclei in these areas were determined. To aid in this process, alternate sections were stained with propidium iodide and BCIP/NBT respectively and the two images merged.
As shown in Figure 2, the brains from Hq mice had more PLAP+ stained areas of all sizes and stained areas in Hq mice were associated with more nuclei. Single
PLAP+ events were three fold more frequent in Hq mice. Summation of all nuclei associated with PLAP+ areas showed that Hq/y mice had nearly six times as many such nuclei as controls. The brains of Hq mice as young as 12 weeks of age exhibited high numbers of PLAP+ areas (Table 1). The frequency of PLAP+ cells did not increase with age (Table 1). By contrast with the effect seen in the
53 brain, Pdcd8 genotype did not affect PLAP+ cell frequency in heart, kidney and
liver (data not shown).
Figure 1. Number of PLAP+ events in brain. Each point shows the number of PLAP+ regions observed per million brain cells examined in an individual mouse with indicated genotype. Bars indicate the mean value. [26]
Locations and morphologies of PLAP+ events in the brain.
Brains were sectioned in the sagittal plane starting at the midline and proceeding approximately 0.5 mm into the organ. Figure 3A illustrates the major regions of the brain typically present in a sagittal brain section [106, 107]. While individual sections lacked PLAP+ cells in some brain regions, when all sections from all mice were included, PLAP+ cells were observed in all brain regions in both wt and Hq mice (data not shown).
54
Figure 2. Sizes of PLAP+ events. The quantity of cellular nuclei located in areas stained by BCIP/NBT were scored in wildtype and Hq/y mice. [26]
Some examples of PLAP-staining are shown in Figure 3. The identity of
PLAP+ cells was often suggested by location and/or shape of a stained area.
Figure 3B shows PLAP+ cells located on the surface of the brain, which is the
pia-arachnoid layer [107, 108]. At least half of the cells in the pia-arachnoid layer
of this animal were PLAP+. This pattern of staining could have been caused by a
mutation that occurred during development. Figure 3C shows a PLAP+ focus in
the Thalamic Reticular Nucleus abutting the fimbria of the hippocampus (axonal
bundle) in the cerebrum [106, 107]. The stained area contains large densely
55 staining cell nuclei with clear soma boundaries, suggesting that these are pyramidal cells. The staining in this region extended through five serial sections.
Figure 3D shows an example of staining in the molecular layer (region marked
“m” in the Figure) of the cerebellum. The pattern of staining is consistent with the anatomy of a Purkinje cell. The cell bodies of Purkinje cells reside at the border between the granular (region marked “g” in the Figure) and molecular layers of the cerebellum. Purkinje cells have an extensive dendritric networks that extend through the molecular layer. The streak of staining resembles such a network.
This pattern of staining was observed in three consecutive serial sections.
Figure 3E shows an example of staining in the granular layer of the cerebellum.
Figure 3F shows a cluster of PLAP+ structures in the brainstem, which is rich in myelinated axons. The myelin surrounding axons of the central nervous system is provided by oligodendrocytes, which can myelinate many segments on multiple axons [108]. The arrow shows what appears to be a cross-cut PLAP- negative axon with PLAP+ myelin wrapped around it. The PLAP+ parallel lines resemble myelinated axons cut longitudinally.
56
Figure 3. Examples of PLAP+ events in different brain regions. Panel A: Illustration of a sagittal section of the mouse brain. The major regions are labeled as follows: CE, cerebrum, CB, cerebellum, BS, brain stem. Lines labeled B-F show the locations of the PLAP+ areas shown in panels B-F, which show Hq brain sections stained PLAP and counter stained with nuclear fast red. Panel B: Arrow indicates PLAP+ labeled cells in the pia-arachniod layer. Panel C: PLAP+ cells in the thalamic reticular nucleus of the cerebrum. The PLAP+ event contains approximately 50 apparent pyramidal cell nuclei. Panel D: Cerebellum white matter (w), granular layer (g) and molecular layer (m) with a PLAP+ Purkinje cell body and dendritic extensions. Panel E: PLAP+ focus localized to the granular layer of the cerebellum. Panel F: PLAP+ region in the brain stem. The arrow indicates a PLAP+ oligodendrocyte wrapped around an unstained axon. [26]
57 Discussion
The locations of cells exhibiting PLAP activity suggests that Aif deficiency adversely affected genome stability in a wide variety of brain cell types in mice as young as 3 months of age. These data contrast with previous studies on Hq mice, which found that granule cells in the cerebellum were present in lower numbers at 7 months of age but not at 3 months involved [95]. No other brain region was noticeably degenerate. On the other hand, the previous studies also showed that a variety of cell types had increased levels of catalase, glutathione, lipid peroxidation and the modified base 8-hydroxydeoxyguanosine (8-OHdG), suggesting that many cell types were subjected to increased oxidative stress, which is known to be mutagenic [95]. Oxidative stress would be expected to cause mutations more readily than cell death. It is possible, however, that the increased mutational load in a broad spectrum of brain cells reflects a process that contributes to the gradual degeneration in brain function seen in these mice.
Previous studies showed that some granular layer neurons in Hq mice express S-phase proteins and incorporate DNA precursors [95]. However, these cells did not proliferate, but instead exhibited signs of apoptosis and declined in number with advancing age. These data led to the hypothesis that the expression of S-phase markers reflects nonproductive entry into the cell cycle, which leads to apoptosis rather than to proliferation [95]. By contrast, the patches of PLAP+ cells observed in the brains of Hq/Y mice suggest that some brain cells may have proliferated after mutation. Proliferation of mutant cells could have occurred either during brain development, or in adult mice.
58 Oligodendrocytes, astrocytes and stem cells can proliferate in adult mammals
[109]. Furthermore, microglial cells, which originate in the bone marrow, can be
recruited to sites of damage. However, the data are insufficient to exclude an
alternative explanation. The larger patches of PLAP+ brain tissue could have been due to a shift in the kinds of cells mutated in Hq/Y mice. Many neuronal
cells have extensive networks of dendrites in the cerebrum and cerebellum. If
the affect of the Hq mutation were to cause mutation in such cells, this would
produce larger PLAP+ areas in Hq mice.
Materials and Methods
Chemicals.
All chemicals were obtained from Fischer Scientific International
(Pittsburgh, PA) unless otherwise noted.
Transgenic Mice
Transgenic FVBN mice carrying the βA-G11PLAP frameshift reporter gene were generated as described [23, 24, 46]. Hq mice (B6CBA Ca Aw-J/A-
Pdcd8Hq/J Stock Number: 000501) were acquired from The Jackson Laboratory
(Bar Harbor, Maine). Female mice heterozygous for Hq were mated with males
hemizygous for Tg(βA-G11PLAP). Genotypes of PLAP mice were determined by
amplification of DNA obtained from mouse tail clippings by methods previously
described [23, 60, 95]. Hq genotypes were determined by amplification of DNA
obtained from mouse tail clippings as described by Klein et al [95]. The following
three primers were used in a single reaction: Forward and reverse primers from
intron 1 of the Pdcd8 gene (5'-AGTGTCCAGTCAAAGTACCGGG-3' and 5'-
59 CTATGCCCTTCTCCATGTAGTT-3', respectively) and a viral primer from the U3 region of the LTR from murine C-type ecotropic virus (Genbank accession number U63133; 5'-CCAGAAACTGTCTCAAGGTTCC-3'). Cycling conditions were 94 °C for 3 min, then 35 cycles of 94 °C for 30 sec, 62 °C for 1 min, 72 °C for 1 min, followed by 72 °C for 2 min.
Preparation of tissues.
Animals between 12 and 31 weeks of age were anesthetized by inhalation of isoflurane and sacrificed. Organs were either frozen in N2-cooled isopentane, or submerged in Tissue-TEK O.C.T. compound (VWR International, West
Chester, PA) and frozen on dry ice. Frozen sections (10 μM) were cut with a cryostat and affixed to slides by warming the glass beneath them by pressing a fingertip to the bottom of the slide. Mounted sections were fixed by immersion in
2% (v/v) formaldehyde and 0.2% (v/v) glutaraldehyde in modified Dulbecco’s phosphate buffered saline, (DPBS, 138 mM NaCl, 3.0 mM KCl, 1.47 mM
KH2PO4, 20 mM Na2HPO4, pH 7.4) for 10 min at 4 C. Endogenous phosphatases were inactivated by incubation in DPBS at 65 C for 1h.
Detection and enumeration of cells.
Sections were stained for PLAP+ cells by incubation for 1h at 37 C with
BCIP/NBT solution. To make this solution, solid 5-bromo-4-chloro-3-indolyl- phosphate-p-toluidine salt (BCIP) was dissolved in 100% N,N - dimethylformamide, and solid 4-nitro-blue tetrazolium chloride (NBT) was dissolved in 7 parts N,N-dimethylformamide plus 3 parts water. The NBT and
BCIP solutions were diluted in 100mM Tris-Cl, pH 9.5 to produce final
60 concentrations of 2.31 mM BCIP and 1.22 mM NBT. Nuclei were stained either
with Nuclear Fast Red (Vector Laboratories, Burlington, CA) for 30 seconds or
with 0.0015 mM propidium iodide (PI) (Molecular Probes, Eugene, OR) in 2X
SSC (300 mM NaCl, 30 mM sodium citrate, pH 7.0) for 10 seconds. Stained
sections were dehydrated with 95% (v/v) ethanol and cleared with Citrisolv
(Fisher). Cover slips were mounted with permount. Cells were counted with a
Nikon eclipse E400 microscope (Nikon Instruments, Melvine, NJ). Images were
captured with Spot Junior digital camera (Diagnostic Instruments, Sterling
Heights, MI). The PI fluorescence was visualized with filters set at 540-580 nanometers (excitation) and 600-660 nanometers (emission). The numbers of
cells and PLAP+ cells were estimated from analysis of serial sections from each
tissue. A minimum of 3 x 106 brain cells were examined for each animal. Single
sections were used to scoring clusters of cells. A PLAP+ event was defined as
BCIP/NBT staining of an area of any size.
61 Chapter 4
Impact of mismatch repair deficiency on genomic
stability in the maternal germline and during early
embryonic development
Reference
This chapter was published in the journal Mutation Research: Fundamental
mechanisms of mutagenesis. “Impact of mismatch repair deficiency on genomic stability in the maternal germline and during early embryonic development”.
November 22, 2004. Vol. 556, Issue 1-2. pp 45-53. Therefore, there is some redundancy among chapter introductions.
Purpose
A goal of this study was to determine the frequency at which mutation occurs in
the oocytes of DNA repair deficient mice and in the embryo of mice lacking DNA
mismatch repair as a maternal effect. Additionally, a goal of this study was to
characterize maximal expression of PLAP. A germline mutation in the PLAPG11
allele was optimal for this purpose.
Abstract
The effects of lack of the mismatch repair protein PMS2 on germline and
maternal-effect mutations were studied in transgenic mice that allow mutant cells
to be visualized in situ. Tg(βA-G11PLAP) mice are transgenic for the G11 allele
of a human Placental Alkaline Phosphatase (PLAP) gene driven by a human
beta-actin promoter. The G11 allele of the PLAP gene does not produce enzyme
62 due to a frameshift induced by a mononucleotide repeat containing 11 G:C
basepairs. Loss of one G:C basepair restores enzyme production. When the
G11 PLAP allele was passed through the germline of female mice lacking PMS2,
approximately 25% of the offspring that inherited the transgene exhibited the
phenotype expected for germline mutation. The mice transmitted the germline-
mutation phenotype normally and their offspring exhibited PLAP enzyme activity
in at least 30% of the cells in each tissue examined. By contrast, only 1 of 32
mice that inherited the G11 PLAP transgene from a wildtype male crossed to a
Pms2-/- female exhibited a high number of PLAP+ cells. Compared to germline
revertants, approximately one half to one quarter as many cells were PLAP+,
suggesting that a mutation occurred in one cell of an embryo containing 2 to 4
cells. These data suggest that the paternally-derived Pms2 gene provided
normal levels of PMS2 protein to embryos by the time they reached the 8-cell
stage, but that smaller embryos formed from PMS2-deificient eggs lacked PMS2
function.
Introduction
The Tg(βA-G11PLAP) transgenic mouse model allows mutant cells to be
visualized in tissue sections in situ. Tg(βA-G11PLAP) mice carry several copies of a mutant allele (G11) of a human Placental Alkaline Phosphatase (PLAP) transgene driven by the human beta-actin promoter (βA) [23, 24, 46, 60]. This allele does not produce functional enzyme due to the presence of a tract of 11
G:C basepairs that shifts translation out of the proper reading frame. Loss of one
G:C basepair from this tract restores the reading frame and enzyme activity,
63 thereby producing a cell that stains positive for PLAP activity (PLAP+) using a
histochemical assay [46].
The G11 tract in the βA-G11PLAP gene mimics mononucleotide repeats in the genome, which are present in a surprisingly high number of coding regions given their inherent instability. More than 20 human genes have been reported to have a mononucleotide repeat (of at least 8 basepairs) in a coding region and many of these have been observed to suffer mutation [9, 21, 48-55, 57, 104, 110,
111]. Surprisingly, the list of genes carrying a coding-mononucleotide repeat includes genes that perform functions critical for preventing mutation. Four of the genes encoding components of the human DNA mismatch repair (MMR) system
(Msh3, Msh6, Pms2, and Mlh3) contain mononucleotide microsatellites in their
coding sequences [9]. Many more human genes are at risk of suffering mutation
due to mononucleotide repeat instability. Searches of the ~33,000 human coding sequences have identified mononucleotide repeats with 9 or more basepairs in
365 entries [9, 53, 110].
Previous studies in G11 mice, PLAP+ cells found in tissues from Tg(βA-
G11PLAP) mice occurred at a frequency of approximately 5 x 10-5 [60]. Further studies showed that defective mismatch repair (lack of either PMS2 or MLH1) increased the frequency of PLAP+ cells dramatically, with greater than 1% of the cells in some tissues, such as brain, staining positive for PLAP activity [60]. The high numbers of PLAP+ cells in repair-deficient mice demonstrated that many cells in diverse tissues were capable of expressing PLAP. However, studies with
64 repair deficient mice can provide only a minimum estimate of the number of cells
capable of expressing PLAP.
A better approach to defining Tg(βA-G11PLAP) expression would be to
derive a germline revertant of the G11 allele. In such an animal, every cell would
carry a functional copy of the PLAP gene that is located at the same locus as in
the Tg(βA-G11PLAP) mouse. Any cells that lack PLAP activity in a germline
revertant must be due to failure to express the functional gene. It seemed
probable that a germline revertant of the βA-G11PLAP allele could be produced by passing this allele through the germline of a mouse that lacks mismatch repair because it has been shown that dinucleotide microsatellites frequently mutate in the germline of female Pms2-/- mice [112]. Nine percent of the offspring from
Pms2-/- females crossed to wild type males exhibited a novel microsatellite allele
[112]. Hence, passing the βA-G11PLAP allele through the germline of female
Pms2-/- mice (Pms2-/- males are sterile) was predicted to produce offspring
carrying a functional allele of βA-G11PLAP.
Derivation of germline revertants from the βA-G11PLAP allele was also of
interest because these experiments would provide an indication of the stability of
mononucleotide repeats in the mouse germline in the absence of PMS2. There
is a possible relationship between mononucleotide repeats in coding sequences,
modulation of mismatch repair, and rate of evolution [5, 6, 9, 113-118].
Determining the rate of mutation in such sequences in the mouse germline is
important in understanding these relationships.
65 Whereas germline reversion was the principal focus of these experiments, the Tg(βA-G11PLAP) model is also suited to studying mutation in the early embryo formed when a PMS2-deficient oocyte is fertilized by a sperm cell carrying a functional Pms2 gene. The previous experiments on dinucleotide microsatellite instability suggested that lack of PMS2 activity in oocytes caused genetic instability in early embryos that were genetically Pms2+/- [112]. These data indicated that Pms2 is a maternal-effect gene in the mouse. The fraction of cells harboring the altered sequence suggested that the maternal-effect mutations observed occurred before the 8-cell stage. Mutations may have also occurred in later embryos, but not have been detected due to the limitations of the PCR assay. The Tg(βA-G11PLAP) mouse model provides a more sensitive assay for detecting lack of PMS2 function. Although PLAP+ cells arise spontaneously in Tg(βA-G11PLAP) mice, the fraction of PLAP+ cells is generally fewer than 1 in 10,000. Therefore, if mutation were to occur in a 32-cell embryo, for example, this event would produce an animal with more than 300 times the background number of PLAP+ cells.
Herein are described the results of experiments designed to detect germline and maternal-effect mutations in Tg(βA-G11PLAP) mice. As expected, the G11 allele was very unstable in the germline of female mice lacking PMS2.
Approximately 25% of the offspring that inherited βA-G11PLAP from a Pms2-/- dam received a revertant allele of the gene. These mice transmitted this allele normally and their offspring produced PLAP enzyme activity in at least 30% of the cells in each tissue examined. By contrast, only 1 of 32 mice that inherited
66 βA-G11PLAP and a wildtype Pms2 allele from a male crossed to a Pms2-/-
female exhibited an unusually high number of PLAP+ cells. Compared to
germline revertants, approximately one half to one quarter as many cells were
PLAP+, suggesting that the mutation occurred in a cell within an embryo
containing between 2 and 4 cells. The rarity of this putative maternal-effect
phenotype among offspring, and the high number of PLAP+ cells in the animal
that exhibited the phenotype, together suggest that the paternally-derived Pms2
gene provided normal levels of PMS2 protein to embryos containing as few as 8 cells.
Results
In the absence of PMS2, the G11 allele was very unstable in the female
germline.
Three Pms2-/- females that were hemizygous for Tg(βA-G11PLAP) were
crossed to wildtype males. The tips of the tails were removed from the 65 F1 offspring and tested for PLAP activity. Seven of the tails exhibited PLAP activity.
All seven of these tails also contained the PLAP gene. Twenty-three other mice carried the PLAP gene but did exhibit PLAP activity in tail tissue. Therefore the overall frequency of putative germline revertants was 23.3% (7/30). The seven mice with the putative revertant allele came from three different dams. One dam produced three revertants; the other two produced two each. These data suggested that in each dam, approximately 25% of the eggs that received the transgene carried a revertant allele.
67 To confirm that the tail-staining phenotype was due to a genetic mutation,
the seven mice that had PLAP activity in tail tissue were each crossed to a
wildtype mate. These crosses produced 93 F2 offspring, 43 (46%) of which
inherited Tg(βA-G11PLAP). All 43 of these mice had tails that stained positive
for PLAP activity while no PLAP activity occurred in mice that did not inherit
Tg(βA-G11PLAP). Further transmission of the PLAP+ phenotype was analyzed
by crossing two female PLAP+ F2 progeny to repair-proficient males. One of the
two mice tested transmitted the PLAP gene to 4 of 10 offspring. The other mouse
transmitted to 7 of 26 offspring. Neither of these two results is statistically
different (p > 0.05) from what would be expected if the average probability of transmission is 0.5. All 11 of the F3 progeny that inherited the PLAP gene also
inherited the PLAP+ phenotype in tail tissue. These data showed that the PLAP+
tail phenotype was transmitted as a dominant Mendelian trait and that there was
close genetic linkage between the tail phenotype and the Tg(βA-G11PLAP)
locus.
Analysis of the germline revertant phenotype.
If the tail phenotype were due to a germline mutation, then PLAP activity
should be present in many if not all tissues. To test this hypothesis, PLAP
activity was examined in seven tissues (brain, heart, kidney, liver, spleen, lung
and colon) from 2 putative germline revertants (F1). All tissues stained
extensively. Some examples of PLAP-stained tissues are shown in Figure 1.
Additionally, 6 offspring (F2) of three putative germline revertants were examined.
Two F2 offspring, which were littermates, were studied most extensively. Nine
68 tissues (brain, heart, kidney, liver, spleen, lung, breast, skeletal muscle and gonad) were tested for PLAP activity. All nine tissues stained extensively in both mice.
69 [25]
70 Approximately 90% of the cells in brain and skeletal muscle were stained.
Nearly half of the cells in breast, lung, heart, kidney, spleen and gonad were
PLAP+. About 30% of the cells of the liver were stained. Nearly all cells in colon
sections were stained with exception to the colonic epithelium, which also does
not stain in mice that carry a wild type PLAP transgene (unpublished
observation). Previous studies on mice that carry a wild type PLAP transgene
demonstrated that PLAP activity was associated with what appeared to be all
cells of the brain, heart, kidney and liver [24, 60]. Therefore, the PLAP-negative cells in germline revertants cannot be attributed to tissue specific differences in the stability or activity of PLAP enzyme. The presence of PLAP-negative cells in the tissues of offspring of germline revertant mice was presumably due to repression of transcription of the transgene in some cells, a trait that is commonly exhibited by transgenes [119].
Phenotype of Pms2+/- mice formed from PMS2-negative eggs suggests a
maternal effect.
To introduce Tg(βA-G11PLAP) into eggs that lacked PMS2 protein,
Tg(βA-G11PLAP) males were crossed to Pms2-/- females. Ten mating pairs
produced 32 Pms2+/- offspring that inherited βA-G11PLAP. All of these animals
were sacrificed and sections from four tissues (brain, heart, kidney and liver)
were examined for PLAP activity. All but one animal exhibited a low frequency of
PLAP+ cells (~10-4) similar to that observed in previous studies on spontaneous
somatic cell mutation in repair-proficient Tg(βA-G11PLAP) mice [60]. However,
one mouse was clearly different and exhibited large numbers of PLAP+ cells in all
71 tissues examined. The PLAP+ cells in this animal tended to be clustered
together in large groups (Figure 2). These data fit with what would be expected
if one cell in an embryo containing 2 to 4 cells were to sustain a mutation in
Tg(βA-G11PLAP).
Figure 2. Phenotypes of hearts from maternal-effect mutation and germline reversion. Cells with PLAP activity are stained purple. Sections were counter- stained with nuclear fast red, which stains all cells red.[25]
The low frequency of the maternal-effect phenotype in the Tg(βA-
G11PLAP) model suggests that instability caused by PMS2 deficiency is confined to embryos containing fewer than 16 cells. This inference rests on the following considerations. One mouse in 32 exhibited a phenotype consistent with mutation having occurred in one of the cells present in either a two-cell or four- cell embryo. Thus, the observed frequency of mutation in the early embryo was between 1/64 and 1/128 events per cell. This frequency range conforms to that observed in previous studies on Pms2-/- mouse tissues [60]. In addition, experiments with cultured embryonic fibroblasts from Pms2-/- mice indicated that the rate of mutation was approximately 0.01 events per cell generation (data not
72 shown). Using a value of 0.01 for the probability that a given cell will sustain a mutation, the probability that PMS2 deficiency persisted in an embryo of given size can be calculated (see methods). Based on these calculations, the probability that PMS2 deficiency extended to the 8-cell stage is low (0.14), and the probability that cells in 16-cell embryos were deficient in PMS2 activity is very low (0.013) (Table 1).
Table 1. Previous data show that βA-G11PLAP mutates at 0.01 events per cell per generation in somatic cells lacking PMS2. Using this value, the probability of sustaining an event, prior to expression of paternal Pms2, was calculated.[25]
Discussion
Approximately 25% of the offspring that inherited Tg(βA-G11PLAP) from a
Pms2-/- dam received a revertant allele of the gene. This frequency of mutation is two and one half times greater than that obtained in previous studies on GT repeats containing at least 32 basepairs [60, 112]. In that report, nine percent of
the dinucleotide repeats derived from Pms2-/- females were of a novel size. The
73 higher frequency of reversion in the Tg(βA-G11PLAP) mice may have been due to the presence of multiple copies of the transgene (~5 copies per haploid
genome) [23, 24]. If all 5 of these copies, which are linked to one another, are
capable of producing transcripts, then the frequency of reversion per copy of
Tg(βA-G11PLAP was half that seen in dinucleotide repeats. In any case, the
rate of reversion of Tg(βA-G11PLAP) was very high.
At least as far as the two types of simple repeats studied so far are
concerned, loss of PMS2 function in the mouse germline raised the germline mutation rate to a very high degree. Hence, Pms2-/- mice offer a model system with which to test hypotheses concerning germline mutation rates and fitness.
The data reported herein suggest that at least one genomic protein-encoding mononucleotide repeat is mutated in every egg made in the absence of PMS2.
Approximately 300 human coding sequences contain mononucleotide repeats containing at least 9 basepairs. The density of mononucleotide repeats is similar in mouse and human, suggesting that there are hundreds of mouse genes that can be inactivated by a frameshift mutation in a mononucleotide repeat [120].
Despite the very high mutational burden imposed on the oocyte genome, Pms2-/-
dams produced normal numbers of progeny. There was no statistically significant
difference in the mean litter sizes produced by Pms2-/- dams and repair
proficient FVB/n dams. In addition, mice originating from eggs formed in the
absence of PMS2 appeared phenotypically normal and were fertile. The vigor of
these animals is presumably due to three factors. Firstly, most genes do not
contain long mononucleotide repeats. Secondly, the mutant alleles generated
74 would be expected to be due to frameshifts and therefore be recessive due to
loss of function. In cultured mouse cells, reversion of the G11 allele has been
shown to be due to deletion of a single basepair to generate a G10 allele [46].
Thirdly, recessive defective genes arising in eggs would be expected to be
complemented by functional copies derived from the paternal genome. Methods
to unmask these predicted defective genes are available [113].
Data from the Pms2-/- Tg(βA-G11PLAP) model are pertinent to discussions concerning the possible role of transient, low-efficiency mismatch
repair in evolution [5, 6, 9, 113-118]. It has been noted that four DNA mismatch
repair (MMR) proteins contain mononucleotide microsatellites in their coding sequences. These four proteins have been termed minor components of the
human MMR system (Pms2, Msh3, Msh6, and Mlh3) because MMR complexes
can still form in the absence of any one of them. By contrast, genes encoding
major components of the DNA MMR system (MSH2 and MLH1) do not contain
intrinsically unstable sequence motifs. The reason for this dichotomy is not clear,
but one possibility is that it contributes to the capacity of species to evolve rapidly
during times of stress. The postulated mechanism for transient accelerated evolution begins with the loss of a minor MMR function, which degrades, but does not destroy the germline MMR system. Theoretically, individuals with the mild mutator phenotype conferred by loss of a minor MMR component would produce more germline mutations, which would provide the variants needed to allow the lineage to survive in an environment unfavorable to the status quo [5, 6,
9, 113-118]. The data from the Tg(βA-G11PLAP) mice show that lack of PMS2
75 in the germline has a major impact on mononucleotide repeat instability. Data
from studies on somatic cells have shown that other types of mutations also
occur at higher frequency when cells lack PMS2, suggesting that germline cells
also incurred high numbers of transitions and transversions [121-127]. However,
the mice that develop from eggs formed in the absence of PMS2 did not exhibit
obvious defects. Thus, these data are consistent with the conjecture that loss of
a minor MMR component can increase germline mutation without causing a
dramatic decrease in fitness. Similarly, Pms2-/- Tg(βA-G11PLAP) mice exhibit
millions of mutant cells in all somatic tissues, yet develop normally [60].
However, the lack of PMS2 in all somatic cells is demonstrably detrimental to the
individual because Pms2-/- mice are predisposed to develop neoplasias as they
age [126].
In the experiments designed to detect a maternal-effect, only 1 of 32 mice that inherited the PLAP gene from a repair proficient male crossed to a Pms2-/-
female exhibited an unusually high number of PLAP+ cells. This animal had
thousands of times more PLAP+ cells than a typical repair-proficient mouse, but
the number of PLAP+ cells was much less than that seen in the germline
revertants. These data suggest that a mutation occurred not in the male
germline, but rather in the embryo, when this animal consisted of only a few cells.
If so, then this animal is an example of a maternal-effect conferred by absence of
PMS2 in the egg. While it seems less likely than a maternal-effect, it is not
possible to exclude the possibility that the putative maternal-effect phenotype
was in fact caused by a germline mutation in the male parent. Of course, in this
76 case, all cells in the offspring would carry the revertant gene, so one would need
to invoke poor expression of this gene to explain the lower number of PLAP cells
observed in this animal. Nevertheless, such a scenario cannot be excluded.
However, the lack of certainty regarding the origin of this mouse poses no
problem for assessing the embryonic stage at which PMS2 derived from a sperm gene becomes functional because this assessment is based on the lack of a maternal effect. It is clear that either 31 of 32 (97%) mice or 31 of 31 (100%) mice did not exhibit this effect. Either outcome would be improbable unless
PMS2 deficiency ends prior to the 16 cell stage.
Our findings regarding PMS2 expression during development are in agreement with those obtained in the study of dinucleotide repeat variation [112].
In these studies, 8 PMS2+/- animals derived from PMS2-deficient eggs exhibited
the following phenotype. The PCR produced three microsatellite bands instead
of the expected two. One band corresponded to that seen in the female parent
and was of the expected intensity. Another band was the same size as that in
the male parent, but was less intense than expected. The third band was of
novel size and relatively faint. These data suggested that these animals were
mosaics caused by mutations to paternal microsatellites during early embryonic
development. A similar phenotype occurred in 4 other animals, but in these
cases it appeared that maternal alleles were mutated in a fraction of the cells. Of
the 816 microsatellites that were amplified in this study, 12 (1.5%) exhibited a
maternal-effect mutation.
77 The frequency of the maternal-effect phenotype in Tg(βA-G11PLAP) mice was similar to that observed in the microsatellite studies [112]. There were approximately five copies of the Tg(βA-G11PLAP) transgene in each of the 32 mice. Hence, 160 potentially scorable mononucleotide repeats were present among the mice analyzed. One mouse exhibited the phenotype expected of a maternal effect. Hence, the frequency was 0.00625. While this frequency is
lower than that observed in the dinucleotide repeat study (where it was 0.015),
the difference in the two frequencies is not statistically significant.
The maternal-effect data contrast with results obtained in mice that lack
PMS2 in all cells throughout development [60]. All such mice exhibited tens of
millions of PLAP+ cells. By contrast, only one in 32 Pms2+/- mice formed from a
PMS2-negative oocyte had a very high number of PLAP+ cells. These data
indicate that the high number of PLAP+ cells seen in previous studies of somatic
tissues from Pms2-/- animals was not primarily due to amplification of a few
mutant cells formed in the very early embryo [60].
Tg(βA-G11PLAP) mice offer a unique means to detect mutation in vivo.
Because mutant cells are detected in situ, theoretically, mutation can be detected in any cell type. However, in practice, the in situ phenotypic assay can be limited because some cells may not express the transgene. Germline revertant
Tg(βA-G11PLAP) mice show that expression of the transgene was widespread, but not universal. This information can be used to aid in interpreting results obtained with Tg(βA-G11PLAP) mice. An improved model can be made by moving the transgene to a locus that supports universal expression. Mice
78 carrying a PLAP gene at the ROSA 26 locus appeared to express PLAP activity
in all cells [127].
Materials and Methods
Transgenic Mice.
Transgenic FVB/N mice carrying the βA-G11PLAP frameshift reporter gene were
generated as described [23, 128] . Mice containing a targeted disruption in Pms2
were a generous gift from R. M. Liskay [121] The Pms2-knockout animals
originally obtained were in a C57BL/6 background. Because Tg(βA-G11PLAP) was in the FVB/N strain, the Pms2-knockout allele was transferred to this
background by four backcrosses to yield mice carrying the disrupted Pms2 allele in a genetic background that was ~97% FVB/N. Mice were analyzed for the presence of Tg(βA-G11PLAP) and for the Pms2 genotype by PCR of tail DNA as described [60, 121, 129]. Mice heterozygous for the Pms2 null mutation were mated with Tg(βA-G11PLAP) mice to produce offspring, some of which were heterozygous for the Pms2 null mutation and hemizygous for βA-G11PLAP.
Matings between these mice produced female Pms2-/- mice, some of which had
Tg(βA-G11PLAP) and some which did not. To obtain germline revertants, three
Pms2-/- females that carried Tg(βA-G11PLAP) were crossed to wildtype males.
To test for a maternal effect, ten Pms2-/- females that lacked Tg(βA-G11PLAP) were mated to MMR proficient males that carried Tg(βA-G11PLAP). Mating strategies are illustrated in Figure 3.
79
Figure 3. Mating strategies used to generate germline revertant and maternal- effect mice. A. To examine mutation in the maternal germline, Pms2-/- females that carried Tg(βA-G11PLAP) were crossed to wildtype males. Twenty five percent of progeny that inherited the PLAP gene exhibited a high number of PLAP+ cells in tissues. B. Mutation in the early embryo was examined by passing Tg(βA-G11PLAP) from the paternal genome into the PMS2-deficient oocyte. Upon fertilization, the βA-G11PLAP allele is susceptible to mutation until paternal PMS2 is expressed. Of thirty-two mice that inherited the PLAP gene, one exhibited a high number of PLAP+ cells in tissues. [25]
Assay for germline revertants.
The tips of the tails of offspring were removed and tested for PLAP activity as follows: tail clippings were submersed in phosphate buffered saline (0.15 M
NaCl/ 2.7 mM KCl/ 1.47 mM KH2PO4/ 4.86 mM Na2HPO4, pH 7.4) (PBS) and heated to 65°C for 1 hour to inactivate endogenous murine phosphatases.
Heated tail tissue was placed in BCIP/NBT solution, which is 0.1M Tris, pH 10 containing 5-Bromo-4-Chloro-3-Indolyl Phosphate (Fisher Scientific, Pittsburgh)
After 8 hours at 4°C, tubes containing tail tips were examined for the presence of the purple precipitate produced by PLAP (stained ndidates for germline
80 reversion, Figure 4). After staining with BCIP/NBT, the tail sections were
removed from the staining solution and processed to extract DNA for analysis by
PCR as described [129].
Figure 4. Tails from germline revertant mice stain purple. Tail A.is from a mouse that harbors the G11PLAP allele. Tails B & C are from mice that inherited mutated G11 PLAP alleles. Tail D is from a mouse that has no G11 PLAP gene. Tail E is a positive control tail expressing the wildtype form of the PLAP transgene.
Test for a maternal effect
To test for mutation in the early embryo, Pms2-/- females were crossed to male mice (4 wildtype and 6 Pms2+/-) that carried Tg(βA-G11PLAP). Ten mating pairs produced 19 litters and 121 offspring. Tail DNA PCR showed that
32 of the 121 mice were Pms2+/- and inherited βA-G11PLAP. These 32 animals were sacrificed and tissues analyzed for PLAP activity. Mice were between 4 and 20 weeks of age at the time of sacrifice.
Analysis of PLAP expression in frozen sections.
Organs were removed and either frozen in OCT embedding compound on dry ice or snap frozen in liquid N2-cooled isopentane. Cryosections were cut 10
81 um thick, mounted on slides, and fixed in 2% formaldehyde (v/v)/ 0.2%
glutaraldehyde (v/v) in PBS for 10 minutes at room temperature. Endogenous
phosphatases in fixed tissues were inactivated by incubation of mounted sections
at 65oC for one hour. Tissues were stained for PLAP activity as follows: Sections
were incubated in BCIP/NBT pH 10 for 30 - 90 minutes at 37°C and counter-
stained with Nuclear Fast Red (Vector Labs, Burlingame). Images were captured
and combined with a Spot Jr digital camera (Diagnostic Instruments).
Probability analysis.
The relationship between the embryonic stage reached in the absence of
PMS2 function and the probability of observing 1 maternal-effect phenotype
among 32 animals was determined as follows: The probability of mutation in a
cell that lacked PMS2 function was set at 0.01. This value was chosen because
previous studies indicated that the rate of mutation of βA-G11PLAP in mouse somatic cells lacking PMS2 is approximately 0.01 events per cell generation [60].
The probabilities of observing 1 event among n cells at risk of sustaining an event were calculated using the Poisson probability function:
T -np P(1 / n) = (e )(np)
Where n is the number of cells at risk of mutation. In the calculations, the value
of n ranged from 32 to 512, which corresponds to the number of cells in 32
embryos containing between 1 and 16 cells, respectively.
82 Chapter 5
Expression and Loss of Alleles in Cultured Mouse
Embryonic Fibroblasts and Stem Cells Carrying Allelic
Fluorescent Protein Genes
Reference
This chapter has been submitted as an article to the BioMed Central Journal of
Molecular Cell Biology. September, 2006.
Purpose
To develop and characterize a novel reporter system which facilitates studying a
broad spectrum of genetic aberrations, which result in loss of gene expression.
This fluorescent based system will report events at corresponding loci on
homologous chromosomes. The ultimate goal is to view said genetic events in
situ within the whole animal.
Abstract
Loss of heterozygosity (LOH) contributes to many cancers, but the rate at
which these events occur in normal cells of the body is not clear. LOH would be detectable in diverse cell types in the body if this event were to confer an obvious cellular phenotype. Mice that carry two different fluorescent protein genes as alleles of a locus would seem to be a useful tool for addressing this issue because LOH would change a cell’s phenotype from bichrome to monochrome.
In addition, LOH caused by mitotic crossing over might be discernable in tissues because this event produces a pair of neighboring monochrome cells that are
83 different colors. As a step in assessing the utility of this approach, we derived
MEF and ES cell lines from mice that carried two different fluorescent protein
(cyan and yellow) genes as alleles at the chromosome 6 locus, ROSA26. FACS showed that the vast majority of cells in each line expressed the two marker proteins at similar levels, but populations exhibited the extrinsic and intrinsic
expression noise similar to bacteria and yeast expressing allelic markers. In
addition, cells with a monochrome phenotype were present at frequencies on the
order of 10-4. MEFs did not proliferate as clones, precluding further analysis, but
45 of 45 stably monochromatic ES cell clones exhibited loss of one allele at the
ROSA26 locus. Approximately half of these clones retained heterozygosity at a
locus between ROSA26 and the centromere, suggesting that mitotic
recombination was a cause of ROSA26 LOH. The remaining clones showed
LOH near the centromere, but were disomic for chromosome 6, suggesting that
uniparental disomy was also a cause of ROSA26 LOH.
Introduction
During malignant progression, cells accumulate multiple genetic and
epigenetic alterations that cause loss of at least one anti-oncogenic function.
Such a loss can be caused by a variety of events including mutation and losses
that take place at the chromosome level, e.g. loss of heterozygosity (LOH), which
is a hallmark of numerous cancers [2, 8, 130-132]. Many cases of LOH are
caused by mitotic recombination (MR) between homologous chromosomes [133].
LOH can also arise via uniparental disomy (UPD), a change that presumably
begins with nondisjunction of sister chromatids, producing trisomy in a daughter
84 cell. Subsequent mis-segregation during mitosis of a trisomic cell can produce a disomic cell where both homologues were derived from the same parental homologue (UPD) [134]. On other occasions, gene conversion (GC) and interstitial deletions cause LOH [135, 136]. In addition, it has recently come to light that some cells in the brain can be monosomic for one or more chromosomes [137].
Tumors serve as indicators of allele loss, but not all allele loss events necessarily
lead to a tumor. LOH in diverse, nontransformed cell types in the body would be
directly detectable if this event were to confer an obvious phenotype (other than
tumorous growth) on an individual cell, its sibling, and their progeny. Mice that
carry two different fluorescent protein genes as alleles of a locus would seem to
be a useful tool for addressing this issue because LOH would change a cell’s
phenotype from dichromatic to monochromatic. If tissue architecture permits, the
cause of LOH would be suggested by the number and arrangement of mutant cells because LOH caused by mitotic crossing over produces a pair of neighboring monochromatic cells expressing different colors [138]. By contrast,
LOH caused by other events, such as UPD, gene conversion, or point mutation would be expected to produce a single monochromatic cell.
As a first step in assessing the utility of the allelic marker approach in mammals, we derived cell lines from mice that carried two different fluorescent protein (cyan and yellow) genes as alleles at the widely expressed ROSA26 locus, which is on
chromosome 6 [139]. Although our studies were primarily motivated by an
interest in LOH in mouse tissues, studies on genetic stability and allelic gene
85 expression in mouse ES cells are of interest in their own right. The totipotent
nature of ES cells has made them a useful tool for manipulating the genome and
a promising prospect for human therapeutic applications. However, introduction of genetically damaged ES cells could lead to adverse outcomes.
The genetic and karyotypic stability of ES cells in general is not entirely clear.
On one hand, aneuploid mouse ES cell lines are fairly common [140, 141]. On
the other hand, hundreds of mice have been made from ES cells, showing that
these cells can maintain genetic stability when handled properly [142]. Some
studies have suggested high rates of allele loss in ES cells [132, 143], while in
others, loss rates were hundreds of fold lower [40]. Different rates of point
mutation have also been reported for ES cells [40, 61]. The reasons for these
different observations are not clear, but could include differences in marker
genes employed, methods used to detect variant cells, cell lines studied, rates at which different chromosomes undergo either nondisjunction or mitotic recombination, and inadvertent selection of cells that proliferate better in culture.
The bichromatic biallelic ES cells described herein differ from others studied
because cells with variant phenotypes can be identified and isolated by FACS.
86 Results
Fluorescence phenotypes of cells in populations of biallelic embryonic
fibroblasts
The two lines of transgenic mice used to make embryonic cell lines were a
gift from F. Costantini, whose work had shown that both fluorescent proteins
were simultaneously widely expressed in mice, which appeared normal [139].
Three R26CY mouse embryonic fibroblast (MEF) populations were
independently derived, each from one 13.5 day post coitus (dpc) embryo. These
cells were cultured for a few passages and then subjected to FACS analysis.
Nearly all of the cells in each population exhibited both CFP and YFP
fluorescence (bright cells) (Figures 1A-C). However, CFP signal intensities
tended to be lower than YFP intensities, which was expected because CFP is
intrinsically less bright [144]. A few percent of the cells exhibited little if any
fluorescence of either color (dim cells). The nature of these cells was not
investigated. Intact embryos appeared to express both fluorescent proteins
uniformly and ubiquitously (Figure 2). Nevertheless, embryos could have contained a small number of cells that fail to express either fluorescent protein.
Figure 3A shows a scatter plot of YFP and CFP fluorescence intensities in 1000
individual MEFs drawn at random from the “bright” population shown in Figure 1.
The points in this figure were plotted using CFP signal intensities that were
normalized to correct for the inherent faintness of this protein. Different cells
exhibited different levels of fluorescence and the fluorescence intensities varied
over a 5 fold range on both axes. Such variation has been termed extrinsic
87 expression noise [20,21]. However, in a given cell, YFP and normalized CFP signal intensities tended to be approximately equal. Coordinate variation in allelic expression has been observed in bacteria and yeast and is expected because the two alleles are exposed to the same intranuclear environment [145,
146]. Nevertheless, lack of complete coordination of expression within a cell, a phenomenon known as intrinsic expression noise, is also expected, based on theory and on observations in E. coli and yeast [145, 146]. MEFs that were
brighter with respect to one color or the other can be seen as off-diagonal points
in the scatter plot shown in Figures 1 and 3A. The intrinsic noise level exhibited by MEFs was approximately 0.2, a value similar to that reported for weakly transcribed loci in E. coli [145]. It was difficult to compare the MEF intrinsic noise level to those reported for yeast because yeast noise levels were measured for several different promoters under a variety of induction-repression conditions and reported in arbitrary units [146]. However, comparison of scatter plot shapes (i.e. the distribution of points relative to a diagonal line, compared to the range of fluorescence signals in the population) suggested that MEF and yeast intrinsic noise levels were generally comparable.
A few cells in each MEF population exhibited more than a 5 fold difference
in CFP and YFP fluorescence and were suspected of being monochromatic
(Figure 1). The number of cells expressing YFP only was approximately the same as the number of cells expressing CFP only (Figure 4). Apparent
monochromatic cells of both colors were selected by sorting and placed in
culture, but did not survive, precluding further phenotypic and genetic analysis.
88 The reason for the failure of these cells to form clonal colonies was not
investigated, but factors that might have caused this result include the low plating
efficiency of single MEFs and the fact that the populations that were sorted had
been passed several times prior to sorting.
Figure 1. Raw FACS dot plots (106 events) showing fluorescent profiles of three R26CY MEF cell lines (A-C) and two ES cell lines (D, E). Arrows in panel A indicate “dim” and “bright” events in MEF plots. Circles in panels D and E indicate areas where monochromatic variants would be expected to be located.
89
90 Fluorescence phenotypes of cells in populations of biallelic ES cells
Two R26CY ES cell lines were independently derived and studied.
Figures 1D and 1E show FACS data obtained from these two cell lines. In addition to the points produced by ES cells, the FACS plots contained a relatively small number of points produced by autofluorescence of the feeder cells present in the ES cell cultures.
The fluorescent phenotypes in the two ES cell lines were similar to those seen in the three R26CY MEF cell lines (described above) except that the fluorescence emitted by each protein in a typical ES cell was about a third as intense as that seen in a typical R26CY MEF cell. Nearly all of the ES cells in each population exhibited both CFP and YFP fluorescence (Figures 3B). In a given cell, normalized CFP and YFP signal intensities tended to be approximately equal, but coordination of expression was not perfect. The two ES cell lines, R26CY2, exhibited intrinsic noise levels of 0.2, similar to that exhibited by R26CY MEFs.
Noise is expected to produce cells that exhibit CFP and YFP fluorescence intensities that differ over time. To determine if this were the case, a population in which the cells were two fold brighter with respect to CFP than YFP was obtained by FACS, placed in culture and passed 3 times. Analysis by FACS showed that the original phenotype (brighter CFP) was not maintained. Instead the population resembled those shown in Figure 3. Hence, the phenotype of the population was transient, as would be expected if it were due to noise.
91 Both R26CY lines contained rare cells that appeared to be monochromatic
(Figure 1). These cells occurred at a frequency of approximately 10-4 and apparent CFP and YFP monochromatic variants occurred in roughly equal numbers in both ES cell populations (Figure 4). To determine if they were stable variants, apparent monochromatic cells were gated into collection tubes, plated at low density and cloned. Examination of 85 clonal cultures by fluorescent microscopy showed that about 80% of them exhibited the expected monochromatic phenotype (Figure 5). The remaining 20% were dichromatic, indicating that the sorting method produced populations of cells that were highly enriched for monochromatic cells, but that some dichromatic cells passed through the gates. Subsequent experiments using mixtures of monochromatic and dichromatic ES cells showed that using more narrow gates reduced the recovery of true monochromatic cells (data not shown). Therefore, quantification of monochromatic cells was more accurate when gates were set wide and dichromatic cells that were misidentified by FACS were later detected by microscopy.
92
Figure 3. Variation in fluorescent signals (noise) from 1000 randomly selected cells. (A) MEF cell line 33e1 (B) ES cell line R26CY1. The fluorescent units are relative values assigned by the sorter. Intensities for CFP have been normalized to correct for its relative faintness when compared to YFP intensity. Arrows at right angles indicate noise axes, extrinsic (E) and intrinsic (I).
Spontaneously arising stable monochromatic ES cells lacked the gene encoding the non-expressed fluorescent protein.
To determine if gene loss contributed to the production of monochromatic
ES cells, PCR was used to amplify fluorescent protein genes and amplicons were analyzed by digestion with the restriction endonuclease Pst1 because the
YFP gene has a Pst1 cleavage site that the CFP gene lacks (Figure 6). All spontaneously occurring monochromatic ES cell clones analyzed by restriction enzyme analysis (n=45) lacked the gene encoding the absent fluorescent protein.
To confirm these results, PCR products from 7 monochromatic ES cell clones (3 expressing only YFP and 4 expressing only CFP) were cloned and sequenced.
At least 5 cloned amplicon copies from each of the 7 monochromatic ES cell clones were sequenced. All of the sequences from a given monochromatic cell
93 line were identical to the gene encoding the fluorescent protein observed in that
cell line. As a control, PCR products from a dichromatic ES cell population were
cloned and sequenced. Sequences from CFP and YFP genes were both present
and equally abundant, as expected. Three additional monochromatic clones
were analyzed by Southern blot hybridization, which showed that each had lost
the gene encoding the absent fluorescent protein (data not shown).
Figure 4. Abundance of spontaneously occurring apparent monochromatic variants. Solid and hatched bars indicate frequencies of cells exhibiting only CFP or only YFP, respectively. R26CY1 and R26CY2 are data from the two ES cells lines. 21e2MEF, 33e1MEF and 37e3MEF are data from the three MEF lines. Error bars indicate the standard errors of the means.
Analysis of a centromeric marker in cells with LOH at ROSA26.
To investigate the nature of the events that produced allele loss at the
ROSA26 locus, a centromeric heterozygous microsatellite (D6Mit159) was
identified. The D6Mit159 locus is 30 Mbp from the centromere and 83 Mbp from
the ROSA26 locus. Heterozygosity was retained at the D6Mit159 locus in 12 of
94 20 spontaneous monochromatic clones examined, suggesting that 60% of
spontaneous monochromatic clones were produced by mitotic crossovers within
the 83 Mbp interval between the D6Mit159 microsatellite marker and ROSA26. In
the other 8 monochromatic clones, heterozygosity was lost at the D6Mit159
locus. Such a result was consistent with loss of one homologue of chromosome
6, although LOH at both D6Mit159 and ROSA26 might have been caused by mitotic recombination taking place in the 30 Mb interval between the centromere and D6Mit159.
Table 1. Mutations produced by EMS in monochromatic ES cell clones. Monochromatic Phenotype Gene mutated Mutation Predicted protein alteration
clone
R26CY 1025 C-8 cyan YFP 1 bp deletion Truncated after 4th amino acid
R26CY 1013 Y-1 yellow CFP G to A G94D
R26CY 1025 Y-1 yellow CFP G to T W60C
The possibility of chromosome 6 monosomy was tested by whole chromosome
painting of metaphase chromosomes in two of the 8 monochromatic clones that
were homozygous at D6Mit159. All metaphase spreads examined were disomic
for chromosome 6 (Figure 7). These data, along with the fact that autosomal
monosomy has not been described in mouse ES cell lines, suggested that
chromosome loss without re-duplication was not a major contributing mechanism
of allele loss.
95
Figure 5. Phenotypic analysis of FACS-isolated ES cell clones. Panels A-C show images of three clones with different phenotypes. (A) colony of cells expressing CFP only, (B) colony of cells expressing YFP only, (C) colony of cells expressing both CFP and YFP. Images shown were are produced by merging three images, CFP epifluorescence, YFP epifluorescence, and phase contrast. Gray cells under the ES clones are a monolayer of non-fluorescent wild type MEF feeder cells. Arrows indicate some of the cellular debris and dead cells that exhibited autofluorescence. Panels D, E and F correlate with panesl A, B and C, respectively, and show FACS analysis of each subclone. Polygons indicate gates used to collect putative monochrome cells. Arrows indicate residual MEF feeder cells present in ES cell populations subjected to FACS.
Copy number of chromosome 6 in parental ES cell lines.
UPD would explain the monochromatic cells that were disomic for
chromosome 6 yet had LOH near the centromere. Development of UPD would
be facilitated by trisomy for chromosome 6 in the parental cell lines. Therefore, it
was of interest to determine if either of the parental R26CY ES cell lines had this
96 karyotype. To that end, 66 cells in cell line R26CY1 were subjected to spectral
karyotyping (Figure 8), and 42 cells in cell line R26CY2 were analyzed by whole
chromosome painting. All cells analyzed contained 2 copies of chromosome 6.
These data established with 95% confidence that the fraction of cells with trisomy
6 was 5% or less in cell line R26CY1 and 7% or less in cell line R26CY2.
Frequency of monochromatic ES cells and estimated rate of LOH
The original FACS experiments showed that monochromatic variants occurred at frequencies of 3.1 x 10-4 and 2.6 x 10-4 in the R26CY1 and R26CY2
ES cell lines, respectively. These data were acquired from populations of cells
that had been derived from blastocysts and kept in culture for several months.
Thus, it was possible that the frequencies of monochromatic variants were
inflated by their accumulation over time. To examine the relationship between
frequencies of monochromatic cells and the rate at which they arise, preexisting
monochromatic cells were removed from populations of R26CY cells by FACS.
These dichromatic cell populations were expanded for 10 population doublings in
culture, and then subjected to FACS analysis. Monochromatic cells were present
on the order of 10-4 in both cell lines. Monochromatic cells expressing only blue
were approximately as frequent as those expressing only yellow.
Variants can be either more of less frequent than otherwise dictated by their rate of formation if they proliferate more or less rapidly than parental cells. To determine if monochromatic cells might proliferate more rapidly than their
dichromatic parents, growth kinetics of clonal populations of different fluorescent
phenotypes were studied. The different clonally derived populations exhibited a
97 variety of growth rates, but fluorescent phenotype and growth rate were not correlated (data not shown). In addition, an increase in proliferation rate upon loss of one fluorescent protein seems an improbable scenario for two reasons.
First, monochromatic cells did not exhibit less fluorescent signal than dichromatic cells, suggesting that the amount of fluorescent protein in monochromatic cells was not less than in dichromatic cells. Second, although it has been reported that it is possible to cause ill effects by over expressing GFP [147], there is little evidence of general toxicity associated with fluorescent proteins, which have been used in many ES cell lines [148-152] and in transgenic mice, which are often generated from ES cells expressing one fluorescent protein or another
[139, 149-154]. In the case of mice carrying CFP and YFP genes at ROSA26, the animals are viable and reproduce normally, as do mice homozygous for either CFP or YFP at ROSA26 [12].
These data suggested that the frequencies of monochromatic cells present in populations of R26CY cells reflected the rate of their production and that the two types of monochromatic variants arose at essentially the same rate, which can be estimated from the relationship between frequency of variants and the number of reproductive cycles (generations) the population has undergone, where rate equals the proportion of variants in a final culture divided by the number of generations that have elapsed [71, 155]. By this calculation, the rate of LOH in both cell lines was approximately 10-5 per cell-generation.
98 Induction of mutation in bichromatic ES cells.
Because all 45 spontaneous monochromatic ES cell clones examined exhibited
LOH at the ROSA26 locus, it was of interest to determine if LOH were the only pathway capable of producing this phenotype. Therefore, R26CY1 ES cells were treated with ethylmethanesulfonate (EMS), which is a strong inducer of point mutations. The cell population exposed to EMS exhibited 1.7 fold more monochromatic cells as assessed by FACS followed by microscopy.
Twenty-five monochromatic clones isolated from EMS-treated R26CY1 ES cells
populations were subjected to DNA analysis by PCR followed by Pst1 digestion.
Five of the 25 clones retained both the CFP and YFP genes, suggesting that a
point mutation had caused loss of expression of one fluorescent protein gene in
20% of the monochromatic cells produced following treatment with EMS. To test
this hypothesis and determine the nature of these mutations, 3 of the clones that
retained both fluorescent protein genes were subjected to sequence analysis. As expected, all 3 monochromatic clones harbored a mutant version of the non-
expressed fluorescent protein gene, and the mutation predicted an alteration in the encoded protein sequence (Table 1). The mutation found in clone R26CY
1025 CFP-8, which expressed CFP but not YFP, explained the lack of YFP fluorescence because the YFP gene contained a frameshift mutation predicted to completely block production of the YFP peptide by stopping translation at codon
4. The mutations observed in the other two biallelic but monochromatic ES cell clones altered the amino acid sequence and were predicted to change the
99 protein in ways that could extinguish fluorescence, as they either resided near
the sequence encoding the fluorophore or truncated the protein.
Figure 6. Genotyping the ROSA26 locus in DNA isolated from monochrome clones. At left are shown maps of the two alleles in parental cells. The YFP gene has an additional cleavage site for PstI restriction endonuclease. Lane 1, a clone lacking the YFP gene. Lane 2, a bichrome clone, which retained both the genes, as expected. Lane 3, a clone lacking the CFP gene. Lanes 4 & 5, data from DNA of unsorted ES cell lines. Lane 6, PCR control containing no DNA in the reaction mix. Markers are 1 Kb DNA ladder (Invitrogen).
Figure 7. Whole chromosome-6-paint FISH analysis of monochromatic ES cell clones that had lost heterozygosity at both the ROSA26 locus and D6Mit159. Metaphase chromosomes were hybridized to a chromosome 6 probe labeled with FITC. Chromosomes were counter stained red with propidium iodide (PI).
100 Images shown were produced by merging FITC and PI epifluorescence images. Copies of chromosome 6 are yellow. Other chromosomes are red. Panels A and B show metaphase chromosomes from cells that expressed CFP only or YFP only, respectively. 400x magnification.
Figure 8. A representative image of a metaphase chromosome spread from a R26CY1 ES cell. Panel A shows R26CY1 ES cell chromosomes following simultaneous hybridization using combinatorial labeled chromosome painting probes. Panel B shows DAPI staining of the same spread for comparative analysis.
Discussion Populations of biallelic dichromatic mouse embryo cells contained numerous cells in which fluorescence of CFP and YFP were not equivalent. In about one in ten thousand cells, CFP and YFP fluorescence were highly disproportionate. This phenotype arose by the spontaneous loss of the gene encoding either CFP or YFP. The number of cells with this type of LOH could be determined by a combination of FACS followed by microscopic observation of populations originating from individual sorted cells.
While rare mutant cells exhibited highly disproportionate fluorescence, similar but nonequivalent CFP and YFP fluorescence was seen in most of the
101 cells in biallelic dichromatic populations. These variations in relative
fluorescence were transient and presumably produced by expression noise.
Genetic instability in mouse ES cells
The rate of LOH inferred from the frequency of monochromatic ES cells is similar to rates reported in most other studies on LOH in mouse ES cells. LOH rates at 11 different mapped loci each carrying an inserted neo gene have been reported to range between 10-3 and 10-5 events per cell generation [132, 143,
156]. Experiments on cells carrying neo genes inserted at unknown loci
produced similar results [157]. Studies using other markers also reported rates
of LOH within the range suggested by the frequency of monochromatic cells [4,
158]. In contrast to the rates reported in most studies, experiments using ES
cells heterozygous at the Aprt locus indicated that spontaneous LOH occurred at
a rate of approximately 1x10-7 events per cell generation [40]. It is not clear why
the rate of allele loss at the Aprt locus differed from those seen at other loci, but it
is possible that chromosome 8 behaves differently from other chromosomes in
this respect.
In mouse ES cells heterozygous at the Aprt locus, about 40% of LOH events had occurred via mitotic recombination and UPD was the most common genetic change associated with LOH at Aprt [40]. Similar results were reported for the FasI locus, where a third of the cells with LOH were produced by mitotic
recombination [156]. Lefevbre et al showed that LOH with respect to integrated
neo genes was accompanied by LOH at linked markers, but did not attempt to
102 distinguish between mitotic recombination and UPD as the cause of these events
[132].
Our studies on biallelic dichromatic mouse ES cells suggest that LOH at
ROSA26 occurred principally by mitotic recombination occurring between
ROSA26 and the DMit159 marker that is 30 cM from the centromere. However, because a heterozygous locus telomeric to ROSA26 could not be found, it is not possible to exclude interstitial deletion encompassing the ROSA26 locus but too small to produce an obvious decrease in the size of chromosome 6.
Nevertheless, interstitial deletion seems an unlikely contributor because such events rarely generate spontaneous LOH in ES cells [133, 159-162].
About 40% of the cells had LOH at both D6Mit159 and ROSA26. This genotype might have been caused by mitotic recombination taking place in the 30 Mb interval between the centromere and D6Mit159. However, the frequency of monochromatic clones with LOH at both D6Mit159 and ROSA26 was two fold higher than would be expected to be produced solely by mitotic recombination, assuming these events occur in proportion to the distance between markers. We would expect recombination in the 30 Mbp interval between the centromere and the D6Mit159 locus to occur 36% (30/83) as frequently as recombination in the
83 Mbp interval between the D6Mit159 locus, where recombination produced
12/20 (60%) LOH events. Therefore, only 4 of the 8 clones with LOH at
D6Mit159 would seem to be attributable to mitotic recombination. Mechanisms other than mitotic recombination that might have contributed to the development of LOH at both loci include UPD and monosomy. UPD seems more likely
103 because monosomy is very rarely seen in mammalian cells and has not been
seen at all in mouse ES cells [159, 162].
Potential of biallelic fluorescent markers for studies on genetic instability
Findings obtained in these studies on biallelic dichromatic ES cells
suggest that this approach can be extended to tissues isolated from mice. GFP has been shown to be useful for detecting chromosome loss in Hela cells and in mouse brain [7, 137]. In a similar fashion, monochromatic cells in tissues from dichromatic mice can be identified and isolated by FACS. In addition, it may be possible to detect monochromatic cells in situ in tissue sections. While other approaches provide data on mutation and mis-segregation events in tissues, only the dichromatic model can provide information about the locations of variant cells within tissues, and this information has the potential to reveal mitotic recombination events because such events can generate twin spots composed of neighboring patches of monochromatic cells of different colors descended from the monochromatic daughter cells produced from a mitotic cell that has undergone crossing over between homologous chromosomes [138].
Sources and features of expression noise
Mouse cells exhibited both extrinsic and intrinsic noise, the two types of
variation seen in expression of allelic genes in yeast and bacteria [145, 146].
Extrinsic noise refers to variation in the fluorescence intensity emitted by a given
fluorescent protein in different cells in the population [145, 146], and is thought to
be due to variation among cells with respect to parameters such as position in
104 the cell cycle. ES cells and MEFs exhibited similar levels of extrinsic noise,
which may seem surprising given that MEF populations are derived from the
numerous cell types present in a 13.5 dpc embryo, while ES cell populations
contain a single cell type. However, many of the diverse cell types present in a dissociated embryo appear to fail to proliferate, and cells resembling fibroblasts quickly predominate in MEF cultures. Another factor that may work to minimize extrinsic noise caused by heterogeneity with respect to cell types is the robust activity of the ROSA26 promoter, which is driving transcription of the fluorescent protein genes. This promoter is known to function in a wide array of cell types
[163-172].
Intrinsic noise refers to the discordance in the intensities of the two
fluorescent proteins within a single cell, and is thought to be due to lack of
coordination with respect to processes such as assembly of transcription
complexes at allelic promoters [145, 146]. This lack of coordination is thought to
result from stochastic variation caused by a scarcity of factors needed to
accomplish gene expression. Intrinsic noise occurred in both types of mouse
cells, and to similar extents. The level of intrinsic noise exhibited by mouse cells
resembled that seen in E. coli when the promoter driving fluorescent gene
transcription was semi-repressed by the lac repressor protein [145]. Intrinsic
noise in E. coli was reduced when repression was lifted and transcription rate
was increased, leading to the suggestion that intrinsic noise is inversely related
to transcription rate [145]. However, studies in yeast showed that the
relationship between intrinsic noise and transcription can be more complex [146].
105 Conclusions LOH at ROSA26 produced monochromatic cells in populations of
dichromatic mouse embryonic cells and was usually accompanied by retention of
heterozygosity at a locus between ROSA26 and the centromere, suggesting that
mitotic recombination was the major cause of ROSA26 LOH. Dichromatic mouse
embryonic cells exhibited expression noise, a phenomenon previously described
in bacteria and yeast carrying different fluorescent protein genes as allelic
markers. Dichromatic mouse embryonic cells provide a novel system for
studying genetic/karyotypic stability and factors influencing expression from
allelic genes in cultured ES cells, and suggest that similar approaches will allow
these phenomena to be studied in tissues.
Methods
Mice
The two lines of transgenic mice used to make embryonic cell lines were a
gift from F. Costantini [139]. One transgenic mouse line (R26R-EYFP) carried a
gene encoding enhanced YFP at the ROSA26 locus. The other mouse line
(R26R-ECFP) carried a gene encoding enhanced CFP at the ROSA26 locus.
We crossed the two lines to produce mice with different fluorescent protein
markers at the ROSA26 locus. These mice were of mixed genetic background
that included alleles from three inbred strains, 129X1/SvJ, C57BL/6J and FVB/n.
This situation was due to the following history of mouse production and maintenance. Strain 129X1/SvJ was the background into which the fluorescent
protein genes were originally integrated into the mouse genome by gene
106 targeting [139]. The targeted 129X1/SvJ ES cells were injected into C57BL/6J
blastocysts and chimeric mice were bred to C57BL/6J females to obtain
transgenic mice that were 129X1/SvJ/ C57BL/6J hybrids. The fraction of alleles
from the 129X1/SvJ background was reduced from 50% because the mouse
lines were maintained by crossing to C57BL/6J mice. The degree of this reduction can only be estimated because the number of crosses that had been
performed since derivation of the original 129X1/SvJ/ C57BL/6J hybrid
transgenic lines was not available. The FVB/n background was introduced when
the R26R-ECFP and R26R-EYFP mice were each crossed to a line of FVB/n mice that expresses the Cre recombinase during early embryonic development
[139]. This cross was necessary to remove a floxed transcriptional stop cassette situated at the beginning of each fluorescent protein gene. We screened the offspring of these crosses and found that some of the mice expressed fluorescent protein ubiquitously, and analysis (by PCR) of the transgenes in fluorescent mice showed that the stop cassette had been removed. These results were as expected based on previous reports [139]. We selected one mouse exhibiting YFP fluorescence and one exhibiting CFP fluorescent and used
them to establish two lines (R26YFP and R26CFP) by inbreeding. Hence, it was expected that most loci in the R26YFP and R26CFP lines would be occupied by
alleles from either C57BL/6J or FVB/n, although alleles from strain 129X1/SvJ
could be present at some loci.
107 Derivation, culture and treatment of cell lines
Mouse embryonic fibroblast (MEF) polyclonal cell lines were derived as
described [173]. Briefly, R26YFP and R26CFP mice were mated and at 13.5 dpc pregnant mice were sacrificed. Embryos were harvested and heart, liver and blood were removed and discarded. The remaining tissue was minced and pieces suspended in 1 mL 0.05% trypsin-EDTA (Invitrogen, Carlsbad, CA) and subjected to 12-18 hour digestion in at 4°C. Following digestion, cells were
mechanically disaggregated by repeated pipetting in Dulbecco’s modified eagle media (DMEM), supplemented with 10% FBS (Invitrogen, Carlsbad, CA) and
200mM L-glutamine (Invitrogen, Carlsbad, CA) (complete DMEM). Cells from
one embryo were divided among four 10 cm cell culture dishes and cultured
overnight in DMEM, 200mM L-glutamine, 10% FBS. These cells were
trypsinized and cells divided among four plates. A day later, the cells were
harvested and cryopreserved. For analysis, a vial of cells was thawed and
placed in complete DMEM in a single 10 cm culture dish, which was incubated at
37°C until the plate was confluent (12 to 24 h).
To obtain ES cells, male R26YFP mice were crossed to female R26CFP
mice. Derivation of ES cells was as described [173]. Briefly, blastocysts were
harvested at embryonic day 3.5 and put into culture in ES cell media (Dulbecco’s
modified eagle media (DMEM) that was supplemented with 15% certified FBS
(Invitrogen, Carlsbad, CA), 200mM L-glutamine (Invitrogen), 0.1mM ß-
mercaptoethanol (Sigma, St. Louis, MO) and leukemia inhibitory factor (LIF),
108 1000U/ml (Chemicon, Temecula, CA)). Two cultured blastocysts each yielded a
cell line that maintained morphology characteristic of mouse ES cells.
ES cell lines were maintained on primary embryonic mouse fibroblast (MEF)
feeder layers, mitotically inactivated by treatment with mitomycin C (Fisher,
Pittsburgh, PA) 0.01 μg/ml in DMEM for 1-2 hours, using standard incubation
conditions: 37°C, 10% CO2 [173].
ES cells were mutagenized by the procedure described by Munroe et al.
[61]. Briefly, approximately 1 million ES cells on a feeder layer were exposed to
culture media containing 0.6 mg/ml ethylmethanesulfonate (EMS) for 20 hours, at
which time culture media was removed, cells were washed several times and
then placed in ES cell media and incubated at 37°C for 10 days. Cells were
harvested and analyzed by FACS.
FACS analysis
Cells were removed from plates by digestion with 0.25% trypsin-EDTA
(Invitrogen, Carlsbad, CA) at 37°C for 5 to 15 min. Cells were suspended in sorting buffer (Ca++ & Mg++ free PBS, 1mM EDTA, 25mM Hepes,1%FBS),
transferred into Falcon FACS tubes with cell strainers (35-2235)(BD Biosciences,
San Jose, CA) and examined by light microscopy to verify the absence of
aggregates.
Cells were sorted using a FACSVantageTM SE (BD Biosciences, San
Jose, CA) flow cytometer equipped with digital DiVa Software, an argon laser
(488 nm), Coherent Innova krypton laser tuned to 407 nm, 90 micron nozzle using a sheath pressure of 26 pounds per square inch, standard optics including
109 a 530/30 nm band-pass filter (FL1) for YFP and 480/30 nm band-pass filter (FL4)
for CFP. Daily alignment was performed using chicken red blood cells (BioSure,
Grass Valley, CA) and AlignFlow Plus UV beads (Molecular Probes, Eugene,
OR) for the 488 nm laser and UV laser, respectively.
Cells were first gated based on forward and side scatter. Then, cells were
analyzed for both CFP and YFP fluorescence. Gates for isolation of monochromatic embryonic mouse cells were established using cells (either MEF or ES) made for the purpose. These cells carried and expressed either a YFP or a CFP gene. Sorted cells were collected into 1.5 ml conical tubes containing cell culture media, supplemented with 1x antibiotic-antimycotic (Invitrogen). Cells recovered from the FACS were cultured in media containing antibiotics.
ES cells subjected to FACS were harvested from plates that contained both
transgenic ES cells and wild type MEF feeder cells. Although MEF feeder cells
did not carry either fluorescent protein, it was expected that these cells might
contribute to the FACS data due to autofluorescence. In fact, the FACS data
produced by ES cell cultures contained a small number of faintly fluorescent points. To determine if these points were from feeder cells, the dim objects were isolated and placed in culture dishes. The next day, the dishes were inspected by microscopy, which showed the presence of adherent cells that were morphologically identical to feeders and were dimly fluorescent.
Data files were saved for every sort and subsequently analyzed using
FlowJo software (TreeStar, San Carlos, CA). Frequencies reported are based on
110 analysis of 2.5x106 events per sort. Statistical comparisons were performed using the Student’s paired T-test.
Noise calculations
Intrinsic and extrinsic expression noise values were calculated using
formulas previously described [145, 146]. Calculations were performed using
CFP fluorescence values that had been adjusted to correct for the intrinsic
faintness of CFP molecules compared to YFP molecules. An adjustment factor
was determined for each cell line by comparing the population mean YFP
fluorescence to the population mean CFP fluorescence. This factor varied
slightly from one cell line to the other, but, generally, CFP fluorescence values
used in noise calculation were approximately 2 fold higher the values reported by
the FACS.
Rate Estimation
Rates of allele loss were obtained by studying the frequency of
monochromatic variants in populations that arose during the expansion of
populations that initially lacked monochromatic variants, which were obtained by
FACS. Rates were estimated by dividing the proportions of variants by the
number of generations estimated to have elapsed during expansion of initially
variant-free populations [71, 155].
Phenotypic analysis of R26CY subclones
FACS-isolated subclones were expanded in culture and then analyzed by
fluorescent microscopy on a Nikon E400 microscope. The filter set for CFP
111 provided excitation light wavelengths between 426 and 446 nm and allowed
detection of emitted light between 460 and 500 nm. The filter set for YFP
provided excitation light wavelengths between 490 and 510 nm and allowed
detection of emitted light between 520 and 550 nm. Experiments with
monochromatic cells showed that the signals from the two fluorescent proteins
could be detected without interference of one with the other. Images were captured using a Spot Jr. CCD camera (Diagnostic Instruments, Sterling Heights,
MI).
DNA analysis of R26CY subclones
ES cells were first separated from feeder cells by FACS. This separation
was easily accomplished because feeder cells were not very fluorescent. DNA to
be used for PCR was isolated from cells by proteinase K digestion in lysis buffer
(5mM EDTA, 200mM NaCl, 100mM Tris, 0.2% SDS) 8-12 hours at 57°C followed
by isopropanol precipitation. PCR was performed in 50μl containing 2.5mM
MgCl2 (Promega, Madison, WI), 1x PCR buffer (Promega), 1.6 mM dNTPs, 0.5U
Taq polymerase (Promega).
To analyze the ROSA26 locus, Primer 1
(5’CAGTAGTCCAGGGTTTCCTTGATG) which is specific to the ROSA26
promoter region, was paired with Primer 2 (5’GTCGCGGCCGCTTTACTTGT),
which is common to both the CFP and YFP genes. Cycling conditions were as
follows: 3 min denaturation at 94°C followed by 35 cycles of 94°C for 1 min; 58°C
for 1 min, 72°C for 1 min, after which there was a final 5 min extension reaction
at 72°C. The 1 Kb (approximate size) product was purified using the QIAquick
112 PCR purification kit (Qiagen, Valencia, CA). CFP and YFP PCR products could
be distinguished by restriction fragment length polymorphism (RFLP) analysis with Pst1 endonuclease (10U/μl)(Invitrogen) using gel electrophoresis (1.5%
agarose). The CFP product contains one Pst1 recognition site yielding bands of
roughly 800 and 200 bp in size. The YFP product has two Pst1 sites and
generates bands of approximately 500, 300 and 200 bp in size. To confirm the
results of the Pst1 RFLP analysis, PCR products were cloned into TOPO-TA
(Invitrogen) and sequenced. Statistical significance was tested using the
binomial distribution.
LOH at ROSA26 might be accompanied by LOH at linked loci. To identify
linked loci that would be informative, i.e. heterozygous in the parental ES cell
lines R26CY1 and R26CY2, were sought. The first step was to identify candidate
loci using the Mouse Genome Informatics (MGI) database
(www.informatics.jax.org). Several loci were then tested by PCR and sequencing
performed by Polymorphic DNA technologies (Alameda, CA)
(www.polymorphicdna.com). The D6MIT159 locus was found to be
heterozygous. Genomic DNA samples were subjected to PCR using the primers
described in the MGI database. PCR products were resolved using gel
electrophoresis through a 3% agarose gel. Parental cells produced the two
bands expected. Cells that had lost heterozygosity at D6MIT159 produced only
one band. In all cases of LOH at D6MIT159, the band that was retained was the
one linked to the ROSA26 allele retained.
113 Cytogenetic analyses
Metaphase chromosome preparations for spectral karyotyping (SKy)
analysis were prepared by treating 50-70% confluent cultures with 10µg/ml
colchicine (Sigma) for 2 hours. Following treatment, the cells were lifted with
0.25% trypsin, re-suspended in ES cell media, collected by centrifugation at 1300
RPM for 5 minutes, and re-suspended in 10ml hypotonic solution (KCl 3g/L,
HEPES 4.8g/L, EGTA 0.2g/L, NaOH 0.36g/L, pH7.4). This suspension was
incubated at 37°C for 50 minutes. Subsequently, 2ml of fixative (methanol acetic
acid 3:1) was added, the cells were collected by centrifugation at 2000 RPM for 5
minutes and the pellet was re-suspended in 12ml fixative. The cells were stored
at -80°C until use. The SKy was performed by the SKy/FISH facility at the
Roswell Park Cancer Institute (Buffalo, NY) (Figure 8).
Whole chromosome painting was performed on metaphase chromosomes
prepared as described above. Chromosome spreads were prepared as
previously described [174]. Chromosome paints for chromosome 6 (starfish
paints) were purchased from Cambio Ltd. (Cambs, UK). Hybridizations were
performed according to the manufacturer’s protocol. Fluorescent images were
collected on a Leica TCS SP2 confocal microscope and edited using Velocity
software.
Estimation of the maximum number of trisomic cells in a population was
obtained from the following relationship: (1-f)x=1-p, where p is the probability of seeing at least 1 trisomic cell, x is the number of cells examined. and f is the frequency of trisomy in the population. When p is set at 0.95, and f at 0.05, X is
114 59. Thus, if one examines 59 cells and does not see a trisomic cell one can
conclude with 95% confidence that the frequency of trisomic cells in the population is 5% or less.
Authors’ Contributions
JSL derived MEF cell lines, maintained all cell lines (MEF & ES),
mutagenized ES cells, performed FACS and analyzed FACS data, cloned variant
ES cells, performed FISH and other fluorescent microscopy techniques, and analyzed cells for allele loss. JSL also contributed to manuscript preparation.
MY performed timed mating of mice, and harvested and cultured blastocysts for the establishment of ES cell lines. JMF assisted in the establishment of ES cell lines and carried out much of the microsatellite analysis. SLS assisted in the establishment of ES cell lines and provided advice on experimental design and interpretation of data. JRS was the principal investigator and corresponding author. JRS conceived the project, helped with interpretation of data, contributed to authorship and final drafting of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors thank the Flow cytometry core at Children’s Hospital,
Cincinnati, the gene targeting and DNA sequencing cores of the University of
Cincinnati, Yuri Nikiforov and Manoj Gandhi for assistance with FISH and
confocal microscopy. Jon S. Larson was supported by the National Institute for
115 Environmental Health Sciences (NIEHS) training grant T32 ES07250. This work was supported by grants P42 ES04908 and P30-ES06096 from the NIEHS.
116 Chapter 6
Dissertation Summary & Directions
Multicellular organisms are mosaic in nature because of genetic
alterations that occur in somatic cells. There are many factors that can contribute
to the formation of such alterations including aberrant DNA repair, aberrant cell
cycle, environmental insults, epigenetic modification, errors in DNA replication
and errors in chromosome duplication/segregation. To further the study of the distributions, frequencies and rates at which some alterations can occur, mouse reporter models were implemented.
The Tg(βA-G11PLAP) transgenic mouse of mutation harbors an allele
(G11 PLAP ) that is rendered incapable of producing its functional enzyme because a frameshift caused by an insertion of 11 G:C basepairs. Spontaneous deletion of one G:C basepair from this mononucleotide repeat restores gene
function, and cells with PLAP activity can be detected histochemically. G11 PLAP
mice enable mutant cells to be visualized in situ and were used to study variation
during early development, in the germline, under oxidative stress and in solid
tumors.
Past and present studies utilized that PLAPG11 transgenic mouse have
been both informative and productive. This approach has effectively
demonstrated the ability to visualize mutant cells in situ in most cell and tissue
types of the whole mouse. However there are some drawbacks that merit
consideration when interpreting the data produced from PLAP mice as well as
proving rationale to improve this model. The primary concern is that transgenic
117 mouse contains approximately five copies (oriented head-to-tail) of the plap
transgene. This feature has made detecting the genetic basis for the reversion phenotype extremely difficult in genomic samples. Discriminating on the basis of a one base difference in mononucleotide tract length variation requires utmost sensitivity which has eluded detection, presumably because of background caused by the presence of multiple copies of the mononucleotide run.
Another consideration is that although there are five reporter genes
present in the PLAPG11 mouse, but effectively only one mutational event can be
detected. Having the ability to visualize multiple events within the same cell
would prove invaluable when determining the degree of MSI, i.e. MSI-H or MSI-L
within tissues and tumors. That is, when the rate of reversion is 1e10-5 the
expected frequency of two independent events would be 1e10-10. However if a
cell(s) acquired a mutator phenotype the frequency at which multiple events are
seen could be orders higher. This could be accomplished by creating a MSI
reporter using the humanized LacZ gene. This reporter has been shown to target
efficiently to the ROSA26 locus in mouse, where it expresses well in all cells and tissues, P. Soriano lab articles and website
(http://www.fhcrc.org/science/labs/soriano/).
To study additional genetic aberrations e.g. LOH in diverse cell types in
the body another reporter model was implemented which distinguishes events by
a detectable cellular phenotype. Mice that carry two different fluorescent protein
genes as alleles of a locus were generated to address this issue because LOH
would change a cell’s phenotype from bichrome to monochrome. As a step in
118 assessing the utility of this approach, we derived MEF and ES cell lines from
mice that carried two different fluorescent protein genes as alleles at the
chromosome 6 locus, ROSA26. FACS showed that the vast majority of cells in
each line expressed the two marker proteins at similar levels, but populations
exhibited extrinsic and intrinsic noise with respect to expression. In addition, cells
with a monochrome phenotype were frequent (10-4). In ES cells, all monochrome
events were accompanied by allele loss. Mitotic recombination appeared to be
the major cause, although UPD also appeared to have contributed to LOH.
These cells provided a novel assay for studying genetic/karyotypic stability of
cultured ES cells and confirm the need for caution regarding the use of cultured
stem cells in therapy.
Studies utilizing the R26CY reporter model are still in their infancy. Future
studies with this model will be conducted in genetic backgrounds at risk for
disease (e.g. bloom’s syndrome) to get a clearer understanding and better resolutions of the genetic components of disease in the whole animal. Whole animal studies are an ultimate goal of this model system. However, to this point such studies have been hindered by the available methods of preparing tissues for microscopy. The fluorescent proteins used in R26CY mice reside in the cytoplasm. Therefore, fixing agents that permeablize the cell membrane abolish fluorescence. Aldehyde based fixatives must be used with caution as they produce autofluorescence. One possibility to overcome this, eyfp and ecfp can be replaced with reporter genes that have cellular localization signals e.g. nucleus or mitochondria. Furthermore, other GFP variants and others e.g.
119 dsRed2 can be tested. Another potential problem with these reporters is that they
don not fluoresce as intensely in non-living cells and are susceptible to the
effects photobleaching. In an effort to overcome this some invetigators have
turned to the use of immunofluorescnece with antibodies raised against GFP or an incorporated tag. This too is problematic because there are not antibodies that
possess sufficient specificity for other GFP variants.
120 References
1. Weinberg RA: How cancer arises. Sci Am 1996, 275:62-70. 2. Knudson AG: Cancer genetics. Am J Med Genet 2002, 111:96-102. 3. Ashton-Rickardt PG, Dunlop MG, Nakamura Y, Morris RG, Purdie CA, Steel CM, Evans HJ, Bird CC, Wyllie AH: High frequency of APC loss in sporadic colorectal carcinoma due to breaks clustered in 5q21-22. Oncogene 1989, 4:1169-1174. 4. Borgdorff V, van Hees-Stuivenberg S, Meijers CM, de Wind N: Spontaneous and mutagen-induced loss of DNA mismatch repair in Msh2-heterozygous mammalian cells. Mutat Res 2005, 574:50-57. 5. Brisson D: The directed mutation controversy in an evolutionary context. Crit Rev Microbiol 2003, 29:25-35. 6. Caporale LH: Natural selection and the emergence of a mutation phenotype: an update of the evolutionary synthesis considering mechanisms that affect genome variation. Annu Rev Microbiol 2003, 57:467-485. 7. Burns EM, Christopoulou L, Corish P, Tyler-Smith C: Quantitative measurement of mammalian chromosome mitotic loss rates using the green fluorescent protein. J Cell Sci 1999, 112 ( Pt 16):2705-2714. 8. Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R, Gallie BL, Murphree AL, Strong LC, White RL: Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 1983, 305:779- 784. 9. Chang DK, Metzgar D, Wills C, Boland CR: Microsatellites in the eukaryotic DNA mismatch repair genes as modulators of evolutionary mutation rate. Genome Res 2001, 11:1145-1146.
121 10. Turker MS: Estimation of mutation frequencies in normal mammalian cells and the development of cancer. Semin Cancer Biol 1998, 8:407- 419. 11. Knudson AG: Antioncogenes and human cancer. Proc Natl Acad Sci U S A 1993, 90:10914-10921. 12. Hethcote HW, Knudson AG, Jr.: Model for the incidence of embryonal cancers: application to retinoblastoma. Proc Natl Acad Sci U S A 1978, 75:2453-2457. 13. Cleton-Jansen AM, Collins N, Lakhani SR, Weissenbach J, Devilee P, Cornelisse CJ, Stratton MR: Loss of heterozygosity in sporadic breast tumours at the BRCA2 locus on chromosome 13q12-q13. Br J Cancer 1995, 72:1241-1244. 14. Beder LB, Gunduz M, Ouchida M, Fukushima K, Gunduz E, Ito S, Sakai A, Nagai N, Nishizaki K, Shimizu K: Genome-wide analyses on loss of heterozygosity in head and neck squamous cell carcinomas. Lab Invest 2003, 83:99-105. 15. Fodde R, Kuipers J, Rosenberg C, Smits R, Kielman M, Gaspar C, van Es JH, Breukel C, Wiegant J, Giles RH, Clevers H: Mutations in the APC tumour suppressor gene cause chromosomal instability. Nat Cell Biol 2001, 3:433-438. 16. Huang Y, Boynton RF, Blount PL, Silverstein RJ, Yin J, Tong Y, McDaniel TK, Newkirk C, Resau JH, Sridhara R, et al.: Loss of heterozygosity involves multiple tumor suppressor genes in human esophageal cancers. Cancer Res 1992, 52:6525-6530. 17. Kinzler KW, Vogelstein B: Lessons from hereditary colorectal cancer. Cell 1996, 87:159-170. 18. Nussbaum RL, McInnes RR, Willard HF, Thompson MW: Thompson & Thompson genetics in medicine, 6th edn. Philadelphia: Saunders; 2001. 19. Bodmer WF: Cancer genetics: colorectal cancer as a model. J Hum Genet 2006, 51:391-396.
122 20. Barnetson R, Jass J, Tse R, Eckstein R, Robinson B, Schnitzler M: Mutations associated with microsatellite unstable colorectal carcinomas exhibit widespread intratumoral heterogeneity. Genes Chromosomes Cancer 2000, 29:130-136. 21. Duval A, Hamelin R: Mutations at coding repeat sequences in mismatch repair-deficient human cancers: toward a new concept of target genes for instability. Cancer Res 2002, 62:2447-2454. 22. Oki E, Oda S, Maehara Y, Sugimachi K: Mutated gene-specific phenotypes of dinucleotide repeat instability in human colorectal carcinoma cell lines deficient in DNA mismatch repair. Oncogene 1999, 18:2143-2147. 23. DePrimo SE, Cao J, Hersh MN, Stringer JR: Use of human placental alkaline phosphatase transgenes to detect somatic mutation in mice in situ. Methods 1998, 16:49-61. 24. DePrimo SE, Stambrook PJ, Stringer JR: Human placental alkaline phosphatase as a histochemical marker of gene expression in transgenic mice. Transgenic Res 1996, 5:459-466. 25. Larson JS, Stringer SL, Stringer JR: Impact of mismatch repair deficiency on genomic stability in the maternal germline and during early embryonic development. Mutat Res 2004, 556:45-53. 26. Stringer JR, Larson JS, Fischer JM, Stringer SL: Increased mutation in mice genetically predisposed to oxidative damage in the brain. Mutat Res 2004, 556:127-134. 27. Stringer JR, Larson JS, Fischer JM, Medvedovic M, Hersh MN, Boivin GP, Stringer SL: Modeling variation in tumors in vivo. Proc Natl Acad Sci U S A 2005, 102:2408-2413. 28. Qian F, Germino GG: "Mistakes happen": somatic mutation and disease. Am J Hum Genet 1997, 61:1000-1005. 29. Smith SA, Easton DF, Evans DG, Ponder BA: Allele losses in the region 17q12-21 in familial breast and ovarian cancer involve the wild-type chromosome. Nat Genet 1992, 2:128-131.
123 30. Tischfield JA: Loss of heterozygosity or: how I learned to stop worrying and love mitotic recombination. Am J Hum Genet 1997, 61:995-999. 31. Cappione AJ, French BL, Skuse GR: A potential role for NF1 mRNA editing in the pathogenesis of NF1 tumors. Am J Hum Genet 1997, 60:305-312. 32. Lodish HF: Molecular cell biology, 5th edn. New York: W.H. Freeman and Company; 2004. 33. Engle SJ, Stockelman MG, Chen J, Boivin G, Yum MN, Davies PM, Ying MY, Sahota A, Simmonds HA, Stambrook PJ, Tischfield JA: Adenine phosphoribosyltransferase-deficient mice develop 2,8- dihydroxyadenine nephrolithiasis. Proc Natl Acad Sci U S A 1996, 93:5307-5312. 34. Turker MS: Autosomal mutation in somatic cells of the mouse. Mutagenesis 2003, 18:1-6. 35. Tischfield JA, Ruddle FH: Assignment of the gene for adenine phosphoribosyltransferase to human chromosome 16 by mouse- human somatic cell hybridization. Proc Natl Acad Sci U S A 1974, 71:45-49. 36. Gupta PK, Sahota A, Boyadjiev SA, Bye S, Shao C, O'Neill JP, Hunter TC, Albertini RJ, Stambrook PJ, Tischfield JA: High frequency in vivo loss of heterozygosity is primarily a consequence of mitotic recombination. Cancer Res 1997, 57:1188-1193. 37. Stambrook PJ, Shao C, Stockelman M, Boivin G, Engle SJ, Tischfield JA: APRT: a versatile in vivo resident reporter of local mutation and loss of heterozygosity. Environ Mol Mutagen 1996, 28:471-482. 38. Shao C, Deng L, Henegariu O, Liang L, Raikwar N, Sahota A, Stambrook PJ, Tischfield JA: Mitotic recombination produces the majority of recessive fibroblast variants in heterozygous mice. Proc Natl Acad Sci U S A 1999, 96:9230-9235.
124 39. Wijnhoven SW, Sonneveld E, Kool HJ, van Teijlingen CM, Vrieling H: Chemical carcinogens induce varying patterns of LOH in mouse T- lymphocytes. Carcinogenesis 2003, 24:139-144. 40. Cervantes RB, Stringer JR, Shao C, Tischfield JA, Stambrook PJ: Embryonic stem cells and somatic cells differ in mutation frequency and type. Proc Natl Acad Sci U S A 2002, 99:3586-3590. 41. Dobrovolsky VN, Casciano DA, Heflich RH: Tk+/- mouse model for detecting in vivo mutation in an endogenous, autosomal gene. Mutat Res 1999, 423:125-136. 42. Tomlinson I, Sasieni P, Bodmer W: How many mutations in a cancer? Am J Pathol 2002, 160:755-758. 43. Lengauer C, Kinzler KW, Vogelstein B: Genetic instability in colorectal cancers. Nature 1997, 386:623-627. 44. Sieber OM, Heinimann K, Tomlinson IP: Genomic instability--the engine of tumorigenesis? Nat Rev Cancer 2003, 3:701-708. 45. Jackson AL, Loeb LA: The mutation rate and cancer. Genetics 1998, 148:1483-1490. 46. Cao J, DePrimo SE, Hersh MN, Stringer JR: In situ detection of frameshift mutations in mouse cells. Mutat Res 1998, 421:163-178. 47. Chakraborty R, Kimmel M, Stivers DN, Davison LJ, Deka R: Relative mutation rates at di-, tri-, and tetranucleotide microsatellite loci. Proc Natl Acad Sci U S A 1997, 94:1041-1046. 48. Iino H, Jass JR, Simms LA, Young J, Leggett B, Ajioka Y, Watanabe H: DNA microsatellite instability in hyperplastic polyps, serrated adenomas, and mixed polyps: a mild mutator pathway for colorectal cancer? J Clin Pathol 1999, 52:5-9. 49. Ikenoue T, Togo G, Nagai K, Ijichi H, Kato J, Yamaji Y, Okamoto M, Kato N, Kawabe T, Tanaka A, Matsumura M, Shiratori Y, Omata M: Frameshift mutations at mononucleotide repeats in RAD50 recombinational DNA repair gene in colorectal cancers with microsatellite instability. Jpn J Cancer Res 2001, 92:587-591.
125 50. Schwartz S, Jr., Yamamoto H, Navarro M, Maestro M, Reventos J, Perucho M: Frameshift mutations at mononucleotide repeats in caspase-5 and other target genes in endometrial and gastrointestinal cancer of the microsatellite mutator phenotype. Cancer Res 1999, 59:2995-3002. 51. Codegoni AM, Bertoni F, Colella G, Caspani G, Grassi L, D'Incalci M, Broggini M: Microsatellite instability and frameshift mutations in genes involved in cell cycle progression or apoptosis in ovarian cancer. Oncol Res 1999, 11:297-301. 52. Yamamoto H, Sawai H, Perucho M: Frameshift somatic mutations in gastrointestinal cancer of the microsatellite mutator phenotype. Cancer Res 1997, 57:4420-4426. 53. Laken SJ, Petersen GM, Gruber SB, Oddoux C, Ostrer H, Giardiello FM, Hamilton SR, Hampel H, Markowitz A, Klimstra D, Jhanwar S, Winawer S, Offit K, Luce MC, Kinzler KW, Vogelstein B: Familial colorectal cancer in Ashkenazim due to a hypermutable tract in APC. Nat Genet 1997, 17:79-83. 54. Markowitz S, Wang J, Myeroff L, Parsons R, Sun L, Lutterbaugh J, Fan RS, Zborowska E, Kinzler KW, Vogelstein B, et al.: Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science 1995, 268:1336-1338. 55. Rampino N, Yamamoto H, Ionov Y, Li Y, Sawai H, Reed JC, Perucho M: Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 1997, 275:967-969. 56. Planck M, Wenngren E, Borg A, Olsson H, Nilbert M: Somatic frameshift alterations in mononucleotide repeat-containing genes in different tumor types from an HNPCC family with germline MSH2 mutation. Genes Chromosomes Cancer 2000, 29:33-39. 57. Wang J, Sun L, Myeroff L, Wang X, Gentry LE, Yang J, Liang J, Zborowska E, Markowitz S, Willson JK, et al.: Demonstration that mutation of the type II transforming growth factor beta receptor
126 inactivates its tumor suppressor activity in replication error-positive colon carcinoma cells. J Biol Chem 1995, 270:22044-22049. 58. Woerner SM, Gebert J, Yuan YP, Sutter C, Ridder R, Bork P, von Knebel Doeberitz M: Systematic identification of genes with coding microsatellites mutated in DNA mismatch repair-deficient cancer cells. Int J Cancer 2001, 93:12-19. 59. Olivero M, Ruggiero T, Coltella N, Maffe A, Calogero R, Medico E, Di Renzo MF: Amplification of repeat-containing transcribed sequences (ARTS): a transcriptome fingerprinting strategy to detect functionally relevant microsatellite mutations in cancer. Nucleic Acids Res 2003, 31:e33. 60. Hersh MN, Stambrook PJ, Stringer JR: Visualization of mosaicism in tissues of normal and mismatch-repair-deficient mice carrying a microsatellite-containing transgene. Mutat Res 2002, 505:51-62. 61. Munroe RJ, Bergstrom RA, Zheng QY, Libby B, Smith R, John SW, Schimenti KJ, Browning VL, Schimenti JC: Mouse mutants from chemically mutagenized embryonic stem cells. Nat Genet 2000, 24:318-321. 62. Dahiya R, Lee C, Zhu Z, Thompson HJ: Microsatellite instability in an animal model of mammary carcinogenesis. Int J Oncol 1998, 13:23-28. 63. Fleisher AS, Esteller M, Harpaz N, Leytin A, Rashid A, Xu Y, Liang J, Stine OC, Yin J, Zou TT, Abraham JM, Kong D, Wilson KT, James SP, Herman JG, Meltzer SJ: Microsatellite instability in inflammatory bowel disease-associated neoplastic lesions is associated with hypermethylation and diminished expression of the DNA mismatch repair gene, hMLH1. Cancer Res 2000, 60:4864-4868. 64. Weinberg HS, Korol AB, Kirzhner VM, Avivi A, Fahima T, Nevo E, Shapiro S, Rennert G, Piatak O, Stepanova EI, Skvarskaja E: Very high mutation rate in offspring of Chernobyl accident liquidators. Proc Biol Sci 2001, 268:1001-1005.
127 65. Guy CT, Cardiff RD, Muller WJ: Induction of mammary tumors by expression of polyomavirus middle T oncogene: a transgenic mouse model for metastatic disease. Mol Cell Biol 1992, 12:954-961. 66. Ichaso N, Dilworth SM: Cell transformation by the middle T-antigen of polyoma virus. Oncogene 2001, 20:7908-7916. 67. Dilworth SM, Hansson HA, Darnfors C, Bjursell G, Streuli CH, Griffin BE: Subcellular localisation of the middle and large T-antigens of polyoma virus. Embo J 1986, 5:491-499. 68. Denko NC, Giaccia AJ, Stringer JR, Stambrook PJ: The human Ha-ras oncogene induces genomic instability in murine fibroblasts within one cell cycle. Proc Natl Acad Sci U S A 1994, 91:5124-5128. 69. Denko N, Stringer J, Wani M, Stambrook P: Mitotic and post mitotic consequences of genomic instability induced by oncogenic Ha-ras. Somat Cell Mol Genet 1995, 21:241-253. 70. Wani MA, Denko NC, Stambrook PJ: Expression of Rap 1 suppresses genomic instability of H-ras transformed mouse fibroblasts. Somat Cell Mol Genet 1997, 23:123-133. 71. Hayes W: The genetics of bacteria and their viruses; studies in basic genetics and molecular biology, 2d edn. New York,: Wiley; 1968. 72. Gonzalez-Garcia I, Sole RV, Costa J: Metapopulation dynamics and spatial heterogeneity in cancer. Proc Natl Acad Sci U S A 2002, 99:13085-13089. 73. Saavedra HI, Fukasawa K, Conn CW, Stambrook PJ: MAPK mediates RAS-induced chromosome instability. J Biol Chem 1999, 274:38083- 38090. 74. Saavedra HI, Knauf JA, Shirokawa JM, Wang J, Ouyang B, Elisei R, Stambrook PJ, Fagin JA: The RAS oncogene induces genomic instability in thyroid PCCL3 cells via the MAPK pathway. Oncogene 2000, 19:3948-3954.
128 75. Miller FR, Pauley RJ, Wang B: Activated c-Ha-ras is not sufficient to produce the preneoplastic phenotype of human breast cell line MCF10AT. Anticancer Res 1996, 16:1765-1769. 76. Ruley HE, Fried M: Clustered illegitimate recombination events in mammalian cells involving very short sequence homologies. Nature 1983, 304:181-184. 77. Ruley HE: Transforming collaborations between ras and nuclear oncogenes. Cancer Cells 1990, 2:258-268. 78. Lowe SW, Jacks T, Housman DE, Ruley HE: Abrogation of oncogene- associated apoptosis allows transformation of p53-deficient cells. Proc Natl Acad Sci U S A 1994, 91:2026-2030. 79. Neuhausen S, Gilewski T, Norton L, Tran T, McGuire P, Swensen J, Hampel H, Borgen P, Brown K, Skolnick M, Shattuck-Eidens D, Jhanwar S, Goldgar D, Offit K: Recurrent BRCA2 6174delT mutations in Ashkenazi Jewish women affected by breast cancer. Nat Genet 1996, 13:126-128. 80. Robson M, Dabney MK, Rosenthal G, Ludwig S, Seltzer MH, Gilewski T, Haas B, Osborne M, Norton L, Gilbert F, Offit K: Prevalence of recurring BRCA mutations among Ashkenazi Jewish women with breast cancer. Genet Test 1997, 1:47-51. 81. Cropp CS, Nevanlinna HA, Pyrhonen S, Stenman UH, Salmikangas P, Albertsen H, White R, Callahan R: Evidence for involvement of BRCA1 in sporadic breast carcinomas. Cancer Res 1994, 54:2548-2551. 82. Desai KV, Xiao N, Wang W, Gangi L, Greene J, Powell JI, Dickson R, Furth P, Hunter K, Kucherlapati R, Simon R, Liu ET, Green JE: Initiating oncogenic event determines gene-expression patterns of human breast cancer models. Proc Natl Acad Sci U S A 2002, 99:6967-6972. 83. Balogh GA, Russo IH, Russo J: Mutations in mismatch repair genes are involved in the neoplastic transformation of human breast epithelial cells. Int J Oncol 2003, 23:411-419.
129 84. Hofseth LJ, Khan MA, Ambrose M, Nikolayeva O, Xu-Welliver M, Kartalou M, Hussain SP, Roth RB, Zhou X, Mechanic LE, Zurer I, Rotter V, Samson LD, Harris CC: The adaptive imbalance in base excision- repair enzymes generates microsatellite instability in chronic inflammation. J Clin Invest 2003, 112:1887-1894. 85. Latham KM, Stanbridge EJ: Examination of the oncogenic potential of a tumor-associated antigen, intestinal alkaline phosphatase, in HeLa x fibroblast cell hybrids. Cancer Res 1992, 52:616-622. 86. Fujii H, Yoshida M, Gong ZX, Matsumoto T, Hamano Y, Fukunaga M, Hruban RH, Gabrielson E, Shirai T: Frequent genetic heterogeneity in the clonal evolution of gynecological carcinosarcoma and its influence on phenotypic diversity. Cancer Res 2000, 60:114-120. 87. Baisse B, Bouzourene H, Saraga EP, Bosman FT, Benhattar J: Intratumor genetic heterogeneity in advanced human colorectal adenocarcinoma. Int J Cancer 2001, 93:346-352. 88. Mora J, Cheung NK, Gerald WL: Genetic heterogeneity and clonal evolution in neuroblastoma. Br J Cancer 2001, 85:182-189. 89. Samowitz WS, Slattery ML: Regional reproducibility of microsatellite instability in sporadic colorectal cancer. Genes Chromosomes Cancer 1999, 26:106-114. 90. Heimann R, Ferguson D, Recant WM, Hellman S: Breast cancer metastatic phenotype as predicted by histologic tumor markers. Cancer J Sci Am 1997, 3:224-229. 91. Heimann R, Hellman S: Aging, progression, and phenotype in breast cancer. J Clin Oncol 1998, 16:2686-2692. 92. Hellman S, Heimann R: The clinical significance of tumor progression: breast cancer as a model. Cancer J 2000, 6 Suppl 2:S131-133. 93. Gatenby RA, Vincent TL: An evolutionary model of carcinogenesis. Cancer Res 2003, 63:6212-6220. 94. Barber BR: Two New Mutations. Mouse News Lett 1971:34-35.
130 95. Klein JA, Longo-Guess CM, Rossmann MP, Seburn KL, Hurd RE, Frankel WN, Bronson RT, Ackerman SL: The harlequin mouse mutation downregulates apoptosis-inducing factor. Nature 2002, 419:367-374. 96. Susin SA, Lorenzo HK, Zamzami N, Marzo I, Snow BE, Brothers GM, Mangion J, Jacotot E, Costantini P, Loeffler M, Larochette N, Goodlett DR, Aebersold R, Siderovski DP, Penninger JM, Kroemer G: Molecular characterization of mitochondrial apoptosis-inducing factor. Nature 1999, 397:441-446. 97. Miramar MD, Costantini P, Ravagnan L, Saraiva LM, Haouzi D, Brothers G, Penninger JM, Peleato ML, Kroemer G, Susin SA: NADH oxidase activity of mitochondrial apoptosis-inducing factor. J Biol Chem 2001, 276:16391-16398. 98. Gasche C, Chang CL, Rhees J, Goel A, Boland CR: Oxidative stress increases frameshift mutations in human colorectal cancer cells. Cancer Res 2001, 61:7444-7448. 99. Busuttil RA, Rubio M, Dolle ME, Campisi J, Vijg J: Oxygen accelerates the accumulation of mutations during the senescence and immortalization of murine cells in culture. Aging Cell 2003, 2:287-294. 100. Cardozo-Pelaez F, Brooks PJ, Stedeford T, Song S, Sanchez-Ramos J: DNA damage, repair, and antioxidant systems in brain regions: a correlative study. Free Radic Biol Med 2000, 28:779-785. 101. Jackson AL, Chen R, Loeb LA: Induction of microsatellite instability by oxidative DNA damage. Proc Natl Acad Sci U S A 1998, 95:12468- 12473. 102. Limoli CL, Giedzinski E: Induction of chromosomal instability by chronic oxidative stress. Neoplasia 2003, 5:339-346. 103. Turker MS, Gage BM, Rose JA, Elroy D, Ponomareva ON, Stambrook PJ, Tischfield JA: A novel signature mutation for oxidative damage resembles a mutational pattern found commonly in human cancers. Cancer Res 1999, 59:1837-1839.
131 104. Duval A, Gayet J, Zhou XP, Iacopetta B, Thomas G, Hamelin R: Frequent frameshift mutations of the TCF-4 gene in colorectal cancers with microsatellite instability. Cancer Res 1999, 59:4213-4215. 105. Wichterle H, Turnbull DH, Nery S, Fishell G, Alvarez-Buylla A: In utero fate mapping reveals distinct migratory pathways and fates of neurons born in the mammalian basal forebrain. Development 2001, 128:3759-3771. 106. Sidman RL, Angevine JB, Pierce ET: Atlas of the mouse brain and spinal cord. Cambridge, Mass.,: Harvard University Press; 1971. 107. Nolte J: The human brain : an introduction to its functional anatomy, 4th edn. St. Louis, MO,: Mosby; 1999. 108. Kierszenbaum AL: Histology and Cell Biology: an Introduction to Pathology. St. Louis, MO: Mosby; 2002. 109. Alvarez-Buylla A, Lim DA: For the long run: maintaining germinal niches in the adult brain. Neuron 2004, 41:683-686. 110. Duval A, Reperant M, Hamelin R: Comparative analysis of mutation frequency of coding and non coding short mononucleotide repeats in mismatch repair deficient colorectal cancers. Oncogene 2002, 21:8062-8066. 111. Duval A, Hamelin R: Genetic instability in human mismatch repair deficient cancers. Ann Genet 2002, 45:71-75. 112. Gurtu VE, Verma S, Grossmann AH, Liskay RM, Skarnes WC, Baker SM: Maternal effect for DNA mismatch repair in the mouse. Genetics 2002, 160:271-277. 113. Rinchik EM, Carpenter DA: N-ethyl-N-nitrosourea mutagenesis of a 6- to 11-cM subregion of the Fah-Hbb interval of mouse chromosome 7: Completed testing of 4557 gametes and deletion mapping and complementation analysis of 31 mutations. Genetics 1999, 152:373- 383. 114. Radman M, Taddei F, Matic I: Evolution-driving genes. Res Microbiol 2000, 151:91-95.
132 115. Tanaka MM, Bergstrom CT, Levin BR: The evolution of mutator genes in bacterial populations: the roles of environmental change and timing. Genetics 2003, 164:843-854. 116. Shaver AC, Sniegowski PD: Spontaneously arising mutL mutators in evolving Escherichia coli populations are the result of changes in repeat length. J Bacteriol 2003, 185:6076-6082. 117. Giraud A, Radman M, Matic I, Taddei F: The rise and fall of mutator bacteria. Curr Opin Microbiol 2001, 4:582-585. 118. Sniegowski PD, Gerrish PJ, Johnson T, Shaver A: The evolution of mutation rates: separating causes from consequences. Bioessays 2000, 22:1057-1066. 119. Shaw-White JR, Denko N, Albers L, Doetschman TC, Stringer JR: Expression of the lacZ gene targeted to the HPRT locus in embryonic stem cells and their derivatives. Transgenic Res 1993, 2:1-13. 120. Dieringer D, Schlotterer C: Two distinct modes of microsatellite mutation processes: evidence from the complete genomic sequences of nine species. Genome Res 2003, 13:2242-2251. 121. Narayanan L, Fritzell JA, Baker SM, Liskay RM, Glazer PM: Elevated levels of mutation in multiple tissues of mice deficient in the DNA mismatch repair gene Pms2. Proc Natl Acad Sci U S A 1997, 94:3122- 3127. 122. Baross-Francis A, Makhani N, Liskay RM, Jirik FR: Elevated mutant frequencies and increased C : G-->T : A transitions in Mlh1-/- versus Pms2-/- murine small intestinal epithelial cells. Oncogene 2001, 20:619-625. 123. Yang Y, Karthikeyan R, Mack SE, Vonarx EJ, Kunz BA: Analysis of yeast pms1, msh2, and mlh1 mutators points to differences in mismatch correction efficiencies between prokaryotic and eukaryotic cells. Mol Gen Genet 1999, 261:777-787.
133 124. Risinger JI, Umar A, Barrett JC, Kunkel TA: A hPMS2 mutant cell line is defective in strand-specific mismatch repair. J Biol Chem 1995, 270:18183-18186. 125. Kato T, Yatagai F, Glickman BW, Tachibana A, Ikenaga M: Specificity of mutations in the PMS2-deficient human tumor cell line HEC-1-A. Mutat Res 1998, 422:279-283. 126. Prolla TA, Baker SM, Harris AC, Tsao JL, Yao X, Bronner CE, Zheng B, Gordon M, Reneker J, Arnheim N, Shibata D, Bradley A, Liskay RM: Tumour susceptibility and spontaneous mutation in mice deficient in Mlh1, Pms1 and Pms2 DNA mismatch repair. Nat Genet 1998, 18:276- 279. 127. Awatramani R, Soriano P, Mai JJ, Dymecki S: An Flp indicator mouse expressing alkaline phosphatase from the ROSA26 locus. Nat Genet 2001, 29:257-259. 128. Menoyo A, Alazzouzi H, Espin E, Armengol M, Yamamoto H, Schwartz S, Jr.: Somatic mutations in the DNA damage-response genes ATR and CHK1 in sporadic stomach tumors with microsatellite instability. Cancer Res 2001, 61:7727-7730. 129. Laird PW, Zijderveld A, Linders K, Rudnicki MA, Jaenisch R, Berns A: Simplified mammalian DNA isolation procedure. Nucleic Acids Res 1991, 19:4293. 130. Scrable HJ, Witte DP, Lampkin BC, Cavenee WK: Chromosomal localization of the human rhabdomyosarcoma locus by mitotic recombination mapping. Nature 1987, 329:645-647. 131. Cavenee WK: The recessive nature of dominance. Genes Chromosomes Cancer 2003, 38:322-325. 132. Lefebvre L, Dionne N, Karaskova J, Squire JA, Nagy A: Selection for transgene homozygosity in embryonic stem cells results in extensive loss of heterozygosity. Nat Genet 2001, 27:257-258. 133. Thiagalingam S, Foy RL, Cheng KH, Lee HJ, Thiagalingam A, Ponte JF: Loss of heterozygosity as a predictor to map tumor suppressor
134 genes in cancer: molecular basis of its occurrence. Curr Opin Oncol 2002, 14:65-72. 134. Engel E: A new genetic concept: uniparental disomy and its potential effect, isodisomy. Am J Med Genet 1980, 6:137-143. 135. Pop R, Zaragoza MV, Gaudette M, Dohrmann U, Scherer G: A homozygous nonsense mutation in SOX9 in the dominant disorder campomelic dysplasia: a case of mitotic gene conversion. Hum Genet 2005, 117:43-53. 136. Lasko D, Cavenee W, Nordenskjold M: Loss of constitutional heterozygosity in human cancer. Annu Rev Genet 1991, 25:281-314. 137. Kaushal D, Contos JJ, Treuner K, Yang AH, Kingsbury MA, Rehen SK, McConnell MJ, Okabe M, Barlow C, Chun J: Alteration of gene expression by chromosome loss in the postnatal mouse brain. J Neurosci 2003, 23:5599-5606. 138. Stern C, Rentschler V: The Effect of Temperature on the Frequency of Somatic Crossing-Over in Drosophila Melanogaster. Proc Natl Acad Sci U S A 1936, 22:451-453. 139. Srinivas S, Watanabe T, Lin CS, William CM, Tanabe Y, Jessell TM, Costantini F: Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev Biol 2001, 1:4. 140. Liu X, Wu H, Loring J, Hormuzdi S, Disteche CM, Bornstein P, Jaenisch R: Trisomy eight in ES cells is a common potential problem in gene targeting and interferes with germ line transmission. Dev Dyn 1997, 209:85-91. 141. Longo L, Bygrave A, Grosveld FG, Pandolfi PP: The chromosome make- up of mouse embryonic stem cells is predictive of somatic and germ cell chimaerism. Transgenic Res 1997, 6:321-328. 142. Friedberg EC, Meira LB: Database of mouse strains carrying targeted mutations in genes affecting biological responses to DNA damage Version 7. DNA Repair (Amst) 2006, 5:189-209.
135 143. Mortensen RM, Conner DA, Chao S, Geisterfer-Lowrance AA, Seidman JG: Production of homozygous mutant ES cells with a single targeting construct. Mol Cell Biol 1992, 12:2391-2395. 144. Hawley TS, Telford WG, Hawley RG: "Rainbow" reporters for multispectral marking and lineage analysis of hematopoietic stem cells. Stem Cells 2001, 19:118-124. 145. Elowitz MB, Levine AJ, Siggia ED, Swain PS: Stochastic gene expression in a single cell. Science 2002, 297:1183-1186. 146. Raser JM, O'Shea EK: Control of stochasticity in eukaryotic gene expression. Science 2004, 304:1811-1814. 147. Liu HS, Jan MS, Chou CK, Chen PH, Ke NJ: Is green fluorescent protein toxic to the living cells? Biochem Biophys Res Commun 1999, 260:712-717. 148. Zhao WN, Meng GL, Xue YF: Labeling of three different mouse ES cell lines with the green fluorescent protein. Yi Chuan Xue Bao 2003, 30:743-749. 149. Kato M, Yamanouchi K, Ikawa M, Okabe M, Naito K, Tojo H: Efficient selection of transgenic mouse embryos using EGFP as a marker gene. Mol Reprod Dev 1999, 54:43-48. 150. Ikawa M, Yamada S, Nakanishi T, Okabe M: Green fluorescent protein (GFP) as a vital marker in mammals. Curr Top Dev Biol 1999, 44:1-20. 151. Ikawa M, Yamada S, Nakanishi T, Okabe M: 'Green mice' and their potential usage in biological research. FEBS Lett 1998, 430:83-87. 152. Hadjantonakis AK, Dickinson ME, Fraser SE, Papaioannou VE: Technicolour transgenics: imaging tools for functional genomics in the mouse. Nat Rev Genet 2003, 4:613-625. 153. Klysik J, Singer JD: Mice with the enhanced green fluorescent protein gene knocked in to chromosome 11 exhibit normal transmission ratios. Biochem Genet 2005, 43:321-333.
136 154. Devgan V, Rao MR, Seshagiri PB: Impact of embryonic expression of enhanced green fluorescent protein on early mouse development. Biochem Biophys Res Commun 2004, 313:1030-1036. 155. Eisenstein BI: Phase variation of type 1 fimbriae in Escherichia coli is under transcriptional control. Science 1981, 214:337-339. 156. Yusa K, Horie K, Kondoh G, Kouno M, Maeda Y, Kinoshita T, Takeda J: Genome-wide phenotype analysis in ES cells by regulated disruption of Bloom's syndrome gene. Nature 2004, 429:896-899. 157. Chan MF, van Amerongen R, Nijjar T, Cuppen E, Jones PA, Laird PW: Reduced rates of gene loss, gene silencing, and gene mutation in Dnmt1-deficient embryonic stem cells. Mol Cell Biol 2001, 21:7587- 7600. 158. Luo G, Santoro IM, McDaniel LD, Nishijima I, Mills M, Youssoufian H, Vogel H, Schultz RA, Bradley A: Cancer predisposition caused by elevated mitotic recombination in Bloom mice. Nat Genet 2000, 26:424-429. 159. Sugawara A, Goto K, Sotomaru Y, Sofuni T, Ito T: Current status of chromosomal abnormalities in mouse embryonic stem cell lines used in Japan. Comp Med 2006, 56:31-34. 160. Shao C, Deng L, Henegariu O, Liang L, Stambrook PJ, Tischfield JA: Chromosome instability contributes to loss of heterozygosity in mice lacking p53. Proc Natl Acad Sci U S A 2000, 97:7405-7410. 161. Donahue SL, Lin Q, Cao S, Ruley HE: Carcinogens induce genome- wide loss of heterozygosity in normal stem cells without persistent chromosomal instability. Proc Natl Acad Sci U S A 2006, 103:11642- 11646. 162. Brown DG, Willington MA, Findlay I, Muggleton-Harris AL: Criteria that optimize the potential of murine embryonic stem cells for in vitro and in vivo developmental studies. In Vitro Cell Dev Biol 1992, 28A:773- 778.
137 163. Gross JB, Hanken J, Oglesby E, Marsh-Armstrong N: Use of a ROSA26:GFP transgenic line for long-term Xenopus fate-mapping studies. J Anat 2006, 209:401-413. 164. Kalin TV, Wang IC, Ackerson TJ, Major ML, Detrisac CJ, Kalinichenko VV, Lyubimov A, Costa RH: Increased levels of the FoxM1 transcription factor accelerate development and progression of prostate carcinomas in both TRAMP and LADY transgenic mice. Cancer Res 2006, 66:1712-1720. 165. Kim JE, Nakashima K, de Crombrugghe B: Transgenic mice expressing a ligand-inducible cre recombinase in osteoblasts and odontoblasts: a new tool to examine physiology and disease of postnatal bone and tooth. Am J Pathol 2004, 165:1875-1882. 166. Kusakabe T, Kawaguchi A, Kawaguchi R, Feigenbaum L, Kimura S: Thyrocyte-specific expression of Cre recombinase in transgenic mice. Genesis 2004, 39:212-216. 167. Chung SS, Cuzin F, Rassoulzadegan M, Wolgemuth DJ: Primary spermatocyte-specific Cre recombinase activity in transgenic mice. Transgenic Res 2004, 13:289-294. 168. Cooley BC: Murine model of neointimal formation and stenosis in vein grafts. Arterioscler Thromb Vasc Biol 2004, 24:1180-1185. 169. Kalinichenko VV, Gusarova GA, Tan Y, Wang IC, Major ML, Wang X, Yoder HM, Costa RH: Ubiquitous expression of the forkhead box M1B transgene accelerates proliferation of distinct pulmonary cell types following lung injury. J Biol Chem 2003, 278:37888-37894. 170. Kohlhepp RL, Hegge LF, Nett JE, Moser AR: ROSA26 mice carry a modifier of Min-induced mammary and intestinal tumor development. Mamm Genome 2000, 11:1058-1062. 171. Narita N, Bielinska M, Wilson DB: Cardiomyocyte differentiation by GATA-4-deficient embryonic stem cells. Development 1997, 124:3755- 3764.
138 172. Bex A, Vooijs M, Horenblas S, Berns A: Controlling gene expression in the urothelium using transgenic mice with inducible bladder specific Cre-lox recombination. J Urol 2002, 168:2641-2644. 173. Nagy A: Manipulating the mouse embryo : a laboratory manual, 3rd edn. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory Press; 2003. 174. Henegariu O, Heerema NA, Lowe Wright L, Bray-Ward P, Ward DC, Vance GH: Improvements in cytogenetic slide preparation: controlled chromosome spreading, chemical aging and gradual denaturing. Cytometry 2001, 43:101-109.
139