information to users - open...
TRANSCRIPT
Selective variegated methylation of thep15/INK4B CpG island is a high frequency
event in acute myeloid leukemia (AML)
Item Type text; Dissertation-Reproduction (electronic)
Authors Dodge, Jonathan Eldon
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this materialis made possible by the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author.
Download date 27/05/2018 16:25:32
Link to Item http://hdl.handle.net/10150/284143
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master UMI films
the text directly from the original or copy submitted. Thus, some thesis and
dissertation copies are in typewriter fece, vwhile others may t)e from any type of
computer printer.
The quality of this reproducticn is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality illustrations
and photographs, print t)leedthrough, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that ttie author dkl rust send UMI a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will irnlicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and continuing
from left to right in equal sections with small overiaps.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6' x 9' black and white
photographic prints are available for any photographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
Bell & Howell Information and Learning 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
800-521-0600
SELECTIVE VARIEGATED METHYLATION OF THE pl5/INK4B CPG ISLAND IS A HIGH FREQUENCY EVENT IN
ACUTE MYELOID LEUKEMIA (AML)
by
Jonathan Eldon Dodge
A Dissertation Submitted to the Faculty of the
COMMITTEE ON PHARMACOLOGY AND TOXICOLOGY (GRADUATE)
In Partial Fulfillment of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2 0 0 0
UMI Number 9972079
<S>
UMI UMI Microfomi9972079
Copyright 2000 by Bell & Howell Information and Learning Company. All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
Bell & Howell Information and teaming Company 300 North Zeeb Road
P.O. Box 1346 Ann Arbor, Ml 48106-1346
2
THE UNIVERSITY OF ARIZONA ® GRADUATE COLLEGE
As members of the Final Examination Committee, we certify that ve have
read the dissertation prepared by Jonathan Eldon Dodge
entitled Selective variegated methylation of the pl5/INK4B CpG island
is a high frequency cMcnt- in acute raveloid leukemia fAMT.">
and recommend that it be accepted as fulfilling the dissertation
requirement for the Degree of Doctor of Philosophy
G. Tim BQii^eAo^ Plj.D. Date
Bernard W. Futscher, Ph.D.
Charl^rte^McQueen,
/ Date
U.~( C -C2£^ (y
Qin Chen, P^D. Date
I. Gle^m Sipes, Ph.D. ^C2Si.
Date
Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.
Dissertation Director Bernard W. Futscher, Ph.D.
4/14- /oi Date
3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial Fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholoarship. In all other instances, however, permission must be obtained from the author.
4
ACKNOWLE DGMENTS
A number of people made significant contributions to this work. Of particular importance are: (1) intellectual and technical contributions, and (2) financial support of my graduate education.
In the first category, the studies originally done by Dr. James G. Herman at Johns Hopkins University provided us with the initial observation of pl5 methylation in AML that stimulated our sequence of studies. Also at Johns Hopkins, Dr. Elizabeth Cameron, a then graduate student in Dr. Stephen B. Baylin's lab helped us establish a protocol for pl5 mRNA detection. An important technical advance was provided by Dr. Vaillancourt's lab, at the University of Arizona, with the their help a pl5 fusion protein was generated and used to establish a protocol for pl5 protein detection. I was fortunate to learn sodium bisulfite sequencing from my advisor Dr. Bernard W. Futscher and then graduate student Dr. George S. Watts. In addition, the studies of pl5 and pl6 introduction into Jurkat were greatly aided by the use of microarray technology developed in Dr. Futscher's lab by current postdoc Dr. George S. Watts. Intellectually, I was helped by my adivsor Dr. Futscher as well as my committee members Dr. Charlene McQueen, Dr. Qin Chen, Dr. G. Tim Bowden, and Dr. I Glenn Sipes.
As a graduate student I was supported financially by: (1) the University of Arizona Graduate College Fellowship in my first year, (2) the Robert S. Flinn Center for Toxicology Fellowship for two years, (3) an NIEHS Center for Toxicology Trainee Grant for two years, and (4) an American Foundation for Pharmaceutical Education (AFPE) Fellowship for three years. These means of support were tremendously helpful and demonstrate the strength of the Pharmacology and Toxicology Department at the University of Arizona.
This work is dedicated to my father and mother, Michael
and Christine Dodge, my brother, Stephen Dodge, and my
sister, Catherine Dodge. They all supported my desire to
understand the world around me.
6
TABLE OF CONTENTS
LIST OF FIGURES 9 LIST OF TABLES 11 LIST OF ABBREVIATIONS 12 ABSTRACT 13
CHAPTER 1 - INTRODUCTION 15 Overview of S-methylcytosine in marranalian cells. . . 15 Dysregulation of 5-methylcytosine in human cancer involves the inactivation of genes 21 Loss of deoxycytidine kinase (dCK) is implicated in drug resistance and dCK is a candidate for inactivation by DNA methylation 23 Potential role of pl5 and pl6 in human cancer. ... 24 Statement of purpose of the present studies 30
CHAPTER 2 - MATERIALS AND METHODS 33 Cell culture 33 Normal cells and patient samples obtained for use in this study 34 DNA isolation and Southern blot analysis 34 RNA isolation and Northern blot analysis 36 Reverse-transcriptase polymerase chain reaction (RT-PCR) 36 Generation of pl5 mRNA by in vitro transcription for use as a positive control 38 Protein isolation and Western blot 39 Generation of pl5 recombinant protein for use as a positive control 40 Isolation of DNA and bisulfite sequencing of dCK, pl5, and pl6 41 Transient transfection with CAT gene reporter constructs 43 Selection of stable expression of pl5 and pl6 in cell lines 43 TGF-I3, PMA, PMA/Ionomycin, cytosine arabinoside, and daunorubicin treatment of cell cultures 45 MTS assay for drug sensitivity 46 FACS analysis of cell cycle by propidium iodide staining, FACS analysis of apoptosis by Annexin V/PI staining, and cell growth in liquid culture 46 Microarray expression analysis 46
7
TABLE OF CONTENTS - Continued
CHAPTER 3 - DETERMINATION OF 5-METHYLCYTOSINE INFLUENCE ON dCK EXPRESSION 51
Drug-sensitive and -resistant model cell lines. . . .51 dCK gene structure and expression 52 DNA methylation mapping of 5' dCK CpG island in model cell lines, normal human peripheral blood lymphocytes, and AML patient samples 58
CHAPTER 4 - DETERMINATION OF 5-METHYLCYTOSINE INFLUENCE ON P15 AND P16 EXPRESSION 64
Established human leukemia cell line models 64 pl5 and pl6 mRNA and protein studies in human leukemia cell lines 66 pl5 promoter studies with the CAT reporter gene. - .70 DNA methylation mapping of pl5 and pl6 5' CpG islands in established human leukemic cell lines, normal human bone marrow, and AML patient samples 72
CHAPTER 5 - INTRODUCTION OF P15 AND P16 PROTEINS INTO JURKAT 81
Stable overexpression of pl5 or pl6 in Jurkat I: establishment of pl5 and pl6 positive subclones. . . 81 Stable overexpression of pl5 or pl6 in Jurkat II: growth is reduced only in pl5 positive cells not pl6 positive cells, retinoblastoma phosphorylation is not constitutively altered, and sensitivity to cytosine-arabinoside, and daunorubicin is unchanged 91 Stable overexpression of pl5 or pl6 in Jurkat III: microarray gene expression analysis of 5700 genes reveals an upregulation of genes involved in cellular differentiation 97
CHAPTER 6 - DISCUSSION 103 Aberrant DNA methylation of the pl5 5' CpG island occurs in the majority of acute myeloid leukemia (AML) patients and might dysregulate normal hematopoietic differentiation 103 Aberrant methylation of pl5 as a marker of AML and potential target for novel drug therapeutic treatments 108 General potential for novel therapeutic intervention of epigenetic defects in human cancer 112
8
TABLE OF CONTENTS - Continued
APPENDIX A. SUMMARY OF STUDIES TO TEST WHETHER P15 CAN BE INDUCED IN KG-1 OR KG-la BY AGENTS NOT ASSOCIATED WITH DNA DEMETHYLATION 115
APPENDIX B. IN PREPARATION TO SUBMIT TO NUCLEIC ACIDS RESEARCH: D-HPLC AS A SIMPLE METHOD TO DETECT ABERRANT DNA METHYLATION 121
REFERENCES 125
9
LIST OF FIGURES
1, Genomic organization of the INK/ARF locus along chromosome 9p21 encoding pl5, pl9ARF, and pl6. . . .27
2, pl5 and pl6 protein homology 28
3, pl5 and pl6 functional pathway 29
4, An overview of the research presented in this dissertation 31
5, Southern blot of dCK in HL60/S, HL60/R, K562/S, and K562/R 53
6, Northern blot of dCK 55
7, Determination of cycle numbers for dCK and histone 3.3 semi-quantitative RT-PCR 56
8, Semi-quantitative RT-PCR for dCK 57
9, Biochemical basis for sodium bisulfite sequencing. .60
10, 5-methylcytosine content of dCK in leukemic model cell lines 61
11, 5-methylcytosine content of dCK in normal human tissue and AML patients 63
12, RT-PCR detection of pl5 and pl6 68
13, Western blots to detect pl5 and pl6 protein 69
14, pl5 promoter constructs with the CAT reporter gene. 71
15, 5-methylcytosine content of pl5 and pl6 in model cell lines 77
16, 5-methylcytosine content of pl5 and pl6 in normal human bone marrow and AML patients, part 1 78
17, 5-methylcytosine content of pl5 and pl6 in normal human bone marrow and AML patients, part 2 7 9
10
LIST OF FIGURES-continued
18, Individual sodium bisulfite clones of normal bone marrow and two AML patients 80
19, PGR to confirm pl5 and pl6 homozygous deletion in Jurlcat 82
20, Explanation of the tetracycline eukaryotic inducible expression system 84
21, Transient transfection of tet-GFP establishes that Tet-Off protein is functional in Jurkat Tet-off. . .85
22, Determination of inducibility of pl5 in Jurkat clones that are pl5 positive 88
23, Western blot analysis of pl5 and pl6 positive Jurkat clones 90
24, Growth of pl5 and pl6 positive Jurkat clones in liquid culture 92
25, MTS assay to determine growth of pl5 and pl6 untreated Jurkat clones 93
26, Annexin V assay for apoptotic/necrotic pl5 and pl6 positive Jurkat cells 95
27, MTS assay to determine effect of araC and daunorubicin sensitivity of pl5 and pl6 positive Jurkat cells. . 96
28, Example of 5,700 cDNA microarray analysis 99
29, Northern blots of differentially expressed genes in pl5 and pl6 positive Jurkat cells 101
30, Summary of microarray versus Northern analysis. . .102
31, Model to explain the observed methylation of pl5 but not pl6 in AML 107
32, Epigenetics/DNA methylation as a potential target of cancer pharmacology Ill
11
LIST OF TABLES
Summary of the cell lines used and tested for pl5 and pl6 mRNA by RT-PCR and pl5 and pl6 protein by immunoblots 32
Characteristics of normal human samples and human acute myeloid leukemia patients obtained for study. 62
Summary of the results obtained using a single plasmid or two separate plasmids in the tetracycline inducible expression system 89
The genes that microarray identified as differentially expressed in the pl5 or pl6 positive Jurkat cells. 100
Summary of the alterations that occur in human cancer cells 114
12
LIST OF ABBREVIATIONS
AML Acute myeloid leukemia ara-C Cytosine arabinoside ATCC American Tissue and Cell Culture 5-aza 5-azacytidine BNIP3 A bcl-2/EiP-interacting gene BSA Bovine serum albumin CAT Chloramphenicol acetyltransferase CDK Cyclin-dependent kinase CpG Cytosine-phosphate-guanosine dinucleotide dCK Deoxycytidine kinase DMEM Dulbecco's modified Eagle's mediiom DOX Doxycycline FACS Flow-assisted cell sorting FBS Fetal bovine serum G418 Neomycin GAPDH Glyceraldehype dehydrogenase GFP Green fluorescent protein HB-EGF Heparin-binding epidermal growth factor Hygro Hygromycin IMEM Iscove's modified Eagle's medium INK4 Inhibitor of cyclin-dependent kinase 4 IRF-1 Interferon-regulatory factor-1 MCLl A gene in the bcl-2 family member MyoD A gene involved in muscle differentiation NBM Normal human bone marrow PAGE Polyacrilimide gel electrophoresis PBL Peripheral blood lymphocytes PBS Phosphate buffered saline PCR Polymerase chain reaction PI Propidium iodide PMA Phorbol-13-myristate acetate PTPNl Protein tyrosine phosphatase RT-PCR Reverse-transcriptase polymerase chain
reaction T-ALL T-cell lymphoblastic leukemia TET Tetracycline TGF-P Transforming growth factor-beta TRE Tetracycline-responsive element
13
ABSTFIACT
We attempted to define target genes that were
inactivated in acute myeloid leukemia (AML) by DNA
methylation. We hypothesized that hypermethylation of 5'
CpG islands is associated with transcriptional silencing of
the corresponding gene and participates in either the
emergence of drug resistance or the conversion of normal
cells to cancer cells. To test this hypothesis the DNA
methylation status of the 5' CpG islands of dCK containing
4 9 CpGs, pl5 containing 80 CpGs, and pi 6 containing 53 CpGs
was determined by sodium bisulfite sequencing of normal
human peripheral blood lymphocytes (PBL) and bone marrow
(NBM), human leukemia cell lines, and cytosine-arabinoside
(ara-C)-resistant adult acute myeloid leukemia (AML)
patients. In PBL and NBM dCK, pl5, and pl6 were all
unmethylated. dCK was unmethylated in the paired ara-C-
sensitive (/S) ara-C-resistant (/R) leukemia cell lines
HL60/S & /R and K562/S & /R, and in the 8 AML patients
analyzed. pl6 was unmethylated in KG-1 and KG-la and both
had detectable pl6 mRNA and protein. None of the 8 AML
patients had aberrant methylation of pl€. For pl5, a
variegated pattern of aberrant methylatlon was found in KG-
1, and complete methylation of pl5 was found in KG-la. The
14
variegated pattern of pl5 methylation seen in KG-1 and the
complete methylation seen in KG-la were both associated
with no detectable pl5 mRNA or protein. pl5 was aberrantly
methylated in 6 of the 8 AML patients, 5 had a variegated
pattern of methylation, and 1 showed complete methylation.
We next introduced ectopic pl5 and pl6 into the pl5 and pl6
negative human T-cell lymphocytic leukemia cell line
Jurkat. The pl5 positive clones grew at a slower rate than
the parent cell or pl6 positive clones as measured by
growth in liquid culture and MTS assay. cDNA microarray
expression analysis differentiated pl5 and pl6 positive
subclones from the parent cell line but not from each
other. This suggests that despite the selective
methylation of pl5 but not pl6 in AML, pl5 and pl6 are
functionally similar.
15
CBASTBR 1 - ZHTRODnCTZON
Ovarvittw of 5-atttiiylcytosi.n* xn •••••• 1i.an cttlls
Genetic change is central to many human pathologies,
however, all human disease is not explained by genetic
change. The remaining changes, those changes that are
heritable and not mediated by a change in DNA nucleotide
sequence, are considered epigenetic changes. The
modification of cytosine to 5-methylcytosine, synonymous
with DNA methylation, is an epigenetic change because it is
heritable but does not have any effect on the triplet amino
acid code contained in DNA.
DNA methylation is the only tolerated modification of
cytosine, guanosine, adenine, and thymine, the four bases
that comprise DNA. Plants, mammals, and fungi methylate a
small percentage of cytosines in their DNA, but the model
organisms yeast, C. elegans, and drosophila lack 5-
methylcytosine (Riggs and Porter 1996). After first
reviewing the enzymatic maintenance of 5-methylcytosine,
the role of 5-methylcytosine in mammalian cell development,
cell differentiation, gene expression, X-chromosome
inactivation, and repression of retroviral sequences will
be discussed.
16
The enzymatic regulation of 5-methylcytosine is not
completely understood. Empirical data support a post-
replication addition of a methyl group to the 5 position of
cytosines located in 5'-CG-3' dinucleotides (Wigler 1980,
Bestor 1996). Evidence supports the almost exclusive
methylation of cytosine located 5' to guanines (Bestor
1996). These 5'-CG-3' are depleted in mammalian genomes
but tend to cluster at the 5' end of genes. These regions
of high 5'-CG-3' density located at the 5' region of genes
are referred to as "CpG islands" that are normally
unmethylated. As defined by Gardiner-Gardner and Frommer
(1987), CpG islands (1) contain > 3 times the normal
density of 5'-CG-3' dinucleotides, (2) are > 200 base pairs
in length, and (3) are located near the transcription start
site of a gene. The first mammalian DNA methyltransferase
discovered, DNMTl, adds a methyl group, using S-
adenosylmethionine (SAM) as a cofactor, to the 5 position
of cytosines in 5'-CG-3' sites. The preferred substrate of
DMNTl is hemi-methylated DNA. Recently, 3 new DNA
methyltransferases have been discovered, DNMT 2, 3a, and 3b
(Okano et al 1998). It is speculated that DNMT 3a and 3b
are de novo methylases, capable of methylating previously
unmethylated regions of DNA. In addition there has long
17
been speculation that an active DNA demethylation process
converts 5-methylcytosine to cytosine independent of DNA
replication. While several reports have been presented
that claim to have found an enzyme with DNA demethylation
activity, some have been retracted and this is still a
controversial area. Bhattacharya et al 1999 has recently
reported an active demethylation activity, but it remains
to be confirmed by other labs.
DNA methylation is implicated in mammalian
development. Mutation of DNMT 1 in mice resulted in
embryonic lethality providing direct evidence for an
essential role of DNA methylation in development (Li et al
1992). This result removed any speculation that 5-
methylcytosine had only a vestigial presence in the
mammalian genome. It has recently been founhd that
knocking out DNMT 3a or 3b (but not DNMT 2) in the mouse
also resulted in embryonic lethality (Okano et al 1999).
Complementing the mouse transgenic experiments, changes in
5-methylcytosine patterns and levels have been observed
during mouse development (Rougier et al 1998). Thus,
direct evidence shows that 5-methylcytosine regulation is
essential to mammalian development.
18
It has also been found that DNA methyiation influences
cellular differentiation. The studies reported in 197 9 by
Taylor and Jones found that the compound 5-azacytidine
resulted in phenotypic alterations in mouse cells lines.
The mechanism of 5-azacytidine's effect on differentiation
was shown to be mediated, at least in part, by inhibition
of DNA methyiation. The altered phenotypes, observed when
Swiss 3T3 or lOTl/2 mouse cells became striated muscle
cells, adipocytes, or chondrocytes, was found to be
associated with an inhibition in the normal methyiation of
DNA. Specifically, in the case of differentiation into
striated muscle cells it was found that the gene MyoDl, a
gene involved in muscle differentiation, was demethylated
by 5-azacytidine and this demethylation was associated with
MyoD transcription (Jones 1985). The discovery of 5-
azacytidine provided a useful pharmacological link between
DNA methyiation and gene transcription.
In 1975 Holliday and Pugh were one of the first to
speculate that DNA methyiation had a direct role in gene
activity. Numerous studies since 197 5 have confirmed that
the presence of DNA methyiation in the 5' CpG islands of
genes is associated with a loss of gene transcription (Bird
1986, Razin and Riggs 1980). One of the earliest studies
19
of a specific gene focused on metallothionein. It was show
that DNA methylation was not only associated with a lack of
metallothionein expression, but heavy metals such as
cadmium did not result in the expected induction of
metallothionein when DNA methylation was present (Compere
and Palmiter 1981). In summary, it is generally believed
that DNA methylation of the 5' proximal CpG islands of
genes is associated with transcriptional silencing of the
corresponding gene.
DNA methylation is implicated in X-chromosome
inactivation in female mammals. Inactivation of one of the
two female X-chromosomes occurs in mammals. In 1975 Riggs
was one of the first to hypothesize that X-chromosome
inactivation involved DNA methylation. This is an
interesting case of an epigenetic modification titrating
gene dosage.. One could imagine a theoretical case where
the female is programmed to extrude the extra X-chromosome.
This does not happen, and it is easy to speculate that
evolving the machinery to actively remove DNA might in the
long run pose more problems than it solves. While current
evidence supports a role for DNA methylation in X-
chromomsome inactivation, the details of how DNA
20
methylation participates in X-chromosome inactivation is
not completely understood (Csankovszki 1999).
Finally, the "host defense hypothesis" put forward by
Bestor states that DNA methylation functions primarily to
repress retroviral genes found throughout mammalian genomes
(Yoder, Walsh, and Bestor 1997). It is well established
that mammalian genomes contain huge regions of repeated DNA
sequences often of retroviral origin. These regions are
usually silent and cytosines in 5'-CpG-3' sites are
methylated. Empirical data show that when expression of
endogenous retroviral sequences occurs it is associated
with a lack of DNA methylation (Walsh et al 1998). Thus, 5-
methylcytosine may have a function in preventing the
expression of both exogenous and endogenous retroviral
sequences.
In conclusion, the base 5-methylcytosine, the only
endogenous alteration of the four bases that comprise DNA,
functions in mammalian development, cell differentiation,
gene activity, X-chromosome inactivation in mammalian
females, and may be important in silencing retroviral
sequences. Germane to the present study, recent advances
link abnormal levels and patterns of 5-methylcytosine to
human cancer (Holliday 1996).
21
Dysragula^on of S-aa-thylcyteaxn* In human eancar Involves
'the Inac^va'fci.on of sp«cl£lc ganas
There are two direct pieces of evidence that 5-
methylcytosine is dysregulated in human cancer: (1) total
5-methylcytosine content in cancer cells is decreased when
compared to normal cells, and (2) regional 5' CpG islands
in cancer cells show increases in methylation levels
(Holliday 1996).
Holliday first postulated in 197 9 that DNA methylation
may be involved in cancer (Holliday 197 9). His argument
stemmed largely from the failure of mutational data to
account for experimental observations of the frequency of
mouse cell transformation into cancer. The proposition
gained support with findings that global DNA methylation
levels were decreased in human cancer when compared to
normal controls. The study of Feinberg and Ehrlich in 1988
showed a decrease in total 5-methylcytosine in colon cancer
when compared to normal controls. It is speculated that
the vast regions of DNA not associated with expressed genes
and containing infrequent 5'-CpG-3' sites are normally
methylated, but in human cancer these cytosine become
demethylated. This might explain the global decrease in
22
DNA methylation observed in human cancer, concurrent to
regional increases in methylation discussed next.
A growing list of human genes that contain 5' CpG
islands have been shown to be inactivated by DNA
methylation in human cancer. Retinoblastoma (pRb) was one
of the first genes where inactivation by DNA methylation
was found to occur in cancer (Ohtani-Fujita et al 1993).
Since pRb is a classic tumor suppressor gene, the idea
emerged that any tumor suppressor that contains a 5' CpG
island may be subject to inactivation by DNA methylation
{Knudsen 1971, Baylin et al 1991). This is an area of
active research and the list of genes that are
hypermethylated in human cancer is still increasing. There
are examples of genes that are inactivated by
hypermethylation in human cancer and alter (1) response to
chemotherapeutics (MGMT), (2) tumor metastasis (E-
cadherin), and (3) cell growth (BRCAl, pl5, pl6, and p73)
(Watts et al 1997, Graff et al 1998, Herman 1999) . In
summary, regional increases in DNA methylation in 5' CpG
islands occur in human cancer in a variety of genes and is
associated with transcriptional silencing of the
corresponding gene.
23
In addition to the absolute changes in 5-
methylcytosine patterns and levels in human cancer, DNA
methylation is also implicated in mutation rates. It is
estimated that one-third of all the observed mutations in
human cancer occur at CpG sites (Denissenko 1997). This
finding links epigenetics to genetics and illustrates the
importance of further research to understand the complete
role of DNA methylation in human cancer. In conclusion, 5-
methylcytosine is definitively dysregulated in human cancer
and these changes influence both the transcriptional
activity of specific genes and the overall mutation rate in
cancer cells.
Loss of daoxycytidln* kinase (dCR) Is io^lxcated In drug
resistance and dCK xs a candidate for inactivetlon by DMA
methylation
Deoxycytidine kinase (dCK) is a gene involved in the
pyrimidine nucleotide salvage pathway. dCK is the rate-
limiting step in the activation of the prodrug cytosine
arabinoside (ara-C), which is used to treat acute myeloid
leukemia (AML). Retroviral introduction of dCK into ara-C-
resistant cells resulted in a ara-C-sensitive phenotype
(Manome et al 1996). While the loss of dCK in AML may
24
represent a mechanism of drug resistance, deletions or
mutations of dCK are infrequent in AML (Flasshove et al
1994) . Two provocative studies, one by Kong et al in 1991
and the other by Antonsson et al in 1987 sugg'ested that 5-
azacytidine, a demethylating agent, reversed drug
resistance in a dCK negative cell line by reactivation of
dCK. These two observations, coupled with the presence of
a 5' CpG island that regulates dCK gene expression, make
dCK a candidate gene for inactivation by DNA
hypermethylation (Chen et al 1995).
Pofcantial rol* of pl5 and pl6 m.m ^uBor si^prassors
The major proteins that regulate eukaryotic cell
cycles have recently been discovered and characterized. It
is beyond the scope of this document to review this area of
research, however, a brief description of one growth
regulatory pathway will be given. Kinases that depend on
the presence of cyclins, referred to as cyclin-dependent
kinases, or cdks, gate the cell cycle. A positive
association has been found between the kinase activity of
cdks and cell cycle progression. How these cdks are
regulated became the focus of gene discovery for much of
25
the 1990s, ultimately linking them to the tumor suppressor
gene retinoblastoma (pEUj) (Weinberg 1995).
pl6 was discovered by Serrano et al in 1993 using a
two-hybrid screen with cdlc4 as the bait. When pl6 bound
cdk4 or cdk6 it was shown to inhibit cdk4/6 kinase
activity. Subsequent to pl6's discovery, pl5 was
discovered by Hannon et al and Jen et al in 1994. pl5 and
pl6 lie adjacent to each other along chromosome 9p21. This
region of DNA produces three potential tumor suppressor
genes, pl6/pl9ARF, where pl9ARF is generated by the use of
a different exon 1 but the same exon 2 and 3 as pl6 (pl9
uses a different reading frame than pl6, hence pl9ARF =
Alternative Reading Frame), and pl5. The pl9ARF gene
product interacts in the p53 tumor suppression pathway by
binding to MDM2. Figure 1 illustrates the genomic
organization of pl5, pl6, and pl9ARF.
pl5 and pl6 share >80% protein homology and both bind
to and inhibit cyclin-dependent kinases 4 and 6 (cdk4/6).
pl5 and pl6 are 2 of the 4 members of the INK4 family
(INhibitor of cyclin-dependent Kinase ±) that also includes
pl8 and pl9. The protein homology between pl5 and pl6 is
shown in figure 2. In addition to the INK4 family, the
CIP/KIP family contains p21, p27, and p57 and each gene
26
functions to inhibit cdk2 paired with a specfic cyclin
(whereas the INK4 family only inhibits cdk4/6, bound to D-
type cyclins 1,2, or 3). The pathway in which pl5 and pl6
are thought to function is depicted in figure 3.
Ultimately it is believed that pl5 and pl6 prevent the
phosphorylation of pRb which in turn prevents the release
of the transcription factor E2F, and E2F dependent
transcription of S-phase genes.
pl5 and pl6 were originally implicated in human cancer
because both lie on a stretch of chromosome 9p21 frequently
deleted in human cancer (Kamb et al 1994). Further, pl6
mutations were shown to occur in familial melanoma
(Hussussian et al 1994) . While deletion of pl5 occurs in
human cancer, it is only rarely a pl5 exclusive deletion
that does not also delete pl6/pl9ARF, and mutation of pl5
has not been observed in any human cancer. Nevertheless,
pl5 and pl6 are both subject to inactivation by
hypermethylation in hematological malignancies.
Specifically, Herman et al 1997 used Southern blots to show
that pl5 but not pl6 is frequently hypermethylated in AML.
Thus, pl5 may have a role in tumor supression in AML
despite an unaltered pl6 gene.
Figure 1. A schematic map of the genomic structure of the region of DNA on human chromosome 9p21 that encodes pl5, pi6, and pl9ARF. The green boxes depict exons 1 and 2 of pl5, the blue boxes depict exons la, 2, and 3 of pl6, and the red box depicts exons ip that comprises pl9ARF along with the blue boxes of exons 2 and 3. The approximate locations of the the three 5' CpG islands is shown in the yellow boxes.
28
NHAR
pl5
pl6 IDAAEGPSDIPD*
Figure 2. The protein homology between pl5 and pl6 is shown as the gray shaded regions. The red shading indicates amino acids in pl6 that bind to cdk6 and are mutated in human cancer, resulting in a loss of pl6 function. Surprisingly, the red shaded amino acids of pl5 are not mutated in human cancer.
p15/p16
6
29
Figure 3. The pathway that pl5 and pl6 participate in is shown in this cartoon. Both pl5 and pl6 bind to cdk4 and cdk6 and inhibit the kinase activity of cdk4 and cdk6. In the absence of pl5/pl6 unregulated phosphorylation of retinoblastoma (pRb) promotes entry of the cell into S-phase.
30
StateBMn^ of puxpoum ot pr«s«nt steditts
The present study was designed to define target genes
that are hypermethylated in AML (figure 4, next page, shows
a summary). We hypothesized that hypermethylation of 5'
CpG islands is associated with transcriptional silencing of
the corresponding gene and participates in either the
emergence of drug resistance or the conversion of normal
cells to cancer cells.
Since dCK was known to have reduced activity in
cytosine-arabinoside (ara-C)-resistant leukemia cell lines,
and no deletion or mutation accounts for the loss of dCK
activity, dCK represents a mechanism of drug resistance and
is a candidate gene for inactivation by hypermethylation.
It was known from Southern blot data that pl5 but not
pl6 is methylated in AML. Both pl5 and pl6 are thought to
be tumor suppressor genes, genes capable of preventing the
conversion of normal cells to cancer cells. pl5 and pl6
are adjacent to each other, have >80% amino acid identity,
and yet only pl5 is hypermethylated in AML. In table 1,
page 33, the cell line models studied are summarized. We
used this situation to describe the methylation of pl5 and
try to understand why such a situation arises in AML.
31
Identification of Candidate Genes that are hypermethyiated in AML
dCK methylated dCK not methylated (Chapter 3)
p15 or pi 6 methylated, p15 methylated (Chapter 4)
pi 5 or pi 6 not methylated
Introduction off pi 5 or pi 6 into Jurkat (Chapter 5)
Figure 4. An overview of the experiments presented in this dissertation. Initially we determined the methylation status of deoxycytidine kinase (dCK) in cell line models and AML patients, (Chapter 3). Once a lack of dCK methylation was demonstrated, the focus shifted to pl5 and pl6 methylation methylation analysis (Chapter 4) . When pl5 methylation of AML cell lines and patients samples in the absence of pl6 alterations was detected, the final Chapter focused on the functional contribution of each (Chapter 5).
32
KG-1 AML No detectable pl5, but pl6 detectable
KG-la AML No detectable pl5, but pl6 detectable
HL60 AML pl5 detectable
U937 AML pl5 detectable
HaCaT Keratinocyte Positive control for pl5 detection
RHEK-1 Keratinocyte Positive control for pl6 detection
Jurkat T-ALL Negative control for both pl5 and pl6 detection
Melanoma UACC 612
Melanoma P15 detectable, but no detectable pl6
Table 1. A summary of the names and types of the cell lines used to determine pl5 and pl6 mRNA levels by RT-PCR and protein levels by immunoblots.
33
CHAPTSR 2 - MXTXRXALS AND METHODS
Call cultux*
HL60-sensitive and HL60-daunorubicin resistant cells
were a gift from Dr. K. N. Bhalla at Emory University, and
K562-sensitive and K562-daunorubicin resistant were a gift
from Dr. S. Grant at Virginia University (Bhalla et al
1984, Grant et al 1995). Jurkat, U937, KG-1, and KG-la
were obtained from ATCC. HaCaT was established by Dr.
Fusenig and we obtained it from Dr. Bowden's lab at the
Arizona Cancer Center (BouJcamp et al 1998). HL60/S,
HL60/R, K562/S, K562/S, U937, and Jurkat were cultured in
RPMI 1640 supplemented with 5% (Jurkat) or 10% (HL60/S,
HL60/R, K562/S, K562/S, and U937) FBS. KG-1 and KG-la were
cultured in IMEM supplemented with 20% FBS. Melanoma
UACC612 was established from a patient with melanoma at the
Arizona Cancer Center core facility. RHEK-1 was obtained
from Dr. Johng S. Rhim (Rhim et al 1985) . HaCaT and RHEK-1
were cultured in DMEM supplemented with 10% FBS All
cultures were maintained in penicillin (100 U/ml), and
glutamine (1% vol/vol) (JRH Biosciences, Lenexa, KS) and
incubated in 95% air-5% C02 atmosphere at 37° C.
34
Normal c«lls and pa^ant mmmplmm obtaxnad for usa in tbxa
study
Peripheral blood or bone marrow specimens were
obtained from healthy allogeneic bone marrow donors and 8
AML patients after informed consent (see table 1 for more
information). From these samples low density mononuclear
cell fractions were isolated by double Ficoll-Hypaque
density centrifugation (Pharmacia, Uppsala, Sweden) or by
the Vacutainer CPT cell preparation system in the case of
the normal PBLs (Becton Dickinson, Franklin Lakes, NJ).
Isolation of AML patient specimens was performed by Dr.
Alan List's lab at the Arizona Cancer Center and yielded
>85% blasts as determined by cytologic examination.
DHA xsola1:i.on and Southam blot analysxs
DNA was isolated with a QIAamp blood kit isolation kit
according to the manufacturer's protocol. The DNA was cut
with appropriate restriction enzymes (SU/fXL) at 37° C. The
digested DNA was phenol/chloroform purified, precipitated
with 0.1 volume 3M sodium acetate and 2.5 volumes 95%
ethanol, and pelleted by 30 minute spin on a microfuge at
maximum speed at 4° C. The DNA was resuspended and
quantitated spectrophotometrically by OD at 260 nm. 10 (ig
of cut DNA was size separated by gel electrophoresis on a
1% agarose gel. The DNA was de-purinated with 0.2 M HCl
35
for 15 minutes, denatured with a solution of 0.5 M NaOH and
1.5 M NaCl for 30 minutes, and neutralized with a pH 7.4
solution of 0.5 M NaOH, 55 mM Trizma-HCl for 30 minutes.
The DNA was then capillary transferred in lOX SSC (3M NaCl,
0.3M sodium citrate) from the gel to a 0.45 nm Nytran
membrane overnight. The membrane and DNA were crosslinked
by UV light, sealed in a membrane with ~5 mL pre-
hybridization solution (50% formamide, 5X SSC, IX
Denhardt's solution [lOOX Denhardt's solution is: 20 mg/mL
Ficoll 400, 20 mg/mL polyvinylpyrrolidone, and 20 mg/mL
bovine serum albumin fraction five], 1% sodium dodecyl
sulfate, 10% dextran sulfate, and 250 ng/mL salmon sperm
DNA) followed by incubation overnight at 42° C with
shaking. radio-labeled DNA probe of the dCK cDNA was
then generated by random primer labeling, briefly -25 ng
template DNA was labeled using the Gibco BRL Radprime kit
with 5 |iCu of a-dCT^^P. The labeled probe was boiled for 5
minutes and added to the pre-hybridization solution and
allowed to hybridize overnight at 42° C. The membrane was
washed at room temperature in 2X SSC, 0.5% SDS for 30
minutes, then 2 washes in 0.IX SSC, 0.5% SDS at 55° C were
performed. The membrane was exposed to x-ray film with
intensifying screens at -80° C.
36
RMA isolation and Northttxn blot analysis
RNA was isolated with Qiagen Rneasy kit according to
the manufacturer's protocol. The isolated RNA was then
quantitated with a spectrophotometer by absorbance at 260
nm. 20 (ig of RNA in ~11 |i.L of H2O was denatured with 4 (iL
lOX MOPS buffer, 7 |iL 37% formaldehyde solution, and 16 }iL
de-ionized formamide and heated at 55° C for 15 minutes.
2 (iL of RNA loading buffer (50% glycerol, 1 mM EDTA, 0.25%
bromophenol blue) were added, and the RNA then loaded onto
a denaturing formaldehyde/agarose gel (IX MOPS buffer, 18%
vol/vol 37% formaldehyde, 1% agarose). The gel running
buffer was IX MOPS with ethidium bromide added to visualize
the RNA. After electrophoretic separation at 60 volts, the
formaldehyde was leached out of the gel by one wash in IX
MOPS for 15 minutes, followed by two 15 minute washes in
DEPC-treated H2O. RNA was capillary transferred to a 0.45
|im Nytran membrane as described in Southern blots above.
The membrane and RNA were UV crosslinked, and then the pre-
hybridization and hybridization were performed as described
for Southern blots above.
Revarsa-transeriptas* polyaarasa chain raaction (RT-PCR)
Total cellular RNA was isolated using the RNeasy minikit
(Qiagen) and quantitated using a spectrophotometer
measurement at 260 nm. For pl5 and pl6 33 cycles were
37
used, and for histone 3.3 25 cycles were used. The primers
used for pl5, pl6, and histone 3.3 (H3.3): were
dCK 5' (5'-CATTTGTGAATATCCTTAAAC-3'),
dCK 3' (5'-ATAAATGATTCCATCCAATTCA-3') ,
p15 5'2 {5'-TGGGGGCGGCAGCGATGAG-3'),
pl5 3' (5'-TGGGTAAGAAAATAAAGTCGTTG-3') ,
pl6 5' (5'-AGCCTTCGGCTGACTGGCTG G -3'),
pl6 3' (5'-CTGCCCATCATCATGACCTGGA-3'),
H 3 . 3 5' (5'-CCACTGAACTTCTGATTCGC-3'),
H3.3 3' (5'-GCGTGCTAGCTGGATGTCTT-3').
These reactions were carried as follows: 2 pg of total
cellular RNA was reverse transcribed by incubating a 20 pi
reaction mixture containing IX PGR buffer (10 mM Tris [pH
8.3], 50 mM KCl, 6 mM MgCl2), 2 mM each dATP, dCTP, dGTP,
and dTTP, 200 pmol random hexamer, 20 U Rnasin (Roche), and
200 U Superscript II reverse transcriptase (Gibco BRL) at
42*^0 for 60 min. The reaction was stopped by incubation at
99^C for 10 min. Analysis of pl5 and pl6 expression was
performed by adding 10 pi of the RT reaction to 40 ul of
amplification reaction mix (IX PGR buffer, and 1.25 U of
Tag DNA polymerase) followed by incubation at 98*^0 for 5
min; 33 cycles of 98^0 for 30 s, 56^C for 30 s, 12^C for 1
min; and a final extension at 12^C for 5 min followed by a
quick chill to 4®C. Analysis of Histone 3.3 was performed
by adding 1 pi of the RT reaction to 4 9 pi of amplification
38
reaction mix (IX PGR buffer, and 1.25 U of Taq DNA
polymerase) followed by incubation at 94^C for 5 min; 25
cycles of 94^C for 1 min, 56^C for 30 s, 72®C for 1 min;
and a final extension at 7 2^C for 5 min followed by a quick
chill to 4^C. For analysis, 20% of the respective PGR
products were separated on 2% agarose gel, visualized by
ethidium bromide staining with the Eagle Eye II Still Video
System (Stratagene, La Jolla, Ga.). In the case of semi-
qunatitative RT-PGR for dGK and His 3.3, the desired cycles
were performed and 20% visualized and staining intenisty
quantitated by ethidium bromide staining with the Eagle Eye
II Still Video System.
Genera^on o£ pl5 aRHIi by in vlteo fcranscripti.on for umm mm
a posx'ti.v'* control
The full length coding sequence of pl5 was obtained
from normal human bone marrow by 35 cycles of RT-PCR as
described above with primers pl5 3' and pl5 5'1, (5'-
CAGGAAAAGGCGGGAGGTAAGGA-3'), and ligated into pGEM T-easy.
The T7 primer generated sense pl5 RNA using the Boehringer
Mannheim in vitro transcription kit according to the
manufacturers protocol. The resulting pl5 RNA was then
used to establish the RT-PGR protocol, and in subsequent
RT-PGR experiments the pl5 cDNA was used as a positive
control.
39
Prot«xn xsola^on and lUmutmxn blo^
Total protein was obtained by treating cells with
lysis buffer (50 mM Tris-HCl, 250 mM NaCl, 0.5% NP-40, 50
mM NaF) containing protease inhibitors (1 mM PMSF, 20 Dg/Dl
N-TPCK, 20 mg/ml STI, 10 mg/ml leupeptin, 0.066 TI units/ml
aprotinin) for 60 minutes on ice. Lysates were spun for 5
minutes at 13,000 rpm to remove debris. Supernatants were
taken for protein quantitation using BSA standards (Pierce,
Rockford, IL). In the case of TGF-B treated cultures,
exponentially growing cultures were treated with 2ng/mL
TGF-Bl for 24 hours (Sigma, St. Louis, MO). After
quantitation, 0.25 mg of protein (except HaCaT -/+ TGF-I5
where 0.1 rag was used) from each cell line was denatured in
SDS-PAGE sample buffer (1 ml glycerol, 0.5 ml 2-
mercaptoethanol, 3.0 ml 10% SDS, 1.25 ml 1.0 M Tris-HCl pH
6.7, 1 mg bromophenol blue) and the samples were subjected
to SDS-PAGE on a 15% pre-cast mini gel (Bio Rad, Hercules,
CA). The resolved proteins were transferred onto a
Immobilon-P membrane by electrophoretic transfer at 60
Volts for 4 hours at 4° C (Millipore, Bedford, MA). The
nonspecific sites were blocked by incubation in 5% nonfat
dry milk in Tris-buffered saline and 0.05% Tween 20 (TEST)
for 1 hour at room temperature. The membrane was then
incubated with pl5 antibody (sc-612 Santa Cruz, CA) 1:50
40
dilution in TBST or pl6 antibody (sc-468 Santa Cruz, CA)
1:100 dilution for 1 hour at room temperature. The
membrane was washed with TBST and incubated with
horseradish-peroxidase-conjugated immunoglobulin G
secondary antibody (Jackson Immunoresearch, West Grove, PA)
for 1 hour at room temperature. Following 3 quick rinses
and 3 15 minute washes with ~10 mL TBST all at room
temperature, detection by chemiluminescence was performed
with the Renaissance Kit (DuPont NEN, Boston, MA) and
exposed to film (XAR-5, Eastman Kodak Co., Rochester, NY).
6«nerat:xon of pl5 rscfbi nanfc protaxn for us* as a posx^va
control
The pl5 cDNA was obtained from normal human bone
marrow by RT-PCR as described above. The pl5 cDNA was
subcloned with two EcoRI sites into the pMAL-c2 vector (New
England Biolabs, Beverly, MA). The resulting vector
pl5MAL-c2 was sequenced to confirm its protein coding
sequence and orientation. pl5MAL-c2 was transfected into
competent E. coll (Invitrogen). A log-phase culture was
used to column purify pl5 and the resulting isolate was cut
with factor Xa to yield pl5 protein.
41
BxauXfxt* ••qu«nexn9 of dCK, pl5, and pl6
DNA was isolated from the cell lines, normal tissue
specimens, and AML patient samples (as previously
described) using the QIAamp blood kit (Qiagen, Santa
Clarita, CA), and quantitated spectrophotometrically at 260
nm. Sodium bisulfite conversion followed the protocol of
Clark et al 1994. Briefly, DNA was denatured in 0.3 M NaOH,
reacted with 3.6 M sodium bisulfite (pH 5.0) at 55° C for
14 hours, desalted with the Wizard Prep kit (Promega,
Madison, WI), desulfonated with 0.3 M NaOH, and ethanol
precipitated. The resulting bisulfite-modified DNA was
then amplified using nested PGR. The first reaction used
-500 ng of DNA with IX PGR buffer, 200 uM dNTPs, 50 pmol of
either Al & A2 (1st round) or B1 & 32 (2nd round), and 2.5
U Taq polymerase (Roche, Indianapolis, IN). The 2nd round
of PGR used 1-10% of the 1st round reaction under identical
conditions. The following bisulfite specific primer sets
were used,
dCK 5' : dCKUl (5'-TTTGTTTATTTTTAATAGGTTTATTAGAG-3' )
dGKDl (5•-AATCCCGTGAAAACTAACTAAAAAAAA-3')
dGKU2 (5'-GGTTTTTGGGGTTTATTTTTTTTTTT-3')
dGKD2 (5'-AAATAAGGATTCGTTAATCTTATA-3')
pl5 5' : P155U1 (5'-TTGTATGATAGGTGTAGAGTTG-3')
P155U2 (5'-GTATTGTTTGGGGATGAATTTAA -3')
P155D1 (5'-GTGTAACAAAATAAAAAAGGAAG-3')
42
P155D2 (5'-AAATCCTCACTACCCCICCTCIC-3' )
pl5 3^: P153U1 (5'-GGTTTTTTGGTTTAGTTGAAAA-3')
P153U2 (5'-AACITAAACTCCTATACAAATCTACAT-3')
P153D1 {5'-TGTIGGTTGGTTTTTTATTTTGTTAGA-3')
P15 3 D2 ( 5 ' -AAACCCTAAAACCCCAACTACCTAAAT - 3 ' )
pl6 5": P165U1 (5'- TTAGTIGTATAGGTGATTTIGATTT -3')
P165U2 (5'-GGTTGAGGGGGTAGGGGGATAT-3')
P165D1 (5'-CCCCACCCTCTAATAACCAAC-3')
P165D2 (5'-TACCCCTCTCCICAACCICCAAAC-3' )
pl6 3' : P163U1 (5 *-GGTTGGTTGGTTATTAGAGGGTGG-3')
P163U2 ( 5 ' -TACAAACCCTCTACCCACCTAAAT-3 ' )
P163D1 (5'-GTTGGTTATTAGAGGGTGGGG-3')
P163D2 {5'-CTCCIACCITAACTATTCIATAC-3')
Amplifications were performed as follows: 95° C 1 min,
followed by 35 cycles of 92° C 1 min, ramp 1.0° C/sec to
56° C, 56° C 3 min, 0.3° C/sec to 72° C, 72° C 1 min, 0.5°
C/sec to 92° C. There was a final extension at 72° C 5 min
and a quick chill to 4° C. All PCRs were performed on a
DNA engine (MJ Research, Watertown, MA). The PCR products
were ligated into the vector pGEM T-easy (Promega) and
transfected into competent E. coli (Invitrogen, San Diego,
CA). Plasmid DNA was isolated from the resultant bacterial
colonies using a QIAprep Spin Plasmid Miniprep kit
(Qiagen), and positive recombinants were sequenced on an
ABI automated sequencer. The resulting sequences were
43
aligned to their respective sequences using AssembyLIGN
software. The number of methylated CpGs at a given site
was divided by the number of clones analyzed (n = 5-11,
except HL60 pl6 where n = 2) to yield a percent
methylation.
Transi«n^ transfttc^on wxth CAT gan* r«port«r cons^ructe
HaCaT cells were plated 24 hours prior to transfection
at ~50% confluency, and transfected with lipofectamine
(Gibco BRL). The wild-type pl5 promoter region was
amplified from normal bone marrow and three regions were
then subcloned into pCAT with Pst I sites after first
adding a Pst I site into the 5' primer and initial TA
cloning into pGEM T-easy. Four micrograms of each
construct were used per transfection and an SV40f3GAL
plasmid served as a control. Activity was detected 48
hours after transfection with CAT and BGAL ELISA assays
(Roche), and total protein was quantitated with BSA
standards (Pierce).
Selec^on of stabl* •xpr«ssion of pl5 and pl6 in coll linos
Jurkat Tet-Off cells and tetracycline inducible
expression system were purchased from Clontech. The pl5
cDNA was subcloned into the EcoRI site of the pTRE vector.
The resulting vector plSpTRE was sequenced. In addition.
44
the hygromycin resistance gene was subcloned from the
pTKHyg vector into the plSpTRE and pl6pTRE vectors with
Xhol sites. The resulting vector, plSpTREpHyg,
pl6pTREpHyg, or pTKHyg (as control for the effects of
hygrom.ycin resistance) were then used for transfections.
Transfection of 1 x 10^ Jurkat Tet-Off cells with 2 fig of
pTKHyg, plSpTREpHyg, or pl6pTREpHyg, and in the case of
double plasmid selection, 2 ̂ g of pTKHyg and 40 |ig plSpTRE
were performed on a Biorad gene pulser set to 250 mV and
960 (IF. The cells were cultured without selection for two
days after electroporation, then bulk selected with 400
(ig/mL hygromycin (Clontech) for 3 days, and finally clonal
populations were obtained by limiting dilution in 96-well
plates. The clonal populations were tested for pl5 or pl6
DNA and RNA presence by PGR or RT-PCR.
45
The individual clones were obtained from plates with
from 4-15 clones per 96-wells. Poisson statistics yield a
probability with the following mathematical argument.
Assuming 9.6 clones/96-well plate, 1 in 10, X = 0.1,
H = Xv, describes the total occurance, where V is the volume
of the well, in this case V = l, soM.= 0.1. Then the
probability, Pn, can be expressed as:
Pn = (i^e-^'/N!, where N = the number of cells that form a
colony, N = 1 for the desired clonal colonies, from this
the probabilities are: (1) no colony forms = 90.5%, (2) a
clonal colony forms = 9.05%, and (3) two cells initiate a
colony = 0.005%. Thus the average 96-well plate yielded 10
colonies which was a > 1000-fold more likely generated by
one cell than by two.
T6F-B, PIA, PIA/IonoHYcxn, eytesln* arablnosld* (ara-C) and
daunorubicin traafant of call cultuxaa
TGF-13, PMA, and lonomycin were all solubilized
according to the manufacturers recommendation (Sigma).
Cytosine arabinoside and daunorubicin were obtained from
the tumor cloning lab at the Arizona Cancer Center. The
46
desired doses were added directly to the cell cultures and
incubated at 37° C.
HTS assay for drug s«nsi.^vi.ty
The cells were counted and plated at 8,000 cells/well,
100 |J.l/well on day 1. The appropriate dose of drug was
delivered 24 hours later on day 2. At day 6 the MTS kit
(Promega) was used according to the manufacturer's protocol
and results were read by absorbance at 470mn.
FACS analysis of e«ll cycla by propidiuB lodxd* s^inxng,
FACS analysis of apoptosls by Annaxln V/PI staining, and
cell groirth In liquid cultura
Cells were washed with PBS, and resuspended in
Krishnan's buffer containing 50 ng/mL propidium iodide, 20
(ig/mL RNase, 0.3% NP-40. Annexin V (Boehringer Mannhiem)
and prodium iodide were added to washed cells according to
the manufacturer's protocol and were analyzed by FACS. The
core facility at the Arizona Cancer Center performed FACS
analysis on both samples. Growth in liquid culture was
performed by plating equal cell numbers in 6-well plates
and counting cells daily with a cytometer.
Mlcroarray axprasslon analysis
Probe Preparation: The human cDNA bacterial clones
used to fabricate the microarray came from Research
47
Genetics (Huntsville, AL). Referred to as gf200, the set
consists of 5,184 sequence-validated IMAGE consortium
bacterial clones, approximately 3,000 of these clones
represent known genes while the remainder are unknown ESTs.
cDNA targets were produced by PGR of the cDNA inserts
directly from bacterial cultures. In brief, individual
IMAGE clones were grown in 96-well plates at 37 °C for 6
hours. One microliter of the bacterial culture was added
to a PGR reaction and the cDNA inserts amplified in a 96-
well format using a PGR primer pair (Research Genetics)
designed to selectively amplify the cDNA inserts contained
within the IMAGE clones. Primers and unincorporated
nucleotides were removed from the PGR products using a 96-
well PGR clean up kit from Qiagen. PGR amplification and
clean up were verified by agarose gel electrophoresis, and
PGR product yield was determined using a PicoGreen-based
fluorescence assay (Molecular Probes, Eugene, OR) in a 96-
well format. After quantitation the purified PGR products
were dried and re-suspended at 300 ng/)jl in 2X SSG for
printing onto slides.
Microarray Fabrication: The purified, PGR-amplified,
cDNAs were printed onto chemically-activated glass slides
48
(CEL industries, Houston, TX) using 4 quill-type pins
(Telechem International, San Jose, CA) mounted onto an
OmniGrid robot (GeneMachines, Menlo Park, CA). In addition
to the 5,184 IMAGE clones, a set of housekeeping genes
{Research Genetics), a set of Mesembzyanthemum
crystallinum (ice Plant) genes, and Cy3 or Cy5 (Amersham)
labeled oligonucleotides were placed strategically to aid
in data normalization, measuring non-specific
hybridization, and locating the corners of the array.
Finally, a set of 103 IMAGE clones representing known genes
that were not included in gf200 were purchased from
Research Genetics and included in the microarray.
Following printing, the slides were placed in a humidity
chamber overnight in the dark to re-hydrate the spots of
cDNA probes. The next day the slides were washed one minute
in 0.1 % sodium dodecyl sulfate (SDS) and one minute in
double distilled water at room temperature. Slides were
then submerged in a 75% v/v water/ethanol solution into
which 0.6 g sodium cyanoborohydride had been dissolved for
five minutes at room temperature followed by four washes in
double distilled water for two minutes each. Slides were
then spun dry at 500g and stored in the dark at room
temperature until use.
49
Target Preparation: poly A+ RNA was isolated from ~50
X 10^ cells using the midi/maxi oligotex direct kit
according to the manufacturer's protocol (Qiagen).
Fluorescent first strand cDNA was made from 4 ng of poly A+
RNA in the presence of 50 HM Cy5-dCTP (green) or Cy3-dCTP
(red) in a 25 |J.l volume containing the following: 500 ng
oligo (12-18) dT, 1 X Superscript Buffer, 400 U Superscript
II, 3.3 U RNAse inhibitor (all from Gibco BRL), 400 |iM each
of dGTP, dATP, dTTP, 100 HM dCTP, and 10 mM dithiothreitol.
All reagents except the Superscript II were mixed on ice,
placed at 65 °C for five minutes, and 25 °C for 5 minutes at
which point the Superscript II was added and the mixture
heated to 42 °C for two hours. The mRNA was then degraded
by heating the reaction for five minutes at 95 °C, adding
6.25 fJ-l of 1 M NaOH, and a final 10 minutes incubation at 37
°C. Labeled cDNA from two reactions (Cy3 or Cy5 labeled)
was purified on a microcon-50 column using four buffer
exchanges (the first three were water, the final exchange
was 10 mM Tris-HCl, pH 8.5), eluted from the column,
lyophylized to dryness, and re-suspended in 10 (4.1
hybridization buffer (2 X SSC, 0.1% SDS, 100 ng/^l Cotl DNA,
50
100 ng/^l oligo dA), denatured by boiling for 2.5 minutes,
and added a denatured (boil slide 2 minutes, plunge into
room temperature ethanol, and spin dry at 500g) microarray.
A cover slip (22 x 22 mm) was used to spread the target
evenly over the probes and the array was placed in a
hybridization chamber (Telechem International) at 62 "C for
18 hours. Following hybridization, slides were washed by
placing them into 50 mL conical tubes containing 2 x SSC,
0.1% SDS for five minutes, 0.06 x SSC, 0.1% SDS for five
minutes, and 0.06 x SSC for two minutes all at room
temperature. Quantitative fluorescent emissions were
collected using a Scan Array 3000 microarray reader
(General Scanning Inc.) and quantitated using the Imagene
software (version 2.0, Biodiscovery Inc.). Three
independent pl5 or pl6 positive Jurkat clones were arrayed
against the pl5 and pl6 negative parent cell line and
cumulative results were used to identify differentially
expressed genes.
51
CHAPTKR 3 - DSTBRMZHATZOT OF S-MBTHTLCTTOSZIIE
INFLUENCE ON dCR EXPRESSION
Dnig-sttnsx^v* and -rMlstanfc modml c«lX IxnM
The hypothesis of our initial efforts (see figure 4
for a complete overview of the research presented in this
dissertation) was that DNA hypermethylation is associated
with decreased dCK expression and this loss of dCK
expression participates in the drug resistant phenotype.
In an effort to better understand the role of dCK
expression levels in the emergence of resistance to
cytosine arabinoside (ara-C), a drug used to treat acute
myeloid leukemia (AML), the paired ara-C-sensitive (/S) and
ara-C-resistant (/R) cell lines, K562/S and K562/R, and
HL60/S and HL60/R were obtained. K562/R was reported by
Grant et al (1995) to contain 70% less dCK activity than
the sensitive parent cell line, and Bhalla et al reported
(1984) that HL60/R was deficient in dCK activity. These
reports showed that loss of dCK activity is central to the
araC-resistant phenotype. Further evidence in the
literature suggests that exposure to 5-azacytidine, an
inhibitor of DNA methylation, re-activates dCK in drug-
resistant cells, supporting the hypothesis that DNA
52
methylation mediates the loss of dCK expression in drug-
resistant leukemia cells (Kong et al 1991, Antonsson et al
1987) .
dCK gene atxuctnxm and •a^pr^ssion in aodal c^ll lin«s
In order to determine if an association between dCK
and ara-C-resistance existed we first performed a Southern
blot using the complete dCK cDNA as a probe. It might be
the case that one or both alleles containing dCK are
deleted or rearranged in the ara-C-resistant cell lines,
and this would explain the loss of dCK activity independent
of DNA methylation changes. The Southern blot, shown in
figure 5, did not reveal any gross changes in dCK gene
structure. The dCK cDNA probe showed similar intensity
and banding patterns in normal human peripheral blood
lymphocytes (PBL) and in the paired ara-C-resistant model
cell lines, K562/S and K562/R, and HL60/S and HL60/R.
These results indicate that the reported loss of dCK
activity is not mediated by deletion or rearrangement of
dCK at the DNA level.
53
9.42Kb
M 1 2 3 4 5 6 7 8 9 1 0
m m
2.03Kb
0.56Kb
Figure 5. Southern blot using the dCK cDNA as a probe with EcoRl digested genomic DNA. Genomic DNA was cut with EcoRl, recovered, and 10 (ig/lane was size separated by gel electrophoresis on a 1% agarose gel. The lane designated 'M' = Hindlll digested lambda DNA and served as a marker, the sizes in kilobases (Kb) are indicated on the left with arrows. Lanes 1-5 correspond to PBL, K562/S, K562/R, HL60/S, and HL60/R of the ethidium bromide stained agarose gel prior to transfer to a nytran membrane. Lanes 6-10 correspond to PBL, K562/S, K562/R, HL60/S, and HL60/R of the nytran membrane blot probed with the ^^P-labeled dCK cDNA.
54
The mRNA levels of dCK were determined in the model
cell lines by two techniques. Northern blot and semi
quantitative RT-PCR. A representative Northern blot is
shown in figure 6 (n of 3 independent blots were performed
for dCK and GAPDH). Expression levels derived from the
Northern assay were normalized to GAPDH expression levels.
Northern analysis revealed an approximate 50% decrease in
dCK mRNA content in K562/R and HL60/R when compared to
K562/S and HL60/S respectively. For semi-quantitative RT-
PCR, various cycle numbers for both dCK and Histone 3.3
were performed to determine the linear range of PGR
amplification, shown in figure 7. The results were
quantitated by ethidium bromide staining of a 2% agarose
gel and detection of staining intensity using the
Stratagene Eagle-eye, shown in figure 8. The results
obtained with semi-quantitative RT-PCR also indicated an
approximate 50% decrease in dCK mRNA in both the K562 and
HL60 ara-C-resistant cells.
55
Figure 6. Northern blot for dCK. 20 \iq of total RNA was size separated by gel electrophoresis on a 1% agarose/ formaldehyde gel, ethidium bromide staining prior to transfer of the gel to a nytran membrane is shown on the right. After the gel was transferred to a nytran membrane, the membrane was probed with the ^^P-labeled dCK cDNA and then GAPDH. Lanes 1 -4 correspond to K562/S, K562/R, HL60/S, and HL60/R. Lanes 5-8 correspond to K562/S, K562/R, HL60/S, and HL60/R, and show the ethidium bromide staining of the 28S and IBS ribosomal RNA bands prior to transfer of the gel to a nytran membrane.
56
Btixidium bromide staxnxng intensity
le+5
le+4 16 18 20 22 24 26
17-25 cycles o£ RT-PCR for Hxstone 3.3
Ethidixwi bromi dn staining intensity
le+4 24 26 28 30
23-31 cycles o£ RT-PCR for dCK
Figure 7. HL60/R total RNA, 100 ng/reaction for histone 3.3 and 500 ng/reaction for dCK was used to determine the linear range of RT-PCR amplification. Shown on the y-axis is arbitrary ethidium bromide staining intensity, and on the x-axis is the number of PGR cycles. The blue arrows indicate the cycle numbers within the linear range chosen for semi-quantitative RT-PCR of dCK, 25 cycles, and Histone 3.3, 20 cycles.
57
M 1 2 3 4 5 6 7 8 9 1 0
Figure 8. Semi-quantitative RT-PCR for dCK. RT-PCR was performed for dCK (500 ng total RNA for 25 cycles/xeaction) and histone 3.3 (100 ng total RNA for 20 cycles/reaction), and then ran on a 2% agarose gel. The ethidium bromide staining intensity was determined with the Stratagene Eagle-eye software. Lanes: M = 100 base pair ladder, 1 = mock dCK, no RNA added, 2-5 = K562/S, K562/R, HL60ys, and HL60/R, 6 = mock Histone 3.3, no RNA added, and 7-10 = K562/S, K562/R, HL60/S, and HL60/R.
S8
DNA mathylatlon napping o£ tha 5' dCR Cp6 Island in nodal
cell Unas, noxnal human parxpharal blood lyaphoeytea
(PBL) , and acuta nyaloxd laukanxa (AML) patxants
The 5' maximal promoter region of the dCK gene is a
CpG island (Chen et al 1995). For this reason, we
hypothesized that dCK may be subject to DNA methylation
inactivation. To test this, we performed sodium bisulfite
sequencing of 49 5'-CG-3' sites in the dCK CpG island of
PBL, the paired ara-C-sensitive and ara-C-resistant human
leukemic cell lines K562/S, K562/R, HL60/S, and HL60/R, as
well as eight AML patients obtained from patients
refractory to ara-C-based chemotherapeutic regimens, see
table 2. The biochemistry underlying the sodium bisulfite
modification that allows for a positive display of those
cytosines that are methylated is shown in figure 9. The
region analyzed covered the maximal promoter region of dCK
from -228 to +143 relative to the transcription start site
(Chen et al 1995). In the paired ara-C-sensitive and ara-
C-resistant cell lines, K562/S, K562/R, HL60/S, and HL60/R,
no increase in cytosine methylation of the dCK CpG island
was seen. In fact, no cytosine methylation of the 5' dCK
CpG island was detected in the ara-C-resistant cells, see
figure 10, In normal PBL, all 49 CpG sites within the 5'
59
dCK CpG island were unmethylated. Similar to normal PBL
and the leukemia cell lines studied, 5 of 8 AML specimens
were completely unmethylated, and 3 of 8 had only 20%
methylation of one CpG site, see figure 11. In addition to
the results presented here, two independent investigations
have also failed to find methylation of dCK in drug-
resistant cell lines (Leegwater et al 1998, Fryxell et al
1999). In summary, the results we obtained indicate that
dCK mRNA is decreased in araC-resistant leukemia cell
lines, however the dCK 5' CpG island was unmethylated in
normal PBL, araC-sensitive and araC-resistant cell lines,
and ara-C-resistant AML patients. Thus, the mechanism for
the observed decrease in dCK expression is idiopathic.
The results presented here clearly demonstrate that
aberrant DNA methylation is not the only mechanism
responsible for decreases in gene expression that act as a
mechanism of drug resistance. The idiopathic decrease in
dCK gene expression is worth trying to understand, however
we decided to focus our effort on understanding aberrant
DNA methylation. For this reason, we chose to switch our
research efforts to pl5 and pl6, two tumor suppressor
genes, that had already been shown to be aberrantly
methylated in human cancer.
60
NH NH
O N ̂ SO," o-̂ N so
cvn,s.„.(0^ /""/ "
NH
CH
H H
5-Methylcytoslne Uracil
Figure 9. The biochemistry underlying the sodium bisulfite modification to define the methylation status of all cytosines within a PGR amplified region of DNA. Only unmethylated cytosines are converted to uracil which in the course of PGR is read as a T (T = 5-methyluracil), this results in a transition from G to T. Thus, only the cytosines methylated in the original DNA sample are retained after PGR. The PGR products are cloned and multiple sequences analyzed to obtain a percent methylation at each GpG site within the PGR product.
61
K562/S
100
80
60 40 20
K562/R
-228 -162 -125
100
80 H 60 40 20
-90 -JO 24 93
HL60/S
-228
100
80
60
40 20
-162 -125 -90 -30 24 93
HL60/R
-228
100 80
60 40 20
-162 -125 -90 -30 24 93
-228 -162 -125 -90 -30 24 93
Figure 10. 5-methylcytosine detection within the dCK 5' CpG island in K562/S, K562/R, HL60/S, and HL60/R. The y-axis is the percent methylation detected (multiple clones of the PGR product were cloned and sequenced, n = 5) and the x-axis is the location relative to dCK transcription start site. At the top, the solid black line depicts schematically the region of dCK sodium bisulfite sequenced, with exon 1 of dCK shown as an open box.
62
AML #1 M r AML HDAC/D
AML #2 M r AML ara-C/D
AML #3 F AML 2nd relapse HDAC/D
AML #4 F r AML HDAC/M
AML #5 M r AML HDAC/M
AML #6 F AML relapse HDAC/D
AML #7 M r AML HDAC/D
AML #8 M AML relapse ara-C/D
Table 2. Suminary of the ara-C-resistant acute myeloid leukemia (AMD samples obtained for the studies presented in this chapter and chapter 4. All the patients obtained for study failed araC-based chemotherapeutic regimes. Abbreviations: AML = acute myeloid leukemia, M = male, F = female, HDAC = high dose cytosine arabinoside, ara-C = cytosine arabinoside, D = daunorubicin, and M = mitoxantrone.
63
AML #1
PBIi
100 -
80 t 60 -40 -20 : 0
-228-162-125 -90 -30 24 93 100 -80 -60 40 -20 • 0 I i
-228-163-125 -90 -30 24 93
AMI. #5
100 80 60 40 20 0
-228-162-125 -90 -30 24 93
100 : 100 80 • 80
AML #2 AML #6m 20 " 50 0 0
-228-162-125 -90 -30 24 93 228-162-125 -90 -30 24 93
100 •
*0 • 100 60 ^ 80
AML #3 : AML #7 4 20
0 • 0
-228-162-125 -90 -30 24 93 -228-162-125 -90 -30 24 93
100 -
~ go * AML *4 S : *ML tSM 1
40 - " ̂ 20 • 20 • »
0 — 0 *
-228-162-125 -90 -30 24 93 -228-162-125 -90 -30 24 93
Figure 11. The sodium bisulfite sequencing of the dCK 5' CpG island in normal human peripheral blood lymphocytes (PBL) and ara-C-resistant adult acute myeloid leukemia (AML) patients. The y-axis is percent methylation and the x-axis is the location relative to the dCK transcription start site.
64
CRJiPTBR 4 - OKTSmcniATZOII OF 5-lfBTHTX.CTTOSZIfS
mrLUENCS att pis ahd pi6 kxprksszon
Es^ablxshad hiT«n lttuk«aixa call linss
After our initial studies on deoxycytidine kinase
(dCK) did identify a role for DNA methylation, we shifted
the focus to tumor suppressor, or genes capable of
preventing the conversion of normal cells to cancer cells,
that were hypermethylated. Fortunately, pl5 and pl6
presented an interesting situation to studying alterations
in DNA methylation. In 1997 Herman reported that Southern
blots (a technique that only determined the methylation
status of one pl5 CpG site) detected pl5 methylation in 88%
of AML patients (n = 60) but an absence of pl6 methylation.
These findings were surprising because pl6 is frequently
methylated in melanoma, breast, colon, lung, and Non-
Hodgkin's lymphoma whereas pl5 is only methylated in AML
and T-ALL. Thus, we already had AML patient samples with
which to do a detailed study of the pl5 and pl6 methylation
patterns in AML patients and cell lines. The initial
studies used established human cell lines to detect
methylation patterns of both pl5 and pl6.
65
We hypothesized that DNA hypermethylation was
associated with decreased expression of pl5 and/or pl6 in
leukemia cells. To address whether DNA methylation of the
pl5 or pl6 5' CpG islands was associated with gene
inactivation we obtained the following established human
leukemia cell lines: U937, monocytic leukemia (Nilsson and
Sundstrom 1974), KG-1, acute myeloid leukemia (Koeffler and
Golde 1978), and KG-la, a less differentiated variant of
KG-1 obtained after subcultivation of KG-1 (Koeffler et al
1980). In addition, the acute myeloid leukemia cell line
HL60/S, in Chapter 3, was also used here.
KG-1 was originally obtained from a patient with AML.
After subcultivation of KG-1, a phenotypically distinct
subline, designated KG-la, emerged. KG-1 and KG-la are an
interesting pair of cell lines. Though both KG-1 and KG-la
originated from one AML patient, some of the phenotypic
differences between KG-la and KG-1 are: (1) KG-la is less
differentiated than KG-1 based on the cellular markers
present on the cells surface, (2) KG-la will form colonies
in soft agar, but KG-1 does not, (3) KG-1 differentiates in
response to phorbol ester treatment while KG-la does not,
and (4) KG-1 is GM-CSF responsive but KG-la is not. The
four myeloid leukemia cell lines, U937, HL60, KG-1, and KG-
66
la were used to determine pl5 and pl6 (1) mRNA by RT-PCR,
(2) protein by immunoblotting, and (3) DNA methylation
levels by sodium bisulfite sequencing. In addition, four
other cell lines were used as controls. HaCaT, a human
keratinocyte cell line, served as a positive control for
pl5 mRNA and protein detection (Boukamp et al 1988). RHEK-
1, a human keratinocyte cell line, served as a positive
control for pl6 protein detection (Rhim et al 1985).
Melanoma UACC612, derived at the Arizona Cancer Center from
a patient with melanoma, served as a positive control for
pl6 methylation detection. Finally, Jurkat, a human T-cell
acute lymphocytic leukemia (T-ALL) cell line served as a
negative control for both pl5 and pl6 as it contains a
homozygous deletion of both pl5 and pl6 at chromosome 9p21
(Neuveut et al 1998). The cell lines and pl5 and pl6 RT-
PCR and immunoblotting results are summarized in table 1.
pl5 and pl6 mRNA. and protein atudx^a in leukemia cell lines
The established human leukemia cell lines were used to
determine if DNA methylation was associated with a loss of
pl5 or pl6 mRNA and protein. We determined pl5 and pl6
mRNA presence by RT-PCR. U937, HL60, HaCaT, and melanoma
UACC612 all had detectable pl5 mRNA by RT-PCR as shown in
67
figure 11. In contrast, pl5 mRNA was not detected by RT-
PCR in Jurkat, KG-1, and KG-la. In all samples Histone 3.3
was detected by RT-PCR as a control for RNA integrity.
Immunoblotting was used to detect pl5 and pl6 protein in
the same cell lines. pl5 protein was detectable in U937,
HL60, HaCaT, and melanoma UACC612, while Jurkat, KG-1, and
KG-la all lacked detectable pl5 protein, shown in figure
12. In addition to pl5, we also determined the status of
pl6 mRNA and protein. pl6 mRNA was detected in KG-1, KG-
la, and HL60 but not in Jurkat, U937, HaCaT, or melanoma
UACC612, shown in figure 12. pl6 protein was detected in
KG-1 and KG-la, but not HL60, U937, Jurkat, HaCaT, or
melanoma UACC612, see figure 13. These results indicate
that pl5 mRNA and protein were not detectable in Jurkat,
KG-1 or KG-la, but both pl5 mRNA and protein were detected
in U937, HL60, melanoma UACC612, and HaCaT. In the case of
pl6, pl6 mRNA and protein were not detected in Jurkat,
U937, melanoma UACC612, HaCaT, while HL60, KG-1, and KG-la
had detectable pl6 mRNA (see Table 1 in Chapter 1 for a
summary).
68
M 8
Figure 12. RT-PCR detection of pl5, pl6, and Histone 3.3 mRNA by ethidium bromide staining of a 2% agarose gel. The lanes are: M = marker, 1 = mock, no RNA, 2 = Jurkat, 3 = KG-la, 4 = KG-1, 5 = U937, 6 = HL60, 7 = melanoma UACC612, and 8 = HaCaT. For pl5 and pl6, 1 (ig of total RNA was reverse transcribed and amplified for 33 cycles. Histone 3.3, 100 ng total RNA amplified for 25 cycles, served as a control for RNA integrity.
69
16.7 Kd 6.7 Kd
1 2 3 4 5 6 7 8
pl5
16.7 Kd 6.7 Kd
•pi 6
Figure 13. Detection of pl5 and pl6 protein in Jurkat (lane 1), KG-la (lane 2), KG-1 (lane 3), U937 (lane 4) HL60 (lane 5), melanoma UACC612 (lane 6), HaCaT (lane 7), and positive controls, recombinant pl5 and RHEK-1 pl6 (lane 8). Each sample, excluding the positive controls, contained 0.25 mg of protein resolved by SDS-PAGE on a 15% acrylamide gel. The arrows at 16.7 and 6.7 kilodaltons show migration of protein standards.
70
Definition o£ th« pl5 msxleal proao-ter In HaCaT calls
To determine the region of the pl5 5' CpG island
responsible for pl5 gene activity, three deletion
constructs of the pl5 promoter were generated and inserted
5' of the chloramphenicol acetyltransferase (CAT) reporter
gene, see figure 13. The pl5 promoter constructs are:
plSPSTl from -1177 to +92, pl5PST2 from -448 to +92, and
pl5PST3 from -126 to +92, in all cases the numbering refers
to position relative to the pl5 transcription start site.
The results, shown in figure 14, demonstrate that pl5PST2
promoter activity in HaCaT was ~3.5 times greater than
pl5PSTl, and ~4.3 times greater than pl5PST3. These
results indicate that the region most relevant to pl5
transcriptional activity is located within the region from
-448 to +92 relative to the pl5 transcription start site.
Regions of the pl5 promoter further 5' of -448 are not
likely responsible for pl5 transcriptional activity, and in
fact may be inhibitory. Based on these results we chose to
sodium bisulfite sequence the pl5 5' CpG island from -440
to +474.
71
plSPSTl -1177 pl5PST2 -448 ̂ mhh^mhh^hm-i-92 pl5PST3 "126 mmmmm'^92 Sodiism bisxilfite aAi-niArt^-i tirr —440 +474 sequencing
20
basic SV40 PSTl PST2 PST3
Figure 14. CAT reporter contructs of the pl5 5' region were transiently transfected into HaCaT. Three constructs containing pieces of the pl5 promoter were cloned into the pCAT vector. pl5PST2, from -448 to +92 relative to pl5 transcription start site has the highest activity normalized to I3GAL activity and total protein isolated. As shown, the region sodium bisulfite sequenced overlaps most of P15PST2 (n = 3).
72
DMA mathyla^on mapping o£ pl5 and pl6 5' CpG Islands
In established human laukamxa call Unas, normal hi,man bona
marrow (MBM) , and AML patients
The pl5 CpG island was sodium bisulfite sequenced from
-440 to +474 relative to the transcription start site; this
region contains 80 CpG sites. The cell lines U937, HL60,
melanoma UACC612, and HaCaT were all primarily unmethylated
in the pl5 CpG island, see figure 15, and all had
detectable pl5 mRNA and protein. The other cell lines
analyzed, Jurkat, KG-1, and KG-la did not have detectable
pl5 mRNA or protein. Jurkat methylation was not determined
because pl5 and pl6 are both deleted in this T-ALL cell
line. In the case of KG-1 and KG-la, two distinct modes of
aberrant methylation of the pl5 CpG island were identified.
KG-1 had a variegated pattern with 23 of 80 CpG sites >40%
methylated throughout the pl5 CpG island. All the alleles
sequenced were partially methylated consistent with
methylation of both pl5 alleles in KG-1. The other form of
methylation, observed in KG-la, was complete methylation
with >80% methylation in 75 of the 80 CpGs sites
distributed throughout the pl5 CpG island. These results
provide in vitro evidence that variegated methylation and
complete methylation of the pl5 5' CpG island observed in
73
KG-1 and KG-la are both associated with transcriptional
silencing of pl5.
In addition to pl5, we also determined the 5-
methylcytosine content of the pl6 5' CpG island in the cell
lines described above- The pl6 CpG island was sodium
bisulfite sequenced from -552 to +156 relative to the
transcription start site; this region contains 53 CpGs.
KG-1, KG-la, and HL60, the three AML cell lines that were
positive for pl6 mRNA, all had only low DNA methylation at
the pl6 CpG island, see figure 15. Similar to the study of
Herman at al (1996) where Q937 was methylated at a Sacll
site in pl6, we detected 80% methylation of U937 in the 3'
region of pl6 from +92 to +156, which contains the Sacll
site. The 5' region of pl6 in U937, however, was only 20%
methylated from -404 to +92. We could not detect pl6 mRNA
or protein in HaCaT. In addition, HaCaT contained >60%
methylation in 5 of 7 sites from -461 to -380, and also
from +29 to +53 in the pl6 5' CpG island. A recent study
corroborates our finding that pi6 is not present in HaCaT
(Munro et al 1999). Melanoma UACC612 lacked detectable pl6
mRNA or protein, and contained >80% methylation in 51 of 53
CpG sites in the pl6 CpG island. In summary, the results
obtained from sodium bisulfite sequencing indicate that
74
while the pl6 CpG island displays more complex regional
patterns of methylation in vitro than the pl5 CpG island
(see U937 and HaCaT figure 15), an association between pl6
gene inactivation in vitro and DNA methylation was observed
in melanoma UACC612.
We also determined the 5-methylcytosine content of the
5' CpG islands of pl5 and pl6 in 2 normal human bone marrow
samples (NBM) and 8 cytosine-arabinoside (araC)-resistant
adult acute myeloid leukemia patients. In the 2 NBM samples
pl5 and pl6 were both predominantly unmethylated, see
figure 16, part 1. pl6 had only 1 methylated cytosine of
the_ 1, 060 CpGs sequenced or ~0.1% of all sites (n of 10 per
sample), while pl5 had 37 methylated cytosines of the 1,600
CpGs sequenced or -2.3% of all sites (n of 10 per sample).
In this study, the pl5 CpG island of NBM was ~25 times more
likely to be methylated at any given CpG compared to the
pl6 CpG island, however, overall both were largely
unmethylated. These results show that the pl5 and pl6 5'
CpG islands contain very low levels of methylation in NBM,
and are in accordance with other studies that have not
found DNA methylation in the 5' CpG islands of genes in
normal human tissue (Fryxell et al 1999).
75
The ara-C-resistant acute myeloid leukemia (AML)
patients had examples of all three types of methylation at
the pl5 CpG island described earlier in normal human bone
marrow (NBM) and the human cell lines KG-1 and KG-la.
Patients #6 and #7 were almost completely unmethylated;
combined patients #6 and #7 had ~2.9% of all CpG sites
methylated (n of 5 per patient) in similar fashion to NBM,
see figure 16, part 2. Patients #1, #2, #4, #5, and #8
displayed a variegated pattern throughout the entire pl5
CpG island similar to the pattern observed in KG-1, see
figure 16. Overall, of the 80 CpG sites analyzed in the
pl5 CpG island, >40% methylation was detected at 12 sites
in patient #1, 20 sites in patient #2, 33 sites in patient
#4, 26 sites in patient #5, and 15 sites in patient #8.
Finally, patient #3 (the only patient of the 8 in this
study that was in 2"*^ relapse) had almost complete
methylation of the pl5 CpG island with 78 of 80 CpGs > 75%
methylated, similar to KG-la.
The pl6 CpG island, in contrast to the pl5 CpG island,
contained only 8 methylated cytosines of the 2,120 CpG
sites analyzed in the 8 AML patients, or ~0.4% (n of 5 per
patient), a result similar to the methylation levels
detected in NBM, see figure 16 parts 1 and 2. In
76
conclusion, the results obtained from sodium bisulfite
sequencing the 8 AML patients showed aberrant methylation
at the pl5 5' CpG island in 75% of the patients, but only
minimal methylation at the adjacent pl6 5' CpG island.
Further, the methylation patterns of the pl5 5' CpG island
in AML were found to extend throughout the region analyzed
from +474 to -440, and can be categorized as either (1)
minimal methylation (2 of 8) (2) variegated methylation (5
of 8) , or (3) complete methylation (1 of 8) .
77
10 0
80
60 •
40
20
0 1 BL60
-440-151-75 28 113 233 322 416
100-
8 0 -
60 :
40 •
2 0 -
0 1 -552 -344 -177 -19 50 140
100
80
60
40
20
0 J L -440-151-75 28 113 233 322 416
100-|
8 0
60
4 0 1 1 • i 1 KG—1
20
-440-151-75 28 113 233 322 416 552 -344 -177 -19 50 140
100
80
60
40
20
0
100
80 •
60 •
40 •
20 •
0
1 0 0 -
8 0 -
60 •
40 -
20 -
0
K6-l«
-440-151-75 28 113 233 322 416
100-
80-^ 60-;
40
20
0 -552 -344 -177 -19 50 140
i -440-151-75 28 113 233 322 416
-440-151-75 28 113 233 322 416
Ml 100-
OACC 80 1
612 6o] 40 1 201
0 ' 100-
HaCaT 8 0 -1 60 ii 40 1 20 j 0
-552 -344 -177 -19 50 140
Figure 15. Methylation analysis of pl5 & pl6 in cell lines.
78
pl5 pl6
100 80 60
40
20 0
100 80
60 40 20
0
100
80
60 40 20
0
NBM «1
kjLA A A -440-151-75 28 113 233 322 416
100
80
60 :
40 •
A °-552 -344 -177 -19 50 140
100^
iLL. -A 4- -A—i -440-151-75 28 113 233 322 416
80^ 60-
NBM «2 40 • 2 0 •
0 -552 -344 -177 -19 50 140 100-
8 0 -
6 0 -
MfL *1 4 0 -2 0 1
-440-151-75 28 113 233 322 416 ^ -552 -344 -177 -19 50 140 100
80 •
6 0 -
MIL *2 20
-552 -344 -177 -19 50 140 100:
80^
-440-151-75 28 113 233 322 416
MIL #3 i 40 1
2 0 1
0 552 -344 -177 -19 50 140
Figure 16, part 1. Sodium bisulfite sequencing of pl5 and pl6 in two normal bone marrow samples (NBM) and 3 of the 8 adult acute myeloid leukemia (AML) patients.
79
lOO:
80 •
60 40 20
0 -440-151-75 28 113 233 322 416
100~
MiL «4 80-
60] 40 • 201
0 -552 -344 -177 -19 50 140
100 80 60 40 20
0
100 MO. #5 80
60
40 20
-440-151-75 28 113 233 322 416 0 -552 -344 -177 -19 50 140
100 80 •
6 0 •
40 • 20 0 1
100"
AML «6 80 60 40'
14 A- 20^__| -440-151-75 28 113 233 322 416 0-552 -344 -177 -19 50 140
100"
80 '
60 •
40 1 20 I 0
100"
MIL «7 8 0 '
60 -
40 1
J A_i i -440-151-75 28 113 233 322 416 -552 -344 -177 -19 50 140
100"
MIL *8 : 601
440-151-75 28 113 233 322 416
40 1
L -552 -344 -177 -19 50 140
Figure 16, part 2. Sodium bisulfite sequencing of pl5 and. pl6 in the remaining 5 adult acute myeloid leukemia (AML) patients.
80
NBM#1
GO •DOO OOO O «D O ooa ID O 0<D OO o o« mtD •o o* OO «DOO OOO O €D o OOJ D O 0<D OO O OO OOO OO OO
OO «DO* OOO O CD o OOO D O ooa eo O OO OOO OO OO «o •DOO OOO O «D o o«a ID O oa» OO O OO OOD OO OO GO «DOO OOO O «D o oa D O 0<D OO O OO OOD OO OO OO «DOO OOO O €D O 009 ID O OOD OO O OO OOO OO OO
OO «B>00 OOO O €D O 003 ID O oa> OO O OO OOO OO OO OO •DOO OCO O <D O OOO 09 O OOO OO O OO OOD OO OO
AML#5
•D mmoo oamo o 0«D • om OO o •• COD •o oo 00 mooo oaoo CD o OOOD o oca OO O OO OOD oo oo «> mooo oaoo CD • • (» • oca 3« o mm mom oo oo «o •DOO oaoo CD o OOOD o OO •O O OO mm •• •• •0 mmmo ooM m o 0«fl» • o<D OO • •• mo •o 00 cm mtmm mmo CD o oa» • • a» OO o •• mo •• oo m mmmo mcmm « o 0<MD • OOD OO • •• mo mm oo
cm mmmm ocmm CD • • CH» • mm •• o •• mam om oo
-14 +385
Figure 17. Individual sodium bisulfite sequences from 1 normal human bone marrow (NBM#1), and 2 AML patients (AML#3 and AML#5). The open circles indicate unmethylated 5'-CG-3' sites and the closed circles indicate methylated 5'-CG-3' sites. The variegated pattern of pl5 methylation pattern is shown in patient #5, and similar results were obtained from AML patients #1, #2, #4, and #8. On the bottom, the position relative to pl5 transcription start site is indicated.
81
CHAPTBR 5 - niTRODUCTION OT P15 AMD P16
INTO TBX T-ALL CELL LZME JURKAT
Stabla ovarM^rassion ot pl5 or pl6 xnte Jurkat Z:
•stablxaluMn ̂ of pl5 and pi6 poal^v* aubclonas
Jurkat is a human T-cell adult acute lymphocytic
leukemia (T-ALL) cell line. In order to confirm previous
reports that Jurkat lacks pl5 and pl6, we performed PGR
that overlaps either pl5 exon 1 or pl6 exon 1, see figure
18. This result corroborated previous reports that Jurkat
is pl5 and pl6 negative due to homozygous deletion at
chromosome 9p21 (Neuveut et al 1998). As demonstrated in
chapter 4, pl5 but not pl6 is methylated in AML. In
addition to AML, methylation of pl5 but not pl6, as well as
homozygous deletion of both pl5 and pl6 (as is the case
with Jurkat), occurs in T-ALL (Batova et al 1997), To
attempt to identify a role for pl5 in leukemogenesis, pl5
or pl6 were introduced into the T-ALL cell line Jurkat-
pl5 or pl6 introduction may result in cell death or a
dramatic decrease in cell growth, for this reason we began
by trying to establish inducible expression of pl5 or pl6.
To achieve inducible expression we used the tetracycline
82
lOObp PBL JE^
lOObp
p15
p16
Figure 18. The PGR for portions of exon 1 of both pl5 and pi6 was performed with normal peripheral blood lymphocyte (PBL) DNA and Jurkat Tet-Off (J) DNA. As indicated with the arrows, PGR products were detected in PBL for both pl5 and pl6 but not in Jurkat Tet-Off consistent with the presence of a homozygous deletion of both pl5 and pl6 in Jurkat.
83
inducible expression system originally described by Gossen
and Bujard (1992).
The cell line Jurkat Tet-Off was purchased along with
the necessary plasmids from Clontech. Jurkat Tet-Off was
generated by stable expression of a tetracycline responsive
protein fused to the VP16 transactivation domain (tetR-
VP16) in Jurkat. With the addition of tetracycline (tet)
or doxycyline (dox) tetR-VP16 no longer transactivates any
gene introduced with a 5' tetracycline-responsive element
(TRE), see figure 19. After obtaining Jurkat Tet-Off, an
initial experiment was performed to confirm that tetR-VP16
was functional. We transiently transfected a green
fluorescent protein (GFP) plasmid with a 5' TRE into Jurkat
Tet-Off. The transfected cells were then cultured in the
presence of 0.001 to 1000 ng/mL dox. The result, shown in
figure 20, demonstrated that tetR-VP16 was inhibited by
increasing dox concentration. We next attempted to
establish stable inducible expression of pl5 and pl6 in
Jurkat Tet-Off cells. Two independent strategies of
electroportion were attempted to establish stable
transfection into Jurkat Tet-Off: (1) two plasmids, the
hygromycin drug resistance gene, pTK-Hyg, plus the pl5 cDNA
84
Inttgrata^ copy •! pT•^Off rcfiiatar piasiRid
i-TctarDox)
TcMrOcK)
TrancriMNs
lalaflraM copy of pTRE rospoaso plasm ii
Figure 19. Explanation of the tetracycline inducible expression system described by Gossen and Bujard in 1992. The TetR-VP16 fusion protein is stably integrated in Jurkat Tet-Off. Introduction of a gene containing a 5' TRE can be induced by the removal of tetracycline or doxycyline.
85
% 6FP-I-Jurkat: | q -Tet-Off
0.4 -
0.001 0.01 0.1 1 10 100 1000 10000
doxycylxne ng/mL
Figure 20. A doxycyline responsive GFP vector was transiently transfected into Jurkat Tet-Off cells. Increasing concentrations of doxycyline (0.001 ng/itiL to 1000 ng/rtiL) decreased the number of GFP positive cells as determined by flow-assisted cytometry (FACS) analysis. This demonstrated that Jurkat Tet-Off contains a functional tetR-VP16 protein.
86
with a 5' TRE, plSpTRE, or (2) the hygromycin drug
resistance gene subcloned into the plSpTRE plasmid to
generate one plasmid, pTKHygplSpTRE. The electroporated
cells were selected for 3 weeks by limiting dilution in the
presence of 400 ng/mL hygromycin. The use of the 2 separate
plasmids was used by Bujard and Gossen (1992) in defining
this system in HeLa cells, and is the procedure recommended
by Clontech (the single plasmid containing the TRE and
hygromycin resistance gene required assembly). The idea to
use a single plasmid came from the few reports in the
literature where the tetracycline inducible expression
system was used with Jurkat cells (Kwon et al 1998).
As shown in table 3, only 1 of 36 independent clones
obtained from limiting dilution of the Jurkat Tet-Off
transfected cells using the two separate plasmids actually
expressed pl5. In contrast, the single plasmid
electroporation yielded 11 of 24 pl5 mRNA positive clones.
As figure 21 illustates, though, the pl5 protein levels
were not shut off in the presence of 1 ng/mL of dox (all
selection was performed in the presence of dox and separate
clones grown in absence of dox were used to determine the
pl5 protein levels). In summary, the results obtained
87
suggest that inducible expression is difficult to obtain in
Jurkat cells likely due to the requirement of a single
plasmid transfection. The clonal sublines of Jurkat were
treated as stable overexpressors of either pl5 or pl6, see
figure 22.
88
+ - + - + - + - p i s
p15
Figure 21. The clonal populations that were hygromycin-resistant and pl5 itiRNA positive without doxycyline were grown in the presence (1 ng/mL) or absence of doxycyline. The pl5 protein levels of the 12 clones were then determined by Western blot, 100 ng per sample. As shown, the expression of pl5 was not inducible.
89
Number of 24 36 clones pl5 DNA 17 of 24 22 of 36 positive pl5 RNA 11 of 24 1 of 36 positive Inducible 0 of 11 0 of 1 expression
Table 3. Summary of the results of the two approaches to obtaining pl5 positive Jurkat Tet-Off cells. The Clontech recommended protocol of two plasmids yielded only 1 of 3 6 pl5 mRNA positive clones. In constrast, the single plasmid, pTKHygplSpTRE, yielded 11 of 24 pl5 positive clones. Our analysis did not show any clones that were inducible due to the high level of "leaky" expression in the presence of doxycyline (l|ig/mL), see figure 20.
90
T-O Pl5 pl6
pl5
pl6 -M
Figure 22. Establishment of multiple clones of Jurkat Tet-Off that stably overexpress either pl5 or pl6. A Western blot was performed with 100 ng of protein per sample, a pl5 fusion protein served as a positive control for pl5 detection, and RHEK-1, a human keratinocyte cell line, served as a positive control for pl6 detection.
91
Stable overttxpresslon of pl5 or pl6 xn Jurkat ZI: r«duc«d
groirth of pl5 positive cells and unchanged axa-C and
daianonabxcln sensitivity
The growth of the pl5 or pl6 positive clones of Jurkat
was measured by plating 400,000 cells per well in 6-well
plates and counting total cell number each day for 6 days.
The results, shown in figure 23, revealed that pl5 clones
grew slower than the pl6 clones or hygromycin only selected
cell line. Interestingly, after the initial slow growth,
the pl5 positive clones displayed a similar rate of growth
from day 4 to day 6. This may represent a cytokine-
dependent growth rate that is slower to "turn-on" in the
pl5 positive clones.
The decreased growth in the pl5 positive clones was
further confirmed by using the MTS assay, which determines
metabolic activity. In the MTS assay, 8,000 cells were
plated out per well in 96-well plates and allowed to grow
for 5 days. The results, shown in figure 24, indicate a
statistically significant decrease in growth only in the
pl5 clones (paired T-test, n = 3, T = 7.24, p = 0.0185).
To determine why the pl5 cells were growing slower we
performed: (1) a Western blot to determine the
phosphorylation status of retinoblastoma (pRb) (2) cell
92
cycle analysis by flow-assisted cell sorting (FACS) with
propidium iodide (PI) staining, and (3) apoptosis analysis
with annexin V/PI FACS staining. The phosphorylation of
retinoblastoma was not altered in either the pl5 or pl6
clones (data not shown)- In addition no differences in the
% growth
4500
4000
3500
3000
2500
2000
1500
1000
500
0
^ "
• •
- -
-i--
/ . . - -
- V
. .V
^ ' .
m
^ T — 1
2 3 4 5
days in culture
6
pTKHgy p15
p16
Figure 23. Cells were plated in 6-well plates at 400,000 cells in 4 mLs of media and allowed to grow for 6 days. Cell counts were performed on days 2-6. The y-axis is percent of cell growth, and the x-axis is days in culture. A hyrogromycin only selected clone, pTK-Hyg, and 2 clones each of pl5 and pl6 are summarized. As shown, pl5 clones had the slowest growth rate (n = 2, one representative result is shown).
93
140
120
0) |3 100 (0
80 -
•S 60
40
20
T
- *
T
-
" • '
" r
1 '
Jurkat Tet-Off pIS-i- p16-i-
Figure 24. Equal cell numbers, 8,000 cells per well, were plated out in a 96-well plate and allowed to grow for 5 days. Metabolic activity was determined at day 5 using the MTS assay (n = 3, the error bars show the standard deviation). The asterisk (*) indicates statistical significance using the Sigma Plot software, paired T-test (paired T-test, n =3, T = 7.24, p = 0.0185).
94
cell cycle profile (Gl, S-phase, and G2/M) were observed in
the pl5 or pl6 positive clones (data not shown). In the
annexin V staining experiments, increased apoptosis and
necrosis was detected in one of the pl5 clones, shown in
figure 25 (not all pl5 positive clones showed an equal
increase, only the most dramatic clone is shown). These
results suggest that pl5 introduction into Jurkat decreases
growth due to increased cell death.
All 8 AML samples used to describe the aberrant
methylation of pl5 in 6 of the 8 AML cases in chapter 4
were ara-C-resistant. Thus, loss of pl5 or pl6 in Jurkat
may represent a mechanism of drug resistance acquired after
drug treatment. To determine if ara-C or daunorubicin, two
drugs currently used clinically to treat AML and T-ALL,
would have increased efficacy in the pl5 or pl6 positive
clones of Jurkat, we performed an MTS assay with ara-C or
daunorubicin doses from 1 nM to 1000 nM. The results,
shown in figure 26, indicate that pl5 or pl6 presence in
Jurkat cells does not alter sensitivity to ara-C or
daunorubicin. The IC50 values of ara-C treatment of Jurkat
Tet-Off, pl5, and pl6 for were 44.8 nM, 45.1 nM, and 44.9
nM respectively. These results suggest that pl5 or pl6
loss occurs independent of drug treatment.
%of total cells
% necrotic
% apoptotic
T-O Hygro pl5 pl6
Figure 25. Annexin V and propidium iodide staining were used to determine the percent of cells that were apoptotic shown in blue, and the percent of cells that were necrotic shown in black. Flow-assisted cytometry (FACS) was used and 10,000 events per sample were scored (n = 2, one representative experiment is shown).
%
AraC Dose Response
0.000001
>1 a. ̂ ̂̂ ^
« 0.6
A—p16
0.00001 0.0001
Molar Concentration of AraC
0.001
Tet-Off
Daunorubicin Dose Response
.2 £ 3 CO
u
0.000001 0.00001 0.0001
Molar Concentration of Daunorubicin
0.001
Tet-Off pis p16
Figure 26. MTS assay of the relative sensitivity of Jurkat Tet-Off, a pl5 clone, and pl6 clone to ara-C and daunorubicin. The parent cell line, Jurkat Tet-Off is shown in blue, the pl5 clone is shown in pink, and the pl6 clone is shown in purple (similar results were obtained for a total of 3 clones each for both pl5 and pl6, n = 3).
97
St:able ovarsxprttsaxon of pl5 or pl6 In Jurkat ZIZ: cDHA
mxcroarray analysis of 5700 gttn«s rsvaals tlia^ pl5 and pl6
differ froa Jurkat bu^ ar« similar te aach other
To obtain a bird's eye view of the alterations that
occurred in Jurkat Tet-Off after pl5 or pl6 introduction,
we performed expression analysis on ~5700 cDNAs spotted
onto a microscope slide, referred to as microarray. As
noted in the previous section a decrease in the growth of
the pl5 positive clones was found. This suggested that
possibly there was a pl5 specific function distinct from
pi6. The use of microarray allowed us to determine the
expression changes that occurred in the pl5 and pl6
positive Jurkat cells.
An example of the results obtained when the mRNA from
a pl5 positive clone was labeled red with Cy3-dCTP and
Jurkat Tet-Off mRNA was labeled green with Cy5-dCTP is
shown in figure 27. A number of cDNAs, ~20 showed altered
expression in all three pl5 or all three pl6 clones. pl5
and pl6 expression profiles were distinguishable from
Jurkat Tet-Off. Interestingly, and counter to the growth
studies, the pl5 and pl6 positive clones had similar
expression changes from the parent cell line Jurkat Tet-
Off. Specifically, we performed northern blot analysis to
98
confirm the expression changes determined using microarray
on the following 5 genes (see table 4): (1) the tyrosine
phosphatase PTPNl; (2) heparin-binding epidermal growth
factor-like growth factor; HB-EGF, (3) the anti-apototic
BCL-2 family gene, MCLl; (4) the pro-apoptotic gene
BCL2/adenovirus ElB 19kD-interacting protein 3, BNIP3; and
(5) the interferon-gamma inducible transcription factor,
IRF-1. In addition, GAPDH mRNA levels were determined for
each blot to ensure equal loading and also for use in
quantitative expression levels. The northern blot results,
shown in figure 28, confirmed the miroarray data. The
expression levels were quantitated using the Stratagene
Eagle-eye software, normalized to GAPDH levels, and the
data obtained from microarray and northern blots are
summarized in figure 29. The results show that microarray
data and northern blot data were similar. The total number
of genes with altered expression in pl5 and pl6 was -20 or
~0.4% of all the genes assayed by microarray (several of
which were ESTs of unknown function). The expression
changes in pl5 and pl6 were concordant. These results
suggest that pl5 and pl6 are functionally equivalent and
interact in the same genetic networks.
99
The red spot below corresponds to BNIP3, accession namber AA446839, indicating a teck of CyS-CTP (green) bnt presence of Cy3-CTP (r^) from the parent ccU line cDNA. Expression of BN1P3 is decreased in pl5/pl6 Jnrkat cells.
Figure 27. A example of the use of a 5,700 cDNA microarray to determine the alterations in expression between pl5 or pl6 and the parent cell line Jurkat. The parent cell line mRNA was labeled red with Cy5-dCTP and the pl5 or pi6 clones were labeled green with Cy3-dCTP. A red or green color indicates differential expression, while yellow indicates similar expression levels.
100
CHANGE IN ABBREVIATION FUNCTION EXPRESSION
IN pl5/pl6 BNIP3 Pro-apoptotic Decreased HB-EGF Growth factor Increased IRF-1 Transcription
factor Increased MCLl Anti-
apoptotic Increased Tyrosine
PTPNl phosphatase Increased
Table 4, A list of the genes that were chosen for further analysis from the microarray data.
lOl
T-O Hyg pl5 pl6
BNIP3 9 •
GAPDH
HB-EGF
GAPDH
MCLl
GAPDH
IRF-1
GAPDH
PTPNl
GAPDH
Figure 28. Northern analysis using 20 (ig of otal RNA per lane. The genes tested were selected based on microarray data of 3 independent pl5 and pl6 clones of Jurkat.
102
• = Microarray result H » Northern result
Fold expression change vs.
Jurkat Tet-Off
B3 HB ML IR PT R^ HR ML IR PT
pl5 pl6
Figure 29. Summary of the results obtained with microarray, in white, and Northern blots, in black. Genes studied: B3 = BNIP3, a pro-apoptotic gene; HB = HB-EGF, a mitogenic gene; ML = MCLl, an anti-apoptotic gene; IR = IRF-1 an interferon-responsive transcription factor; and PT = PTPNl, a protein tyrosine phosphatase.
103
CBAPTKR 6 - DISCUSSION
Aberrant Dim of th* pl5 5' CpG Island occurs iJi
the majority ot acuta Hyaloxd lanlrssii a (AML) patiants and
might dysragulata normal haeatopoietic dlf£arantlatlon
The hypothesis of the preceding experiments is that
hypermethylation of 5' CpG islands is associated with
transcriptional silencing of the corresponding gene and
participates in either the emergence of drug resistance or
the conversion of normal cells to cancer cells. The
hypothesis that decreased dCK activity, a mechanism of drug
resistance, was mediated by hypermethylation of the 5'
maximal promoter of dCK was tested and disproved.
Next, the hypothesis shifted to the role of DNA
methylation in the transcriptional silencing of pl5 and
pl6, representing a mechanism of avoiding normal tumor
surveillance and promoting the conversion of normal cells
to cancer cells.
The results obtained from studying the DNA methylation
patterns in leukemia cell lines and AML patients indicate
that while pl6 was not aberrantly methylated in the
leukemia cell lines or the AML patients, the aberrant
methylation of pl5 occurred both in vitro and in vivo. Two
104
distinct modes of aberrant pl5 methylation, variegated
methylation and complete methylation, were detected both in
vitro and in vivo. KG-1, an AML cell line, had variegated
methylation of pl5 associated with no detectable pl5 mRNA
or protein, and KG-la a subclone of KG-1 had complete
methylation of pl5 also associated with no detectable pl5
mRNA or protein. Six of the eight AML patients studied had
aberrant methylation of pl5, five of eight, 62.5%,
displayed a variegated pattern of pl5 methylation, and one
of eight, 12.5%, displayed a complete pattern of pl5
methylation. Thus, based on KG-1 and KG-la where
variegated and complete methylation of pl5 were both
associated with no detectable pl5 mRNA or protein, it is
reasonable to infer that in vivo aberrant DNA methylation
of pl5 in AML inactivates pl5 in 75% of AML patients.
The selective inactivation of pl5 but not pl6 is
surprising given that pl6 is located adjacent to pl5 on
chromosome 9p21 and shares >80% protein identity to pl5.
How such a situation arises in AML is not immediately
apparent. To try to understand whether pl5 has a specific
function that pl6 lacks we introduced pl5 or pl6 into the
pl5 and pl6 negative T-ALL cell line Jurkat.
lOS
We did detect reduced growth in the pl5 but not pl6
positive clones. In addition, while a 5,700 cDNA
microarray distinguished pl5 and pl6 positive clones from
the parent cell line, the pl5 and pl6 expression profiles
were very similar. This suggests, despite the decreased
growth seen only in the pl5 positive cells, that pl5 is
functionally equivalent to pl6. It is important to note
that while the use of microarray did provide enormous power
in understanding the expression changes in the pl5 or pl6
positive cells, it only measured -6% of all human genes.
It is possible th.at among the 94% of genes not studied
there are pl5 specific changes in expression that provide
evidence for a functional distinction between pl5 and pl6
that would explain the selective methylation of pl5. Also,
it may simply be the case that pl5 overexpression is
significantly higher than the pl6 expression levels.
Determination of the absolute amount of pl5 protein
relative to pl6 protein is not a trivial experiment because
of differences in antibody specificity. In summary, while
the growth of the pl5 clones suggested that pl5 may have
some functional independence from pl6, microarray
expression analysis of 5,700 cDNAs did not find any
evidence to support a pl5 specific function.
106
Several genes that were differentially expressed in
the pl5 and pl6 clones were identified from the microarray
data and confirmed by Northern blots. Among these, HB-EGF,
heparin-binding epidermal growth factor, has previously
been implicated in monocytic differentiation.
Specifically, phorbol ester treatment of HL60 results in
monocytic differentiation, associated with induced
expression of pl5 and HB-EGF (Schwaller et al 1997, Vinante
et al 1999). This provides a possible explanation for the
observed selective methylation of pl5 in AML. As shown in
figure 30, pl5 methylation may prevent the induction of pl5
that would normally mediate monocytic differentiaion by
triggering a cascade that induces other genes such as HB-
EGF. The blockade of this cascade may participate in the
genesis of acute myeloid leukemia (AML).
Di f f e ren t i a t i ng s i g n a l s
G r o w t h dys regu la ted i n A M L p l 5 me thy la ted ^ ^ ^
/ V / V
HB-EGF
107
Figure 30. A model of why pl5 is methylated in AML but pl6 is wild-type. A cytokine might direct normal hematopoietic differentiation by inducing pl5, which results in changes in the expression of genes such as HB-EGF. DNA methylation of pl5 might block this monocytic differentiation pathway.
108
Aberrant B«tbyla^on o f pl5 as a markar o f AML and
po^n^iaX targafc for noval drug traal«aiifca
To date, the predominant genetic and epigenetic
modifications described in acute myeloid leukemia (AML)
are: (1) the oncogenes ras and fms are mutated at
frequencies of ~30% and ~15% respectively (Bos et al 1987,
Tobal et al 1980) (2) several chromosomal translocations
occur that create fusion proteins; one of the most frequent
AMLl-ETO occurs in ~15% of AML patients (Look 1997), and
(3) DNA hypermethylation occurs in AML (Herman 1999).
Aberrant methylation of pl5 occurred in 6 of 8, or 75% of
the AML patients analyzed in this study. This percentage
is consistent with a larger study (n = 60) that found a
higher percentage, 88% of AML patients, that had aberrrant
methylation of pl5 (Herman 1997). Thus, pl5 methylation is
the most frequent alteration described to date in AML.
Aberrant methylation of pl5 is a potential marker for
AML. The detection of hypermethylation at the pl5 5' CpG
island could be important for: (1) the detection of AML,
(2) accurate diagnosis of AML, or (3) monitoring residual
disease after treatment. The high frequency of alteration
at the pl5 CpG island makes it a potentially important
future tool to monitor AML.
109
As shown simplistically in figure 31, the efficacy of
prokaryotic drugs, such as chloramphenicol, is explained by
their targeting of the unique prokaryotic structures that
differ from euJcaryotic structures, such as the ribosome in
the case of chloramphenicol. The treatment of cancer is
complex because both normal and cancerous cells are
eukaryotic and use the same basic cellular machinery. Most
cancer drugs in current use do not kill cancer cells based
on specfic defects in cancer cells. Ara-C and
daunorubicin, two drugs currently used to treat AML, kill
all rapidly dividing cells including normal cells.
Further, ara-C and daunorubicin treatment fails in ~40% of
all AML cases. Novel pharmaceutical drugs that target
specific defects in cancer cells might have higher efficacy
than current treatments.
Presently, 5-azacytidine is the only widely explored
inhibitor of DNA methylation. If demethylation of pl5
results in reactivation of pl5 and this puts AML into
remission, then this is a potential avenue for drug
exploration. A drug that demethylates pl5 would likely act
at numerous other gene promoters potentially resulting in
deliterious pleiotropic effects. Such a situation could be
avoided if it were possible to design drugs that only
no
demethylate specific genes such as pl5. Of course the
design of drugs that could specfically demethylate pl5 and
not other genes may not be possible. In summary, given the
frequency of aberrant methylation of pl5 in AML and the
need for targetted therapies to treat AML, reactivation of
pl5 by DNA methylation inhibitors represents an avenue for
potential drug discovery.
p r c k a ' a n s
111
Figure 31. Epigenetic changes such as DNA methylation represent a potential drug target in cancer. As prokaryotic cells demonstrate, specific alterations such as the completely different ribosomal architecture, provide a target for the selective killing of prokaryotic cells but not eukaryotic cells. Her-2 a receptor overexpressed in breast cancer is a paradigm case of the targeted pharmacology of cancer. DNA methylation is frequently altered in human cancer, thus it represents a target for novel pharmaceutical approaches.
112
G«n«ral po^antial fox nov«l th«r«p«u^e i.nterv«nti.on of
apxgene^c dimt»ctM xn bwmmxx cmnemx
Cancer is a genetic disease and the alterations that
occur in cancer can be classified as: (1) viral, (2)
chromosomal rearrangements, (3) mutation, and (4)
epigenetic/DNA methylation. As shown in table 5,
epigenetics/DNA methylation is the only category that both
has an endogenous role and is a potential target for cancer
drugs. In the case of viruses, no known endogenous role
exists. Chromosomal translocations of DNA are an
endogenous process used in V(D)J recombination, however
despite the occurrence of chromosomal translocations in
cancer it does not appear to be an avenue for drug
treatment. If the process of shuffling DNA that occurs
physiologically in V(D)J recombination could be harnessed
to any end, it is not apparent how this would be useful to
treat cancer. Mutation has no well documented endogenous
role within the lifespan of individual mammalian organisms.
It is entirely possible that mutation enables selection and
ultimately helps ensure the survival of a given species,
but within the lifetime of an organism there is no current
evidence that mutations in DNA have a physiological role.
In fact, mutation may be considered a complete failure of
113
the endogenous processes of DNA repair systems that attempt
to maintain DNA sequences intact within the organisms
lifespan. The case of epigenetics/DNA methylation is
unique. Because DNA methylation is a endogenous process
with physiological consequences, alterations in DNA
methylation represent an exciting possibility for future
cancer pharmacology that might be able to use the
endogenous methylation machinery to revert a cancer cell to
a normal cell. Again, as stated previously there are many
potential drawbacks of such a strategy such as the lack of
specificity in DNA methylation changes that could have
deleterious consequences. Nevertheless, based simply on
the existance of endogenous processes that both remove and
add methylation to DNA, particularly during development,
this is an area of cancer research worthy of future
investigations.
114
Viral No None/ Yes HTLV-1
Chromsomal Yes V(D) J No rearrangement rearrange
ments/ AMLl-ETO
Mutation No None/p53 Gene therapy
Epigenetics, Yes Mammalian 5-aza, DNA development future
methylation /pl5 in AML drugs?
Table 5. A general classification of the alterations that occur in human cancer cells. DNA methylation represents an intriguing possibility because it is an endogenous process with physiological consequences. Thus pharmacological manipulation of the endogenous DNA methylation machinery might be able to revert cancer cells to normal cells. Currently, 5-aza = 5-azacytidine, is the only drug widely explored as an inhibitor of DNA methylation.
115
APPENDIX A. SUMnRT OF STODZSS TO TKST WBKTBSR KG-1 OR KG-
la CAN BE INDUCED WITHOUT DEMBTBTIATZON
One of the areas of research that was pursued but not
resolved involves whether the methylation of pl5 (observed
in the cell lines KG-1 and KG-la and in AML patients)
prevents the induction of pl5. From the original
description that TGF-P induces pl5 in HaCaT, it has been an
attractive idea that DNA methylation of pl5 blocks TGF-P
mediated pl5 induction. It has been published that
treatment with 5-azacytidine will result in reactivation of
pl5 in KG-la. The question here is whether an agent such
TGF-P, that is not a DNA methylation inhibitor, could induce
pl5.
First it was determined if leukemia cell lines also
show pl5 induction after TGF-p treatment. Because previous
studies demonstrated that HL60 lacks the TGF-p type II
receptor, HL60 was not included in this study. The cell
lines HaCaT (a positive control for TGF-P mediated induction
of pl5), U937, KG-1, and KG-la were tested. The results,
shown in figure A, show that only HaCaT had induced pl5
protein levels 24 hours after TGF-P (2ng/mL) treatment. It
could be argued that U937, KG-1, and KG-la are not
responsive to TGF-P, and this explains the lack of pl5
116
induction. To address this, cell cycle position v;as
determined by propidium iodide (PI) staining and flow-
assisted cell sorting (FACS), see figure B. This showed a
marked growth arrest only in HaCaT. Thus it may be the
case that the leukemia cell lines lack the necessary
signaling for TGF-P to induce pl5. Currently, there is no
precedent in the literature of a leukemia cell Xine that
shows induced pl5 after TGF-p treatment. Thus, these
results suggest that TGF-p mediated induction of pl5 may
only occur in specific cell types and not in myeloid cell
lines.
In addition to TGF-P inducing pl5 in HaCaT, it was
reported that PMA induces pl5 in HL60. Whether PMA would
also induce pl5 in U937, KG-1, and KG-la was determined.
The results shown at the bottom of figure A demonstrate
that both HL60 and U937 do induce pl5 in response to PMA
treatment, but KG-1 and KG-la, the two cell lines that
contain methylation of pl5, do not show pl5 induction (in
addition, no pl5 mRNA was detectable in KG-1 or KG-la, data
not shown). Also, PI staining followed by FACS analysis
showed that U937 and HL60 are both growth arrested after
PMA treatment (15ng/mL), see figure C. In constrast, KG-1
and KG-la did not show any difference in cell cycling. The
117
results obtained after PMA treatment were interesting
because KG-1 but not KG-la has previously been reported to
be PMA responsive, and KG-1 differentiates in response to
PMA treatment. Thus, the results show that after PMA
treatment of myeloid cell lines that lack aberrant pl5
methylation, pl5 is induced (HL60 and U937), but when
methylation is present pl5 (KG-1 and KG-la ), pl5 is not
induced. It still may simply be that KG-1 and KG-la are
lack responsiveness to PMA and a more direct experiment
would be homologous recombination of a methylated pl5 into
either HL60 or U937 to determine if this ablates PMA-
mediated pl5 induction.
118
1 2 3 4 5 6 7 8
pl5
8
Figure A. pl5 is induced by TGF-P in HaCaT, at the top is a Western blot for pl5, lanes: 1 = HaCaT, 2 = HaCaT +TGF-P, 3 = KG-1, 4 = KG-1 +TGF-P, 5 = KG-la, 6 = KG-la +TGF-p, 7 = U937, and 8 = U937 +TGF-P. Below is also shown Western blots for, from top to bottom, HL60, U937, KG-1, and KG-la. The lanes correspond to: 1 = HaCaT +TGF-p, 2 = untreated, 3-6=2, 4, 24, and 48 hour PMA treatment, 7 = 48 vehicle control treated, and 8 = pl5 protein.
119
HaCaT
+TGF-B *
vehicle
KG-la
vehicle KG-1
vehicle U937
vehicle
Figure B. Propidium iodide (PI) staining of cells followed by flow-assisted cytometry (FACS) after TGF-P treatment. As depicted with the asterisk (*), only HaCaT cells were growth arrested after TGF-P treatment.
120
61
vehicle
U937
+PMA*
vehicle
KG-I
vehicle
KG-la
vehicle
Figure C. Propidium iodide (PI) staining of cells followed by flow-assisted cytometry (FACS) after PMA treatment. As depicted with the asterisk (*), U937 and HL60 cells were growth arrested after PMA treatment.
121
APPENDIX B. SUMART OF D-BPLC AS A SZMPLB MSTBOD TO DSTXCT
ABBRRANT DNA MBTBTLATZCMf
We hypothesized that a cheaper and quicker method to
detect aberrant methyiation could be achieved by combining
sodium bisulfite modification and PGR amplification with
denaturing-high pressure liquid chromatography, D-HPLC. In
brief, we obtained sodium bisulfite modified DNA from U937,
KG-1, and KG-la. The 5' pl5 region then was amplified in
U937, KG-1, and KG-la by PGR. Figure A shows the
differences in the pl5 DNA methyiation patterns detected by
obtaining individual sequences of the cloned 5' pl5 sodium
bisulfite PGR products. In the 5' pl5 PGR product of
sodium bisulfite modified DNA there are ~10-12 base changes
in KG-1 and ~23 base changes in KG-la when compared to
U937. The PGR products were purified by ethanol
precipitation and quantitated with picogreen dye. The DNA
was then given to the D-HPLG core facility at the
University of Arizona. They ran 5 (iL of 15 ng/]xL of the pl5
sodium bisulfite modified PGR products from U937, KG-1, and
KG-la on a 55-70% 0.1 M triethylammonium acetate (TEAA) and
25% acetonitrile (AGN), 30-45% 0.1 M TEAA and 0.1% AGN,
gradient for 8 minutes at 58 degrees G on a DNAsep D-HPLG
column. The results obtained, shown in figure B, indicate
122
that D-HPLC is a valid method for distinguishing between
the three identified pl5 methylation states. The cost per
sample was $1.50. These results indicate that the use of
D-HPLC is a useful cheap and quick assay for determining
DNA methylation changes in any DNA sequence.
123
U937
100
80
60
:|La -MO -185
4u 120 -9
i i 49 113 182 297 367 436
KG-l
100
•440 -1*5 -120 "3 182 297 367 436
Figure A. The pl5 DNA methylation patterns segregates into unmethylated observed in U937, in blue, variegated methylation observed in KG-l, in red, and complete methylation observed in KG-la, in black. These three cell lines provide an ideal model to test whether D-HPLC can distinguish between the 5' pl5 PGR products of sodium bisulfite modified DNA.
124
I
I
r-r-f ̂ » 7 T
I I i i t I t < I i 1 I r-r
Figure C. Validation of concept: D-HPLC can distinguish between the 3 pl5 methylation patterns of U937, blue arrow, KG-1, red arrow, and KG-la, black arrow. The retention time of the KG-la PGR product is ~25 seconds greater than KG-1 and ~45 seconds greater than U937.
125
REFERENCES
Antonsson BE, Avramis VI, Nyce J, Holcenberg JS. (1987) Effect of S-azacytidine and congeners on DNA methylation and expression of deoxycytidine kinase in the human lymphoid cell lines CCRF/CEM/0 and CCRF/CEM/dCk-1. Cancer Res, 47, 3672-3678.
Batova A, Diccianni MB, Yu JC, Nobori T, Link MP, Pullen J, Yu AL. (1997) Frequent and selective methylation of pl5 and deletion of both pl5 and pl6 in T-cell acute lymphoblastic leukemia. Cancer Res, 57, 832-836.
Baylin SB, Makos M, Wu JJ, Yen RW, de Bustros A, Vertino P, Nelkin BD. (1991) Abnormal patterns of DNA methylation in human neoplasia: potential consequences for tumor progression. Cancer Cells, 3, 383-390.
Bestor TH. (1996) DNA methyltransferases in mammalian development and genome defense. In Epigenetic Mechanisms of Gene Regulation, Riggs AD, Martienssen RA, and Russo VEA editors. Cold Springs Harbor Laboratory Press.
Bhalla K, Nayak R, Grant S. (1984) Isolation and characterization of a deoxycytidine kinase-deficient human promyelocytic leukemic cell line highly resistant to 1-beta-D- arabinofuranosylcytosine. Cancer Res, 44, 5029-5037 .
Bhattacharya SK, Ramchandani S, Cervoni N, Szyf M. (1999) A mammalian protein with specific demethylase activity for mCpG DNA. Nature, 397, 579-583.
Bird AP. (1986) CpG-rich islands and the function of DNA methylation. Nature, 321, 209-213.
Bos JL, Verlaan-de VM, van der Eb AJ, Janssen JW, Delwel R, Lowenberg B, Colly LP. (1987) Mutations in N-ras predominate in acute myeloid leukemia. Blood, 69, 1237-1241.
Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Markham A, Fusenig NE. (1988) Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. J Cell Biol, 106, 761-771.
126
Chen EH, Johnson EE, Vetter SM, Mitchell BS. (1995) Characterization of the deoxycytidine kinase promoter in human lymphoblast cell lines. J Clin Invest, 95, 1660-1 6 6 8 .
Clark SJ, Harrison J, Paul CL, Frommer M. (1994) High sensitivity mapping of methylated cytosines. Nucleic Acids Res, 22, 2990-2997.
Compere SJ, Palmiter RD. (1981) DNA methylation controls the inducibility of the mouse metallothionein- I gene lymphoid cells. Cell, 25, 233-240.
Csankovszki G, Panning B, Bates B, Pehrson JR, Jaenisch R. (1999) Conditional deletion of Xist disrupts histone macroH2A localization but not maintenance of X inactivation. Nat Genet, 22, 323-324.
Denissenko MF, Chen JX, Tang MS, Pfeifer GP. (1997) Cytosine methylation determines hot spots of DNA damage in the human P53 gene. Proc Natl Acad Sci USA, 94, 3893-3898 .
Feinberg AP, Gehrke CW, Kuo KC, Ehrlich M. (1988) Reduced genomic 5-methylcytosine content in human colonic neoplasia. Cancer Res, 48, 1159-1161.
Flasshove M, Strumberg D, Ayscue L, Mitchell BS, Tirier C, Heit W, Seeber S, Schutte J. (1994) Structural analysis of the deoxycytidine kinase gene in patients with acute myeloid leukemia and resistance to cytosine arabinoside. Leukemia, 8, 780-785.
Fryxell KB, McGee SB, Simoneaux DK, Willman CL, Cornwell MM. (1999) Methylation analysis of the human multidrug resistance 1 gene in normal and leukemic hematopoietic cells. Leukemia, 13, 910-917.
Gardiner-Garden M, Frommer M. (1987) CpG islands in vertebrate genomes. J Mol Biol, 196, 261-282.
Gossen M, Bujard H. (1992) Tight control of gene expression in mammalian cells by tetracycline- responsive promoters. Proc Natl Acad Sci USA, 89, 5547-5551.
127
Graff JR, Greenberg VE, Herman JG, Westra WH, Boghaert ER, Ain KB, Saji M, Zeiger MA, Ziinmer SG, Baylin SB. (1998) Distinct patterns of E-cadherin CpG island methylation in papillary, follicular, Hurthle's cell, and poorly differentiated human thyroid carcinoma. Cancer Res, 58, 2063-2066.
Grant S, Turner A, Nelms P, Yanovich S. (1995) Characterization of a multidrug resistant human erythroleukemia cell line (K562) exhibiting spontaneous resistance to 1-beta-D- arabinofuranosylcytosine. Leukemia, 9, 808-814.
Hannon GJ, Beach D. (1994) pl5INK4B is a potential effector of TGF-beta-induced cell cycle arrest. Nature, 371, 257-2 6 1 .
Herman JG, Jen J, Merlo A, Baylin SB. (1996) Hypermethylation-associated inactivation indicates a tumor suppressor role for pl5INK4B. Cancer Res, 56, 722-727.
Herman JG, Civin CI, Issa JP, Collector MI, Sharkis SJ, Baylin SB. (1997) Distinct patterns of inactivation of pl5INK4B and pl6INK4A characterize the major types of hematological malignancies. Cancer Res, 57, 837-841.
Herman JG. (1999) Hypermethylation of tumor suppressor genes in cancer. Semin Cancer Biol, 9, 359-367.
Holliday R,- Pugh JE. (1975) DNA modification mechanisms and gene activity during development. Science, 187, 226-232.
Holliday R. (1979) A new theory of carcinogenesis. Br J Cancer, 40, 513-522.
Holliday R, (1996) DNA methylation in eukaryotes: 20 years on. In Epigenetic Mechanisms of Gene Regulation, Riggs AD, Martienssen RA, and Russo VEA editors. Cold Springs Harbor Laboratory Press.
Hussussian CJ, Struewing JP, Goldstein AM, Higgins PA, Ally DS, Sheahan MD, Clark WH, Tucker MA, Dracopoli NC. (1994) Germline pl6 mutations in familial melanoma. Nat Genet, 8, 15-21.
128
Jamal R, Thomas NS, Gale RE, Linch DC. (1996) Variable expression of pl6 protein in patients with acute myeloid leukemia without gross rearrangements at the DNA level. Leukemia, 10, 629-636.
Jen J, Harper JW, Signer SH, Papadopoulos N, Markowitz S, Willson JK, Kinzler KW, Vogelstein B. (1994) Deletion of pl6 and pl5 genes in brain tumors. Cancer Res, 54, 6353-6358 .
Jones PA. (1985) Altering gene expression with 5-azacytidine. Cell, 40, 485-486.
Kamb A, Gruis NA, Weaver-Feldhaus J, Liu Q, Harshman K, Tavtigian SV, Stockert E, Day RS, Johnson BE, Skolnick MH. (1994) A cell cycle regulator potentially involved in genesis of many tumor types. Science, 264, 436-440.
Knudson AGJ. (1971) Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci USA, 68, 820-823.
Koeffler HP, Golde DW. (1978) Acute myelogenous leukemia; a human cell line responsive to colony- stimulating activity. Science, 200, 1153-1154.
Koeffler HP, Billing R, Lusis AJ. Sparkes R, Golde DW. (1980) An undifferentiated variant derived from the human acute myelogenous leukemia cell line (KG-1). Blood, 56, 265-273.
Kong XB, Tong WP, Chou TC. (1991) Induction of deoxycytidine kinase by 5-azacytidine in an HL-60 cell line resistant to arabinosylcytosine. Mol Pharmacol, 39, 250-257 .
Kwon H, Pelletier N, DeLuca C, Genin P, Cisternas S, Lin R, Wainberg MA, Hiscott J. (1998) Inducible expression of IkappaBalpha repressor mutants interferes with NF-kappaB activity and HIV-1 replication in Jurkat T cells. J Biol Chem, 273, 7431-7440.
Leegwater PA, De Abreu RA, Albertioni F. (1998) Analysis of DNA methylation of the 5' region of the deoxycytidine kinase gene in CCRF-CEM-sensitive and cladribine (CdA)- and 2-chloro-2'- arabino-fluoro-2'-deoxyadenosine (CAFdA)-resistant cells. Cancer Lett, 130, 169-173.
129
Li E, Bestor TH, Jaenisch R. (1992) Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell, 69, 915-926.
Look AT. (1997) Oncogenic transcription factors in the human acute leukemias. Science, 278, 1059-1064.
Manome Y, Wen PY, Dong Y, Tanaka T, Mitchell BS, Kufe DW, Fine HA. (1996) Viral vector transduction of the human deoxycytidine kinase cDNA sensitizes glioma cells to the cytotoxic effects of cytosine arabinoside in vitro and in vivo. Nat Med, 2, 567-573.
Munro J, Stott FJ, Vousden KH, Peters G, Parkinson EK. (1999) Role of the alternative INK4A proteins in human keratinocyte senescence: evidence for the specific inactivation of pl6INK4A upon immortalization. Cancer Res, 59, 2516-2521.
Neuveut C, Low KG, Maldarelli F, et al. (1998) Human T-cell leukemia virus type 1 Tax and cell cycle progression: role of cyclin D-cdk and pllORb. Mol Cell Biol, 18, 3620-3632.
Nilsson K, Sundstrom C. (1974) Establishment and characteristics of two unique cell lines from patients with lymphosarcoma. Int J Cancer, 13, 808-823.
Ohtani-Fujita N, Fujita T, Aoike A, Osifchin NE, Robbins PD, Sakai T. (1993) CpG methylation inactivates the promoter activity of the human retinoblastoma tumor-suppressor gene. Oncogene, 8, 1063-1067.
Okano M, Xie S, Li E. (1998) Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat Genet, 19, 219-220.
Okano M, Bell DW, Haber DA, Li E. (1999) DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell, 99, 247-257 .
Razin A, Riggs AD. (1980) DNA methylation and gene function. Science, 210, 604-610.
130
Rhim JS, Fujita J, Arnstein P, Aaronson SA. (1986) Neoplastic conversion of human keratinocytes by adenovirus 12-SV40 virus and chemical carcinogens. Science, 232, 385-388 .
Rhim JS, Jay G, Arnstein P, Price FM, Sanford KK, Aaronson SA. (1985) Neoplastic transformation of human epidermal keratinocytes by AD12-SV40 and Kirsten sarcoma viruses. Science, 227, 1250-1252.
Riggs AD. (1975) X inactivation, differentiation, and DNA methylation. Cytogenet Cell Genet, 14, 9-25.
Rougier N, Bourc'his D, Gomes DM, Niveleau A, Plachot M, Paldi A, Viegas-Pequignot E. (1998) Chromosome methylation patterns during mammalian preimplantation development. Genes Dev, 12, 2108-2113.
Schwaller J, Pabst T, Koeffler HP, Niklaus G, Loetscher P, Fey MF, Tobler A. (1997) Expression and regulation of G1 cell-cycle inhibitors (pl6INK4A, pl5INK4B, pl8INK4C, pl9INK4D) in human acute myeloid leukemia and normal myeloid cells. Leukemia, 11, 54-63.
Serrano M, Hannon GJ, Beach D. (1993) A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4. Nature, 366, 704-707.
Taylor SM, Jones PA. (1979) Multiple new phenotypes induced in lOTl/2 and 3T3 cells treated with 5- azacytidine. Cell, 17, 771-779.
Tobal K, Pagliuca A, Bhatt B, Bailey N, Layton DM, Mufti GJ. Mutation of the human FMS gene (M-CSF receptor) in myelodysplastic syndromes and acute myeloid leukemia. Leukemia, 4, 486-489.
Vinante F, Rigo A, Papini E, Cassatella MA, Pizzolo G. (1999) Heparin-binding epidermal growth factor-like growth factor/diphtheria toxin receptor expression by acute myeloid leukemia cells. Blood, 93, 1715-1723.
Walsh CP, Chaillet JR, Bestor TH. (1998) Transcription of TAP endogenous retroviruses is constrained by cytosine methylation. Nat Genet, 20, 116-117.
131
Watts GS, Pieper RO, Costello JF, Peng YM, Dalton WS, Futscher BW. (1997) Methylation of discrete regions of the 06-methylguanine DNA methyltransferase (MGMT) CpG island is associated with heterochromatinization of the MGMT transcription start site and silencing of the gene. Mol Cell Biol, 17, 5612-5619.
Weinberg RA. (1995) The retinoblastoma protein and cell cycle control. Cell, 81, 323-330.
Wigler M, Levy D, Perucho M. (1981) The somatic replication of DNA methyation. Cell, 24, 33-40.
Yoder JA, Walsh CP, Bestor TH. (1997) Cytosine methylation and the ecology of intragenomic parasites. Trends Genet, 13, 335-340.