HHS, NCI, J Natl Cancer Inst, May 18, 1994 (1)

Full text

Page
1
1 of 1 DOCUMENT
1994, U.S. Department of Health and Human Services;
Journal of the National Cancer Institute
J Natl Cancer Inst 1994; 86: 753-759
May 18, 1994
SECTION: COMMENTARY
LENGTH: 3993 words
TITLE: Genomic Imprinting, DNA Methylation, and Cancer
AUTHOR: Shirley, Rainier, Andrew P. Feinberg <1>
TEXT:
Our basic concepts of segregation and assortment of two alleles of a gene have been shaped by the work of Gregor
Mendel [n1]. Similarly, our notions of genetic disorders were set more than 75 years ago by Garrod [n2], who distin-
guished patterns of inheritance that depended on whether one or both copies of a gene are needed for normal function.
Fig. 1 illustrates these classical concepts of mendelian inheritance of dominant and recessive genetic traits. There are
two important assumptions implicit in mendelian inheritance: 1) that the materially and paternally inherited alleles of a
gene are identical and 2) that two working expressed copies of a normal gene are always associated with normal function.
Genomic imprinting challenges both of these assumptions.
Genomic imprinting is defined as a reversible modification of DNA that causes differential expression of maternally
and paternally inherited homologous chromosomes or genes (Fig. 2). The nature of this modification is unknown but
must 1) involve the gone itself or the chromosome on which it resides, 2) occur during gametogenesis or shortly after
fertilization, 3) be heritable by somatic cells during cell division, and 4) be potentially reversible, as the imprint can be
changed during reproduction of the organism.
Genomic imprinting has been known to exist in lower organisms for many years, but it has attracted attention recently
because it is now known to be essential for normal development in the mouse and because a molecular mechanism, DNA
methylation, has been tentatively identified [reviewed in [n3]]. While there had been reasons to suspect that imprinting
was important for humans as well, only recently has molecular evidence for imprinting been reported in humans [n4-n8].
Evidence of Genomic Imprinting in Other Species
The greatest insight into mammalian imprinting tonics from several lines of experimentation in mice (Fig. 3). First,
androgenetic embryos (with two complete paternal chromosomal complements) fail to develop normal embryonic tissues
[n9], while parthenogenetic embryos (with two complete maternal chromosomal complements) fail to develop extraem-
bryonic tissues [n9,n10]. These experiments demonstrate that maternal and paternal genome complements are necessary
for normal embryonic development in the mouse. Second, mice harboring certain balanced translocations that result in
uniparental disomy (both copies of specific chromosomal regions are derived from a single parent) develop growth and
developmental abnormalities [n11]. A direct relationship of imprinting to the regulation of cell growth is suggested by
the flier that, for some chromosomal regions, paternal uniparental disomy leads to increased growth, and maternal unip-
arental disomy leads to decreased growth [n12]. These findings have led to the hypothesis of Moore and Haig [n13],
which suggests that imprinting evolved because of opposing paternal and maternal influences in the growth of the em-
bryo. These studies have also provided important mapping data for imprinted regions in the mouse and, by virtue of
comparative mapping, have suggested potentially imprinted regions in the human as well.
Work on transgenic animals, which contain a stably integrated foreign gene, has provided further evidence of im-
printing. Many transgenes are expressed only when inherited from a specific parent [n14]. This imprinting of the
transgene is also associated with specific differences in DNA methylation between the active and imprinted (inactive)

Page
2
HHS, NCI, J Natl Cancer Inst, May 18, 1994
alleles [n15,n16], thus suggesting a role for DNA methylation in the control of imprinting. DNA methylation is the
enzymatic addition of a methyl group to the 5-position of the cytidine ring in genomic DNA by DNA methyltransferase.
Other examples of imprinting include the preferential inactivation of the paternal X chromosotne in kangaroos and other
marsupials [n17] and in the extraembryonic tissue of rodents [n18].
Finally, there is direct molecular evidence for imprinting of the following six mouse genes: 1) the insulin-like growth
factor 2 gene (Igf2), expressed only from the paternal allele [n19,20): 2) H19, expressed only from the maternal allele
[n21]; 3) the Igf2 receptor gene (Igf2), expressed only from the maternal allele [n22]: 4) Snrpn (small nuclear ribonu-
cleoprotein-associated polypeptide SmN), expressed only from the paternal allele [n23,n24]; 5) insulin 1 and 2, expressed
only from the paternal allele in the yolk sac [n25]: and 6) U2afbp-rs or SP2, expressed only from the paternal allele
[n26,n27]. Since genomic imprinting plays an important role in normal development in other species, it might play a
similarly critical role in humans.
Evidence for Genomic Imprinting in Humans
Compelling evidence for genomic imprinting in human disease comes from two disorders that show uniparental
disomy (Fig. 3): Prader-Willi syndrome (PWS) and Beckwith-Wiedemann syndrome (BWS). PWS involves either
maternal uniparental disomy for band 15q11-12 [n28] or visible cytogenetic deletion of the same region from the paternal
chromosome [n29]. PWS is characterized by short stature, mental retardation, and uncontrolled appetite leading to
massive obesity. BWS is characterized by prenatal overgrowth and an increased incidence of childhood tumors such as
Wilms' tumor, hepatoblastoma, and rhabdomyosarcoma. BWS involves paternal uniparental disomy of 11p15 in some
patients [n30]. Other BWS patients show balanced rearrangements involving the maternally inherited chromosome
[n31] or deletions involving the maternal chromosome [n32]. The genetics of both PWS and BWS imply that the ma-
ternal and paternal alleles of these genes are expressed differentially in normal development, and the disorder arises from
an abnormal dose of these genes (either decreased or increased). For example, if the BWS gene is maternally imprinted
(i.e., the paternal copy is expressed), then paternal uniparental disomy would result in a net increased dose of the BWS
gene. A gene has been isolated in the smallest region of deletion in PWS patients at 15q11-12, SNRPN [n33]. This gene
encodes a small nuclear RNA molecule that is maternally imprinted in mice [n23] and humans [n34]. The location and
imprinting of SNRPN suggest a role for this gene in PWS, although there is no direct evidence of its involvement.
The existence in the mouse of several imprinted genes, Igf2 and H19, in the region of mouse chromosome 7 ho-
mologous to the human BWS region [n11p15] has suggested a role for these genes in this syndrome [n13,n35,n36].
Moreover, genomic imprinting of these two genes in humans has recently been demonstrated directly in our laboratory
[n4] and in other laboratories [n5-n8] by examining RNA from a variety of tissues. Messenger RNA isolated from
normal kidney tissue was reverse transcribed into complementary DNA, amplified using polymerase chain reaction, and
restricted with an enzyme that recognizes a transcribed polymorphism, to determine which parental allele was expressed.
IGF2 was expressed from the paternal allele, and H19 was expressed from the maternal allele.
Several hereditary diseases, including Huntington's disease and fragile X syndrome, show variable penetrance, de-
pending on the parent from whom the gene is inherited, suggesting a role for imprinting in the disease process. However,
molecular investigation of these diseases has shown that these are special cases involving expansion of trinucleotide
repeat regions in genomic DNA [n37,n38]. In both Huntington's disease and fragile X syndrome, the two parental alleles
are not known to be differentially expressed normally. In Huntington's disease, the effect of the expansion of the trinu-
cleotide repeat is unknown. However, in fragile X syndrome, a mutant allele undergoes further mutation (additional
lengthening of a trinucleotide repeat) at a higher frequency when passed through the maternal gamete, resulting in the
inactivation of the FMR1 gene [n39]. This differential allele mutability, while loosely referred to as a form of imprinting,
does not represent differential allele expression.
Alterations in Genomic Imprinting in Human Cancer
The potential role of genomic imprinting in human cancer is suggested by several lines of evidence (Fig. 3). Two
unusual human tumors appear to arise from an abnormally imprinted genome. Hydatidiform mole, a uterine mass of
cysts resembling a cluster of grapes, arises from fertilization of an "empty egg," resulting in a totally androgenetic ge-
nome, i.e., 46 chromosomes all from the father [n40]. Conversely, complete ovarian teratoma, an unusual tumor that
presents as an ovarian cyst often containing teeth and hair, develops in nonovulated oocytes in the ovary, resulting in a
totally parthenogenetic genome, i.e., 46 chromosomes from the mother [n41]. Additional evidence for imprinting in
cancer comes from hereditary cancer syndromes such as hereditary paraganglioma, which is inherited exclusively from

Page
3
HHS, NCI, J Natl Cancer Inst, May 18, 1994
the father [n42], and possibly familial adenomatous polyposis (FAP), which may show a preference for paternal inher-
itance [n43].
Another major line of evidence is the preferential involvement of a specific parental allele in loss of tumor suppressor
genes. Molecular evidence for the loss of specific tumor suppressor genes comes from a decade of observations begin-
ning with two childhood tumors, Wilms' tumor and retinoblastoma. These tumors frequently show loss of heterozygosity
(LOH) of polymorphic markers on specific chromosomal arms. For example, in Wilms' tumor, using restriction frag-
ment length polymorphisms that distinguish the maternal and paternal chromosomes, one frequently finds LOH or loss of
one of the polymorphic markers of lip in a tumor, indicating that the chromosome arm has been lost and that it contains a
tumor suppressor gene that has also been lost [n44]. Subsequently, LOH has been found in a wide variety of common
tumors and has led to the localization and eventual cloning of a series of important tumor suppressor genes, such as DCC
and FAP in human colorectal cancer [n45,n46]. Interestingly, during the analysis of LOH in tumors, several neoplasms
have shown LOH that preferentially involves a specific parental allele. For example, maternal chromosome 13 was lost
in sporadic osteosarcoma [n47], paternal chromosome 7 was lost in acute myelogenous leukemia (AML) [n48], and
maternal chromosome 11 was preferentially lost in Wilms' tumor, rhabdomyosarcoma, and hepatoblastoma [n49]. These
results are consistent with unequal expression of two alleles, depending on the parental origin, with loss of the active allele
which encodes a tumor suppressor. Chronic myelogenous leukemia (CML) also showed strong indirect evidence for
genomic imprinting in carcinogenesis. In a recent study [n50], the Philadelphia chromosome translocation was shown to
involve exclusively the paternal chromosome 9 and the maternal chromosome 22. In summary, considerable indirect
evidence suggests that genomic imprinting may play an important role in a variety of human cancers.
The first direct evidence that abnormally imprinted genes may play a role in tumorigenesis was found by examining
imprinting of IGF2 and H19 in Wilms' tumors. We [n4] and others [n6] have shown that normal imprinting of IGF2 is
lost during tumorigenesis. In normal human kidney, the paternal allele of IGF2 is expressed and the maternal allele is
transcriptionally inactive. Seventy percent of Wilms' tumors show loss of imprinting (LOI) and thus express both the
maternal and paternal alleles (Fig. 2) [n4,n6]. Our laboratory [n4] also has shown that LOI of H 19 occurs in these
tumors but at a lower frequency (< 20%). LOI was present in the earliest stage of tumors as frequently as in later stage
tumors. Thus, LOI appears frequently in at least some human cancers and, of course, early in the development of ma-
lignancy. H19 is a gene that apparently encodes an RNA, the function of which is still unknown [n51]. The importance
of this gene in normal development was demonstrated in transgenic mouse studies [n52]. Overexpression or deregulated
expression of H19 in these transgenic mice caused embryonic lethality early in development. Therefore, LOI of HI9,
although not as common as LOI of IGF2, may also be an important step in tumorigenesis.
LOI is apparently not restricted to Wilms' tumor, since we [n4] have also demonstrated LOI of IGF2 in a rhabdoid
tumor of the kidney. Furthermore, Weksberg et al. [n53] and our laboratory (Rainier S, Feinberg AP: unpublished data)
have recently found that LOI is present in normal fibroblasts of BWS patients (Fig. 2). These patients have an increased
risk for Wilms' tumor and for other tumors. Thus, LOI may predispose these patients to tumor development.
What might be the potential effects of LOI in carcinogenesis? Fig. 4 summarizes some possibilities. The most
obvious potential effect is an increase in expression of the gene that has undergone LOI, namely from expression of one
copy to expression of two copies of the gene. Many other genetic alterations in cancer can cause abnormal expression of
normal cellular genes, such as DNA amplification of N-myc in neuroblastoma [n54] and MDM2 in sarcoma [n55] and
translocation of c-myc in Burkitt's lymphoma [n56,57]. LOI of IGF2 could be an important step in carcinogenesis by
causing increased levels of IGF2. That such a gene activation mechanism may be important is suggested by the role of
IGF2 as an important autocrine growth factor in a wide variety of tumors, including lung [n58-n60], breast [n61,n62],
colon [n63-n66], thyroid [n67], liver [n68,n69], and brain [n70,n71] tumors.
Possible therapeutic implications of LOI are suggested by the fact that the blockade of Igf2 by suramin at the insu-
lin-like growth factor 1 receptor (Igflr), the receptor that carries out the growth function of Igf2 [n72], causes growth
inhibition in vitro of human rhabdomyosarcoma [n73]. Clinical trials are under way to inhibit IGF2 at its receptor in
rhabdomyosarcoma. In Wilms' tumor, one proposed mechanism for the activation of IGF2 is mutation of the WT1 gene
on 11p13, a transcriptional repressor of IGF2 and other genes [n74]. However, the frequency of mutation of WTI in
Wilms' tumor is less than 10% [n75]. Thus, LOI may be a more common change leading to deregulated expression of
IGF2 in Wilms' tumor.
An alternative mechanism, first suggested by Sapienza [n76] is that abnormal imprinting could inactivate one copy of
a tumor suppressor gene. Thus, preferential LOH may be due to abnormal imprinting and inactivation of the retained
allele during early embryogenesis [reviewed in [n77]]. If the lost gene is a tumor suppressor gene, its inactivation may

Page
4
HHS, NCI, J Natl Cancer Inst, May 18, 1994
lead to cancer. For example, if H19 is a tumor suppressor gene, as is suggested by the recent report [n78], loss of the
maternal allele in Wilms' tumors would cause loss of the active copy of H19 and presumably decreased H19 messenger
RNA levels.
A third potential mechanism is a more complex interaction between neighboring loci that are reciprocally imprinted.
At least in the case of IGF2 and H19, one gene is expressed from the paternal chromosome and the other expressed from
the maternal chromosome. Thus, a change in imprinting of either gene could affect the expression of the other. In mice,
Igf2 and H19 map to a region of less than 90 kilobases; thus, a regulatory domain lying between them may act as an
allele-specific switch controlling the exclusive expression of Igf2 or H19 [n79,n80]. Hence, LOI of IGF2 may affect the
expression of H19. Therefore, LOH and LOI of IGF2 may lead to the same result, decreased HI9 expression. Whatever
the mechanism, LOI would be expected to lead to abnormal expression of imprinted genes in the cell.
Potential Molecular Mechanisms for Imprinting
What is the mechanism of maintenance of an imprint during normal development and LOI in tumorigenesis? DNA
methylation has been proposed as one possible mechanism [n14], and there are a number of studies [n81-n87] linking
DNA methylation with gene expression. Three lines of evidence suggest that DNA methylation may be at least in part
responsible for imprinting. First, imprinted transgenes often demonstrate hypomethylation of the transgene when
transmitted on the parental allele that is expressed, when compared with the parental allele that is not expressed
[n16,n88,n89]. Second, recent studies of Igf2r in mouse embryos with maternal duplication and paternal deletion of
distal chromosome 7 reveal a specific site of DNA methylation that distinguishes the paternal (active) and maternal (in-
active) alleles [n90], although it remains to be proven that DNA methylation controls the imprint. Interestingly, IGF2r is
not imprinted in humans [n91]: thus, there is not a complete correlation between imprinted genes in mice and humans.
Third, DNA methylation also appears to be an important mechanism in the regulation of genomic imprinting in plants, and
5-azacytidine (a DNA demethylating agent) relaxes these imprints, revealing previously masked variation in plant hybrids
[n92].
Studies of X-chromosome inactivation in humans also suggest that DNA methylation plays a role in the inhibition of
transcription of genes on the inactive X chromosome. Many genes show differential methylation on the active and
inactive X chromosomes [n93]. The regulation of X-chromosome inactivation is complex and involves factors in addi-
tion to DNA methylation [n94]. Nevertheless, reduction in DNA methylation with 5-azacytidine can cause reactivation
of genes on the inactive X chromosome [n95]. Thus, if X-chromosome inactivation is used as an analogy, even if DNA
methylation is not the primary mechanism for genomic imprinting, altered DNA methylation might still be a factor in its
loss in cancer.
Since cancer may involve the deregulated expression of normal cellular genes and the loss of DNA methylation is
associated with gene expression, Holliday [n96] first suggested that a failure to methylate DNA could disrupt gene reg-
ulation and lead to cancer. Experimental studies of this subject, using cell lines, gave contradictory results [reviewed in
[n97]]. Feinberg and Vogelstein [n98] addressed this problem several years ago by comparing DNA methylation of
specific genes in primary human colon and lung cancers to that in the adjacent normal tissue from the same patient. With
the use of methylcytosine-sensitive restriction enzymes, substantial hypomethylation of a wide variety of genes was found
in all 23 tumors studied. These genes included gamma-crystallin, whose hypomethylation was surprising in these tu-
mors, since this gene is not expressed and fully methylated in the normal tissues [n98]. Subsequently, 10 premalignant
benign colonic adenomas, from which colon cancers arise, were found to be similarly hypomethylated [n99,n100]. In
addition, several cellular oncogenes are also hypomethylated [n101]. Thus, abnormal DNA methylation is an attractive
potential mechanism for LOI in cancer. The changes in DNA methylation that occur in normally unexpressed genes,
such as gamma-crystallin, may lead to a cryptic response [a change in the competence or stability of a gene without an
obvious change in transcription [n102]] due to the absence of factors necessary for expression. In contrast, changes in
DNA methylation of imprinted genes may lead to a more visible response [n102], deregulated gene expression, since the
transcription factors necessary for the expression of imprinted genes should be present in the cell expressing those genes.
Recently, normal DNA methylation has been shown to be essential for normal imprinting in mice [n103]. By ex-
amining imprinted genes in DNA methyltransferase knock-out mice. Li et al. [n103] found activation of the normally
silent paternal H19 allele and inactivation of normally active paternal Igf2 allele and maternal Igf2r allele. These studies
support the hypothesis that abnormal imprinting is related to abnormal DNA methylation in cancer.
Another proposed mechanism for maintenance of imprinting comes from studies of Drosophila heterochromatin
[n104], transcriptionally silent regions of the Drosophila genome, by virtue of its folding into a tight complex with spe-

Page
5
HHS, NCI, J Natl Cancer Inst, May 18, 1994
cific proteins. At least two protein components essential for heterochromatin formation in Drosophila are known. A
normally active gene, when translocated next to heterochromatin, can be transcriptionally silenced. This silencing is
heritable. When one of the factors necessary for heterochromafin formation is limiting, it affects the scope of hetero-
chromatin formation and can release transcriptionally silenced genes from the silencing effects of heterochromatin
[n105]. Increasing the expression of this limiting factor can also increase the scope of heterochromatin formation and
can repress neighboring germs that are normally expressed in the cell. Therefore, there is a critical concentration of these
factors necessary to maintain heterochromatin and thus the normal expression of genes in these cells. One can envision
an analogous mechanism in humans. In fact, the human homologue of one of the components of Drosophila hetero-
chromatin (HP1) has been cloned [n106]. These observations in other organisms help to predict models and to design
experiments to explain the mechanism of imprinting in humans.
Recent work [n107] examining replication timing in imprinted genes indicated that there are large domains sur-
rounding imprinted genes that show early replication of the paternal allele. Unimprinted genes located within these
domains are also early replicating: therefore, early replication of the paternal allele does not correlate with allele-specific
expression of the imprinted genes. For example, H19 and Igf2 are expressed from different alleles, yet show early rep-
lication of the paternal allele. Therefore, the authors [n107] suggest a more complex mechanism for allele-specific
expression involving local trans-acting factors. Finally, another possible mechanism for imprinting derived from studies
in yeast is gene conversion, the transposition and substitution of one gene locus for another gene locus. In yeast, gene
conversion plays a role in controlling mating type [n108]. The region of the yeast genome that controls mating type
contains two genes, separated by a control or "switch" region. Only one of these genes is active in a cell, but a daughter
cell can switch mating types by inactivating the active gene and activating the inactive gene. The daughter cell can then
pass on the new phenotype to its progeny.
One of the most exciting aspects of the study of genomic imprinting and its role in carcinogenesis is its potential
therapeutic value. If the allele-specific information necessary for normal genomic imprinting is still present in the cancer
cell with LOI, it may be possible to manipulate these cells pharmacologically or genetically to restore the imprint.
Certainly, LOI is already important in cancer prevention, since LOI is seen in the normal tissues of patients with hered-
itary predisposition to cancer [n53]. It may become possible in the near future to perform screening for patients at risk for
cancer and possibly in the population at large.
Genomic imprinting and its role in cancer constitute an exciting new area of investigation that represents a conver-
gence of the fields of human cancer molecular genetics and mouse, plant, yeast, and Drosophila developmental genetics.
Fig. 4 summarizes just a few of the many questions to be answered.
SUPPLEMENTARY INFORMATION: A. P. Feinberg, Departments of Internal Medicine and Human Genetics,
University of Michigan Medical School, and Howard Hughes Medical Institute, University of Michigan.
<1> Affiliations of authors: S. Rainier, Department of Internal Medicine, University of MIchigan Medical School, and
Howard Hughes Medical Institute, University of Michigan, Ann Arbor.
Correspondence to present address: Andrew P. Feinberg, M.D., M.P.H., Department of Internal Medicine, Johns
Hopkins School of Medicine, 720 Rutland Ave., 1064 Ross Bldg., Baltimore, MD 21205.
Manuscript received November 15, 1993; revised January 19, 1994; accepted February 18, 1994.
REFERENCES:
[n1.] Mendel G: Versuche uber Pflanzen-Hybriden, Verh Naturforsch Ver Brun 4:3-47, 1866
[n2.] Garrod AE: The incidence of alkaptonuria: a study of chemical individuality. Lancet 2:1616-1620, 1902
[n3.] Monk M: Genomic imprinting. Genes Dev 2:921-925, 1988
[n4.] Rainier S, Johnson LA, Dobry CJ, et al: Relaxation of imprinted genes in human cancer. Nature 362:747-749, 1993
[n5.] Rachmilewitz J, Goshen R, Ariel I, et al: Parental imprinting of the human H19 gene. FEBS Lett 309:25-28, 1992

Page
6
HHS, NCI, J Natl Cancer Inst, May 18, 1994
[n6.] Ogawa O, Eccles MR, Szeto J, et al: Relaxation of insulin-like growth factor II gene imprinting implicated in Wilms'
tumour. Nature 362:749-751, 1993
[n7.] Zhang Y, Tycko B: Monoallelic expression of the human H19 gene. Nat Genet 1:40-44, 1992
[n8.] Giannoukakis N, Deal C, Paquette J, et al: Parental genomic imprinting of the human IGF2 gene. Nat Genet
4:98-101, 1993
[n9.] Surani MA, Kothary R, Allen ND, et al: Genome imprinting and development in the mouse. Dev Suppl: 89-98, 1990
[n10.] Clarke HJ, Varmuza S, Prideaux VR, et al: The developmental potential of parthenogenetically derived cells in
chimeric mouse embryos: implications for action of imprinted genes. Development 104:175-182, 1988
[n11.] Cattanach BM, Beechey CV: Autosomal and X-chromosome imprinting. Dev Suppl:63-72, 1990
[n12.] Barton SC, Ferguson-Smith AC, Fundele R, et al: Influence of paternally imprinted genes on development. De-
velopment 113:679-687, 1991
[n13.] Moore T, Haig D: Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet 7:45-49,
1991
[n14.] Swain JL, Stewart TA, Leder P: Parental legacy determines methylation and expression of an autosomal transgene:
a molecular mechanism for parental imprinting. Cell 50:719-727. 1987
[n15.] Sasaki H, Hamada T, Ueda T, et al: Inherited type of allelic methylation variations in a mouse chromosome region
where an integrated transgene shows methylation imprinting. Development 111:573-581, 1991
[n16.] Chaillet JR, Vogt TF, Beier DR, et al: Parental-specific methylation of an imprinted transgene is established during
gametogenesis and progressively changes during embryogenesis. Cell 66:77-83, 1991
[n17.] Cooper DW, Vandenberg JL, Sharman GB, et al: Phosphoglycerate kinase polymorphism in kangaroos provides
further evidence for paternal X inactivation. Nature New Biol 230:155-157, 1971
[n18.] Takagi N, Sasaki M: Preferential inactivation of the paternally derived X chromosome in the extraembryonic
membranes in the mouse. Nature 256:640-642, 1975
[n19.] DeChiara TM, Robertson EJ, Efstratiadis A: Parental imprinting of the mouse insulin-like growth factor II gene.
Cell 64:849-859, 1991
[n20.] Rappolee DA, Sturm KS, Behrendtsen O, et al: Insulin-like growth factor II acts through an endogenous growth
pathway regulated by imprinting in early mouse embryos. Genes Dev 6:939-952, 1992
[n21.] Barlolomei M, Zemel S, Tilghman SM: Parental imprinting of the mouse H19 gene. Nature 351:153 155, 1991
[n22.] Barlow DP, Stoger R, Herrmann BG, et al: The mouse insulin-like growth factor type 2 receptor is imprinted and
closely linked to the Tme locus. Nature 349:84-87.1991
[n23.] Leff SE, Brannan CI, Reed ML, et al: Maternal imprinting of the mouse Snrpn gene and conserved linkage ho-
mology with the human Prader-Willi syndrome region. Nat Genet 2:259-264, 1992
[n24.] Cattanach BM, Barr JA, Evans EP, et al: A candidate mouse model for Prader-Willi syndrome which shows an
absence of Snrpn expression. Nat Genet 2:270-274, 1992
[n25.] Giddings SJ, King CD, Harman KW, et al: Allele specific inactivation of insulin 1 and 2 in the mouse yolk sac
indicates imprinting. Nat Genet 6:310-313, 1994

Page
7
HHS, NCI, J Natl Cancer Inst, May 18, 1994
[n26.] Hatada I, Sugama T, Mukai T: A new imprinted gene cloned by a methylation-sensitive genome scanning method.
Nucleic Acids Res 21:5577-5582, 1993
[n27.] Hayashizaki Y, Shibata H, Hirotsune S, et al: Identification of an imprinted U2af binding protein related sequence
on mouse chromosome 11 using RLGS method. Net Genet 6:33-40, 1994
[n28.] Nicholls RD, Knoll JH, Butler MG, et al: Genetic imprinting suggested by maternal heterodisomy in nondeletion
Prader-Willi syndrome. Nature 342:281-285, 1989
[n29.] Butler MG, Meaney FJ, Palmer CG: Clinical and cytogenetic survey of 39 individuals with Prader-Labhart-Willi
syndrome. Am J Med Genet 23:793-809, 1986
[n30.] Brown KW, Williams JC, Maitland NJ, et al: Genomic imprinting and the Beckwith-Wiedemann syndrome. Am J
Hum Genet 46:1000-1001.1990
[n31.] Weksburg R, Teshima I, Williams BR, et al: Molecular characterization of cytogenetic alterations associated with
the Beckwith-Wiedemann syndrome (BWS) phenotype refines the localization and suggests the gene for BWS is im-
printed. Hum MOl Genet 2:549-556, 1993
[n32.] Schroeder WT, Chao L-Y, Dao DD, et al: Nonrandom loss of maternal chromosome II alleles in Wilms tumor. Am
J Hum Genet 40:413-420. 1987
[n33.] Ozcelik T, Leff S, Robinson W, et al: Small nuclear ribonucleoprotein polypeptide N (SNRPN), an expressed gene
in the Prader Willi syndrome critical region. Nat Genet 2:265-269, 1992
[n34.] Reed ML, Leff SE: Maternal imprinting of human SNRPN, a gene deleted in Prader-Willi syndrome. Nat Genet
6:163-167, 1994
[n35.] Engstrom W, Lindham S, Schofield P: Wiedemann-Beckwith syndrome. Eur J Pediatr 147:450-457, 1988
[n36.] Little M, Van Heyningen V, Hastie N: Dads and disomy and disease. Nature 351:609-610, 1991
[n37.] The Huntington's Disease Collaborative Research Group: A novel gene containing a trinucleotide repeat that is
expanded and unstable on Huntington's disease chromosomes. Cell 72:971-983, 1993
[n38.] Fu Y, Kuhl DP, Pizzuti A, et al: Variation of the CGG repeat at the fragile X site results in genetic instability:
resolution of the Sherman paradox. Cell 67:1047-1058, 1991
[n39.] Pieretti M, Zhang FP, Fu YH, et al: Absence of expression of the FMR-1 gene in fragile X syndrome. Cell
66:817-822, 1991
[n40.] Wake N, Fujino T, Hoshi S, et al: The propensity to malignancy of dispermic heterozygous moles. Placenta
8:319-326, 1987
[n41.] Linder D, McCaw B, Kaiser X, et al: Parthenogenetic origin of benign ovarian teratomas. N Engl J Med 292:63-66,
1975
[n42.] Van Gils AP, Van der Mey AG, Hoogma RP, et al: MRI screening of kindred at risk of developing paragangliomas:
support the genomic imprinting in hereditary glomus tumours. Br J Cancer 65:903-907, 1992
[n43.] Costello CC, Decosse JJ: Possible genomic imprinting in familial adenomatous polyposis. Lancet 340:918, 1992
[n44.] Fearon ER, Vogelstein B, Feinberg AP: Somatic deletion and duplication of genes on chromosome 11 in Wilms'
tumours. Nature 309: 176-178. 1984

Page
8
HHS, NCI, J Natl Cancer Inst, May 18, 1994
[n45.] Kinzler KW, Nilbert MC, Vogelstein B, et al: Identification of a gene located at chromosome 5q21 that is mutated
in colorectal cancers. Science 251:1366-1370, 1991
[n46.] Groden J, Thiveris A, Samowitz W, et al: Identification and characterization of the familial adenomatous polyposis
coli gene. Cell 66:589-600, 1991
[n47.] Toguchida J, Ishizaki K, Sasaki M, et al: Preferential mutation of paternally derived RB gene as the initial event in
sporadic osteosarcoma. Nature 338:156-158, 1989
[n48.] Katz F, Webb D, Gibbons B, el al: Possible evidence for genomic imprinting in childhood acute myeloblastic
leukaemia associated with monosomy for chromosome 7. Br J Haematol 80:332-336, 1992
[n49.] Schroeder W, Chao L, Dao D, et al: Nonrandom loss of maternal chromosome 11 alleles in Wilms tumors. Am J
Hum Genet 40:413-420, 1987
[n50.] Haas OA, Argyriou-Tirita A, Lion T: Parental origin of chromosomes involved in the translocation t(9;22). Nature
359:414-416, 1992
[n51.] Brannan CI, Dees EC, Ingram RS, et al: The product of the H19 gene may function as an RNA. Mol Cell Biol
10:28-36, 1990
[n52.] Brunkow ME, Tilghman SM: Ectopic expression of the H19 gene in mice causes prenatal lethality. Genes Dev 5:
1092-1101, 1991
[n53.] Weksberg R, Shun DR, Fei YL, et al: Disruption of insulin-like growth factor 2 imprinting in Beck-
with-Wiedemann syndrome. Nat Genet 5:143-150, 1993
[n54.] Schwab M, Alitalo K, Klempnauer KH, et al: Amplified DNA with limited homology to myc cellular oncogene is
shared by human neuroblastoma cell lines and a neuroblastoma tumour. Nature 305:245-248, 1983
[n55.] Oliner JD, Kinzler KW, Meltzer PS, et al: Amplification of a gene encoding a p53-associated protein in human
sarcomas. Nature 358:80-83, 1992
[n56.] Joos S, Haluska FG, Falk MH, et al: Mapping chromosomal breakpoints of Burkitt's t(8;14) translocations far
upstream of c-myc. Cancer Res 52:6547-6552, 1992
[n57.] Taub R, Kirsch I, Morton C, et al: Translocation of the c-myc gene into the immunoglobulin heavy chain locus in
human Burkitt lymphoma and murine plasmacytoma cells. Proc Natl Acad Sci USA 79:7837-7841, 1982
[n58.] Havemann K, Rotsch M, Schoneberger H, et al: Growth regulation by insulin-like growth factors in lung cancer. J
Steroid Biochem Mol Biol 37:877-882, 1990
[n59.] Reeve JG, Brinkman A, Hughes S, et al: Expression of insulinlike growth factor (IGF) and IGF-binding protein
genes in human lung tumor cell lines. J Natl Cancer Inst 84:628-634, 1992
[n60.] Jaques G, Kiefer P, Rotsch M, et al: Production of insulin-like growth factor binding proteins by small-cell lung
cancer cell lines. Exp Cell Res 184:396-406, 1989
[n61.] Arteaga CL: Interference of the IGF system as a strategy to inhibit breast cancer growth. Breast Cancer Res Treat
22:101-106, 1992
[n62.] Yee D, Cullen KJ, Paik S, et al: Insulin-like growth factor 11 mRNA expression in human breast cancer. Cancer
Res 48:6691-6696, 1988

Page
9
HHS, NCI, J Natl Cancer Inst, May 18, 1994
[n63.] Lambert S, Carlisi A, Collette J, et al: Insulin-like growth factor II in two human colon-carcinoma cell lines: gene
structure and expression, and protein secretion. Int J Cancer 52:404-408, 1992
[n64.] Lahm H, Suardet L, Laurent PL, et al: Growth regulation and co-stimulation of human colorectal cancer cell lines
by insulin-like growth factor I, II and transforming growth factor alpha. Br J Cancer 65:341-346, 1991
[n65.] Lambert S, Collette J, Gillis J, et al: Tumor IGF-II content in a patient with a colon adenocarcinoma correlates with
abnormal expression of the gene. Int J Cancer 48:826-830, 1991
[n66.] Lambert S, Vivario J, Boniver J, et al: Abnormal expression and structural modification of the insulin-like
growth-factor-II gene in human colorectal tumors. Int J Cancer 46:405-410, 1990
[n67.] Yashiro T, Tsushima T, Murakami H, et al: Insulin-like growth factor-11 (IGF-II)/Mannose-6-phosphate receptors
are increased in primary human thyroid neoplasms. Eur J Cancer 27:699-703, 1991
[n68.] Shapiro ET, Bell GI, Polonsky KS, et al: Tumor hypoglycemia: relationship to high molecular weight insulin-like
growth factor-II. J Clin Invest 85:1672-1679.1990
[n69.] Su TS, Liu WY, Han SH, et al: Transcripts of the insulin-like growth factors I and II in human hepatoma. Cancer
Res 49:1773-1777, 1989
[n70.] Glick RP, Unterman TG, Van der Woude M, et al: Insulin and insulin-like growth factors in central nervous system
tumors. Part V: production of insulin-like growth factors I and II in vitro. J Neurosurg 77:445-450, 1992
[n71.] Melino G, Stephanou A, Annicchiarico-Petruzzelli M, et al: IGF-II mRNA expression in LI human glioblastoma
cell line parallels cell growth. Neurosci Lett 144:25-28, 1992
[n72.] Liu JP, Baker J, Perkins AS, et al: Mice carrying null mutations of the genes encoding insulin-like growth factor I
(Igf-I) and type 1 IGF receptor (Igflr). Cell 75:59-72, 1993
[n73.] Minniti CP, Maggi M, Helman LJ: Suramin inhibits the growth of human rhabdomyosarcoma by interrupting the
insulin-like growth factor II autocrine growth loop. Cancer Res 52:1830-1835, 1992
[n74.] Madden SL, Cook DM, Morris JF, et al: Transcriptional repression mediated by the WTI Wilms tumor gene
product. Science 253:1550-1553, 1991
[n75.] Little MH, Prosser J, Condie A, et al: Zinc finger point mutations within the WT1 gene in Wilms tumor patients.
Proc Natl Acad Sci USA 89:4791-4795, 1992
[n76.] Sapienza C: Genome imprinting, cellular mosaicism and carcinogenesis. Mol Carcinog 3:118-121, 1990
[n77.] Feinberg AP: Genomic imprinting and gene activation in cancer. Nat Genet 4:110-113, 1993
[n78.] Hao Y, Crenshaw T, Moulton T, et al: Tumor-suppressor activity of H19 RNA. Nature 365:764-767, 1993
[n79.] Zemel S, Bartolomei MS, Tilghman SM: Physical linkage of two mammalian imprinted genes, H19 and insu-
lin-like growth factor 2. Nat Genet 2:61-65, 1992
[n80.] Bartolomei MS, Webber AL, Brunkow ME, et al: Epigenetic mechanisms underlying the imprinting of the mouse
H19 gene. Genes Dev 7:1663-1673, 1993
[n81.] Razin A, Szyf M: DNA methylation patterns. Formation and function. Biochim Biophys Acta 782:331-342, 1984
[n82.] Wilks AF, Cozens PJ, Mattaj IW, et al: Estrogen induces a demethylation at the 5' end region of the chicken vi-
tellogenin gene. Proc Natl Acad Sci USA 79:4252-4255, 1082

Page
10
HHS, NCI, J Natl Cancer Inst, May 18, 1994
[n83.] Haigh LS, Owens BB, Hellewell S, et al: DNA methylalton in chicken alpha-globin gene expression. Proc Natl
Acad Sci USA 79:5332-5336, 1982
[n84.] Groudine M, Eisenman R, Weintraub H: Chromatin structure of endogenous retroviral genes and activation by an
inhibitor of DNA methylation. Nature 292:311-317.1981
[n85.] Wigler M, Levy D, Perucho M: The somatic replication of DNA methylation. Cell 24:33-40, 1981
[n86.] Stein R, Razin A, Cedar H: In vitro methylation of the hamster adenine phosphoribosyltransferase gene inhibits its
expression in mouse L cells. Proc Natl Acad Sci USA 79:3418-3422, 1982
[n87.] Waechter DE, Baserga R: Effect of methylation on expression of microinjected genes. Proc Natl Acad Sci USA
79:1106-1110, 1982
[n88.] Reik W, Collick A, Norris ML, et al: Genomic imprinting determines methylation of parental alleles in transgenic
mice. Nature 328:248-251. 1987
[n89.] Sapienza C, Peterson AC, Rossant J, et al: Degree of methylation of transgenes is dependent on gamete of origin.
Nature 328:251-254, 1987
[n90.] Sasaki H, Jones PA, Chaillet JR, et al: Parental imprinting: potentially active chromatin of the repressed maternal
allele of the mouse insulin-like growth factor II (Igf2) gene. Genes Dev 6:1843-1856, 1992
[n91.] Kalscheuer VM, Mariman EC, Schepens MT, et al: The insulin-like growth factor type-2 receptor gene is imprinted
in the mouse but not in humans. Nat Genet 5:74-78, 1993
[n92.] Heslop-Harrison JS: Gene expression and parental dominance in hybrid plants. Dev Suppl:21-28, 1990
[n93.] Pfeifer GP, Riggs AD: Chromatin differences between active and inactive X chromosomes revealed by genomic
footprinting of permeabilized cells using DNase I and ligation-mediated PCR. Genes Dev 5:1102-1113, 1991
[n94.] Lock LF, Takagi N, Martin GR: Methylation of the Hprt gene on the inactive X occurs after chromosome inacti-
vation. Cell 48:39-46, 1987
[n95.] Hockey AJ, Adra CN, McBurney MW: Reactivation of hprt on the inactive X chromosome with DNA demethyl-
ating agents. Somat Cell Mol Genet 15:421-434, 1989
[n96.] Holliday R: A new theory of carcinogenesis. Br J Cancer 40:513-522, 1979
[n97.] Riggs AD, Jones PA: 5-Methylcytosine, gene regulation, and cancer. Adv Cancer Res 40:1-311, 1983
[n98.] Feinberg AP, Vogelstein B: Hypomethylation distinguishes genes of some human cancers from their normal
counterparts. Nature 301:89-92, 1983
[n99.] Goelz SE, Vogelstein B, Hamilton SR, et al: Hypomethylation of DNA from benign and malignant human colon
neoplasms. Science 228: 187-190, 1985
[n100.] Feinberg AP, Gehrke CW, Kuo KC, et al: Reduced genomic 5-methylcytosine content in human colonic neo-
plasia. Cancer Res 48:1159-1161, 1988
[n101.] Feinberg AP, Vogelstein B: Hypomethylation of ras oncogenes in primary human cancers. Biochem Biophys Res
Commun 111:47-54, 1983
[n102.] Jablonka E, Lamb MJ: The inheritance of acquired epigenetic variations. J Theor Biol 139:69-83, 1989

Page
11
HHS, NCI, J Natl Cancer Inst, May 18, 1994
[n103.] Li E, Beard C, Jaenisch R: Role for DNA methylation in genomic imprinting. Nature 366:362-365, 1993
[n104.] Tartof KD, Bremer M: Mechanisms for the construction and developmental control of heterochromatin formation
and imprinted chromosome domains. Dev Suppl:35-45, 1990
[n105.] Locke J, Kotarski MA, Tartof KD: Dosage-dependent modifiers of position effect variegation in Drosophila and a
mass action model that explains their effect. Genetics 120:181-198, 1988
[n106.] Saunders WS, Chue C, Goebl M, et al: Molecular cloning of a human homologue of Drosophila heterochromatin
protein HPI using anti-centromere auto antibodies with anti-chromo specificity. J Cell Sci 104:573-582, 1993
[n107.] Kitsbert D, Selig S, Brandeis M, et al: Allele-specific replication timing of imprinted gene regions. Nature
364:459-463, 1993
[n108.] Klar AJS: Regulation of fission yeast mating-type interconversion by chromosome imprinting. Dev Suppl:3-8,
1990
GRAPHIC: Figure 1, Mendelian inheritance of dominant and recessive genes.; Figure 2, Inheritance pattern of a ma-
ternally imprinted gene and LOI in tumorigenesis and Beckwith-Wiedemann syndrome.; Figure 3, Evidence for genomic
imprinting.; Figure 4, Cause and effects of abnormal imprinting.