MOLECULAR AND CELLULAR BIOLOGY, Sept. 2007, p. 6555–6568
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Vol. 27, No. 18
Epigenetic Activation of the Human Growth Hormone Gene Cluster
during Placental Cytotrophoblast Differentiation?
Atsushi P. Kimura,1† Daria Sizova,1Stuart Handwerger,2
Nancy E. Cooke,1and Stephen A. Liebhaber1*
Departments of Genetics and Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104,1and Department of
Endocrinology, Children’s Hospital Medical Center, and Department of Pediatrics, University of Cincinnati College of
Medicine, Cincinnati, Ohio 452292
Received 14 February 2007/Returned for modification 29 March 2007/Accepted 4 July 2007
The hGH cluster contains a single human pituitary growth hormone gene (hGH-N) and four placenta-
specific paralogs. Activation of the cluster in both tissues depends on 5? remote regulatory elements. The
pituitary-specific locus control elements DNase I-hypersensitive site I (HSI) and HSII, located 14.5 kb 5? of the
cluster (position ?14.5), establish a continuous domain of histone acetylation that extends to and activates
hGH-N in the pituitary gland. In contrast, histone modifications in placental chromatin are restricted to the
more 5?-remote HSV-HSIII region (kb ?28 to ?32) and to the placentally expressed genes in the cluster, with
minimal modification between these two regions. These data predict distinct modes of hGH cluster gene
activation in the pituitary and placenta. Here we used cell culture models to track structural changes at the
hGH locus through placental-gene activation. The data revealed that this process was initiated in primary
cytotrophoblasts by histone H3K4 di- and trimethylation and H4 acetylation restricted to HSV and to the
individual placental-gene repeat (PGR) units within the cluster. Later stages of transcriptional induction were
accompanied by enhancement and extension of these modifications and by robust H3 acetylation at HSV, at
HSIII, and throughout the placental-gene regions. These data suggested that elements restricted to HSIII-HSV
regions and each individual PGR might be sufficient for activation of the hCS genes. This model was tested by
comparing hCS transgene expression in the placentas of mouse embryos carrying a full hGH cluster to that in
placentas in which the HSIII-HSV region was directly linked to the individual hCS-A PGR unit. The findings
indicate that the HSIII-HSV region and the PGR units, although targeted for initial chromatin structural
modifications, are insufficient to activate gene expression and that this process is dependent on additional,
as-yet-unidentified chromatin determinants.
Mammalian development is initiated by differentiation of
the primordial blastocyst into the trophectoderm and inner cell
mass (12, 54, 60). The inner cell mass gives rise to all somatic
tissues, whereas the trophectoderm generates extraembryonic
tissues that will constitute the fetal placenta. Early in develop-
ment, the trophoblast stem cells of human extraembryonic
structures diverge into two distinct lineages: villous and extra-
villous cytotrophoblasts (CTBs) (13, 53, 54, 67). The extravil-
lous CTBs differentiate into invasive trophoblasts that migrate
into the uterine wall in the process of embryo implantation.
The villous CTBs proliferate and fuse to form a layer of syn-
cytial cells that extend over the surface of the placental villi
(24, 34, 47, 51). This multinucleate syncytiotrophoblast (STB)
layer constitutes the interface between fetal and maternal cir-
culation, mediating essential gas and nutrient exchange. In
addition to having these exchange functions, CTBs and STBs
synthesize gestational hormones that are secreted directly into
the maternal circulation.
The morphological and functional transition from a CTB to
a multinucleate STB has been the focus of intense study (24,
38, 50, 55, 60). Although expression profiling has identified an
array of proteins associated with this process (5), the molecular
mechanisms that drive trophoblast differentiation and STB
formation remain to be defined (13, 14, 38, 53, 54). The critical
roles of the STB in multiple aspects of human embryonic
development and maternal-fetal communication and symbiosis
make it likely that defects in this process underlie pathological
abnormalities in placentation affecting fetal growth and devel-
A number of gene induction processes are tightly linked to
the formation and function of human STBs. Among the most
robust of these processes is the induction of the human growth
hormone (hGH) gene cluster (Fig. 1A). This cluster, located
on chromosome 17q22-24, contains five genes generated by a
series of segmental duplications in the primate genome (7, 11).
The most 5? gene in the cluster is the hGH-N gene. hGH-N is
a major hormonal determinant of postnatal growth of bone
and soft tissues (48). The remaining four genes in the cluster,
from 5? to 3?, are the human chorionic somatomammotropin-
like (hCS-L), hCS-A, hGH variant (hGH-V), and hCS-B genes
(11, 40). Expression of these four genes is restricted to the STB
layer of placental villi. hCS-L is considered a pseudogene; its
expression is blocked by multiple splice site mutations (45).
hGH-V encodes fetal GH, which is expressed exclusively in
STBs and secreted into maternal serum beginning early in the
second trimester. hGH-V has growth-promoting, lactogenic,
* Corresponding author. Mailing address: 428 Clinical Research
Building, University of Pennsylvania, 415 Curie Boulevard, Philadel-
phia, PA 19104. Phone: (215) 898-7834. Fax: (215) 573-5157. E-mail:
† Present address: Laboratories of Functional Biology, Department of
Biological Sciences, Faculty of Science, Hokkaido University, Sapporo
?Published ahead of print on 16 July 2007.
and other activities similar to those of hGH-N (2, 22, 39); its
robust expression in the second and third trimesters represses
maternal pituitary hGH-N expression, thus replacing much of
the hGH-N in maternal serum (18, 36). The two hCS genes,
hCS-A and hCS-B, are coexpressed and encode identical pla-
cental chorionic somatomammotropins, also known as placen-
tal lactogen. hCS constitutes the most abundant maternal se-
rum protein (23, 61). Thus, the hGH-V, hCS-A, and hCS-B
genes undergo a selective induction during the CTB-to-STB
transition in the developing placenta. Of these three, hCS-A is
by far the most robustly expressed gene in the term placenta
(40). Defining the molecular events that underlie this induction
should identify molecular mechanisms involved in STB forma-
tion and function.
In previous studies, we demonstrated that activation of the
hGH cluster in both the pituitary and the placenta is depen-
dent on a set of remote regulatory elements (Fig. 1A) (8, 31,
58, 59, 62). These elements, located between 14.5 and 32 kb 5?
of the cluster, were identified initially in chromatin prepara-
tions by their sensitivity to DNase I (DNase I-hypersensitive
site I [HSI] to HSV ). HSV and HSIII, at kb ?32 and kb
?28, respectively, are detected in pituitary somatotrope and
placental STB chromatin; HSIV, at kb ?30, is specific to STB
chromatin; and HSI and HSII, at kb ?14.5 to kb ?15.5, are
specific to somatotrope chromatin. In addition to the specific-
ity of HS formation, the activation of the hGH cluster in
pituitary and placenta is distinguished by observed patterns of
histone modifications. Histone acetyltransferase activity re-
cruited to HSI establishes a continuous 32-kb domain of his-
tone acetylation connecting the locus control region (LCR)
and the hGH-N promoter (16, 26). This acetylated domain
facilitates trans-factor binding at the hGH-N promoter and the
transcriptional activation of hGH-N (26, 27). Activation of the
placental genes in term placental STBs is marked by activating
histone modifications that are restricted to the HSV-HSIII
region and to the placental genes (32); the extensive regions
between HSIII and the cluster remain unmodified. These dif-
ferences in HS formation and histone modifications predict
two distinct pathways of long-range control by the LCR, a
“spreading” or “tracking” mechanism in the pituitary (16, 17,
26) and a “looping” mechanism in the placenta (32). Based on
the present knowledge of cellular differentiation and epige-
netic alterations, it seems reasonable to propose that chroma-
tin structures in the placenta are altered during the terminal
transition of CTBs to STBs and the corresponding robust in-
duction of gene expression from the hGH cluster. Analysis of
this process of epigenetic modifications should address molec-
ular mechanisms at critical stages of human placental devel-
opment and further clarify the corresponding long-range in-
teractions. Here we explore this process utilizing ex vivo cell
culture models of STB differentiation. The data support a
FIG. 1. Expression of the hGH-hCS cluster genes in human placental cell lines. (A) Structure of the hGH gene cluster and its LCR. The hGH
cluster is composed of five conserved genes, including the pituitary-specific hGH-N gene and placenta-specific hCS-L, hCS-A, hGH-V, and hCS-B
genes. The expression of the hGH cluster genes is regulated by its LCR, which is far upstream of the cluster and overlaps two other tissue-specific
genes, the B-lymphocyte-specific CD79b gene and the muscle-specific SCN4A gene. The LCR includes five DNase I-HSs, indicated as vertical
arrows. The bent arrows above each gene indicate the transcriptional direction of the gene. (B) Expression of the early placental marker gene
hCG? was detected in BeWo cells by Northern analysis. Total RNAs from CCDsk-25, JEG3, and BeWo cell lines were fractionated in a
formaldehyde-agarose gel and transferred to a nitrocellulose membrane. The blot was hybridized with a32P-labeled hGH-hCS probe that detects
all five genes. The blot was also hybridized with a probe for the early placental hCG? mRNA. The constitutive rpL32 gene was used as a loading
control. The signals were visualized by autoradiography. (C) Expression of the placental hGH-hCS genes is activated in BeWo cells. Total RNAs
were purified and amplified by PCR. These starting PCR products are shown (top panel). The products were next digested with TaqI to distinguish
the mRNAs corresponding to each of the hGH cluster genes (middle panel). Amplification of rpL32 mRNA was used as a positive control (bottom
6556KIMURA ET AL.MOL. CELL. BIOL.
simplified model of placental-gene activation, and that model
is tested by in vivo transgenic analysis.
MATERIALS AND METHODS
Cell culture. The human choriocarcinoma cell lines BeWo and JEG3 and a
human skin fibroblast line, CCDsk-25, were purchased from the American Type
Culture Collection. JEG3 and CCDsk-25 cells were cultured at 37°C with 5%
CO2in minimum essential medium (Invitrogen Life Technologies, Carlsbad,
CA) supplemented with 10% fetal bovine serum (HyClone, Logan, UT), 100
U/ml penicillin, and 0.1 mg/ml streptomycin (Invitrogen Life Technologies).
BeWo cells were maintained at 37°C with 5% CO2in F12 medium (Invitrogen
Life Technologies) supplemented with 15% fetal bovine serum (HyClone), 0.2%
glucose (Sigma-Aldrich, St. Louis, MO), 100 U/ml penicillin, and 0.1 mg/ml
streptomycin (Invitrogen Life Technologies). To induce hGH/CS expression,
BeWo cells were treated with either 80 ?M forskolin (FSK) or 10 to 100 nM
trichostatin A (TSA) and with the carrier dimethyl sulfoxide or ethanol in
parallel controls. For long-term culture, 0.7 ? 106cells were spread on a 60-mm
dish with or without a Matrigel coating (BD Biosciences, Palo Alto, CA). These
were incubated overnight to allow cell attachment. Additional surface treatments
that were assessed included collagen type I, collagen type IV, fibronectin, lami-
nin, and poly(D) lysine (BD Biosciences). The day after plating was designated
day 0, and the medium was subsequently changed every 2 days for the remainder
of the culture period.
Isolation and culture of primary CTBs. Purification of CTBs from human term
placentas and their primary culture was conducted as previously described (5,
33). Third-trimester placentas were obtained from women with normal pregnan-
cies and deliveries, and CTBs were isolated by enzymatic disaggregation and
cultured as described previously (52). The CTBs were purified to ?98% homo-
geneity by negative CD9 selection. The protocol for obtaining placentas was
approved by the Human Investigation Committees of the University of Cincin-
nati and the Children’s Hospital Medical Center. Second-trimester maternal
serum was added to the culture medium to maximize hCS gene induction (52).
RT-PCR analyses. Total RNA was extracted from cultured cells, term placentas,
or mouse tissues using TRIzol reagent (Invitrogen Life Technologies) according
to the manufacturer’s instructions. The RNAs were treated with RNase-free
DNase I (Invitrogen Life Technologies) at 37°C for 5 h and reextracted with
TRIzol reagent. For the reverse transcriptase-PCR (RT-PCR) analysis of hGH
gene cluster expression, the resulting RNAs were reverse transcribed with an
oligo(dT) primer and Superscript III transcriptase (Invitrogen Life Technolo-
gies), and 2.5% of this RT reaction mixture was used for PCR. PCR was
performed with a sense primer32P labeled with T4 polynucleotide kinase (5?-G
TCCCTGCTCCTGGCTTTTG-3?) and an antisense primer (5?-AGCAGCCCG
TAGTTCTTGAG-3?). PCR was carried out for 25 to 35 cycles under the fol-
lowing conditions: 30 s at 94°C, 30 s at 57°C, and 2 min at 72°C. The amplified
cDNA segments were 546 bp for hGH-N, hGH-V, hCS-A, and hCS-B and 492 bp
for hCS-L. Digestion of the amplification products with TaqI generated an
hGH-V fragment of 546 bp, an hGH-N fragment of 494 bp, an hCS-L fragment
of 251 bp, and hCS-A and hCS-B fragments of 305 bp. The amplified and
digested products were separated on 4% polyacrylamide gels, and the signals
were visualized by autoradiography. As an internal control, a transcript of human
ribosomal-protein large-subunit 32 (rpL32) mRNA was amplified using a specific
primer pair (5?-GTGAAGCCCAAGATCGTCA-3? and 5?-TGTTGCACATCA
GCAGCAC-3?) (70). For the RT-PCR analysis of the relative levels of hCS-A
and hCS-B gene expression, a primer pair of 5?-AGAACTACGGGCTGCTC
T-3? and 5?-AGGGCCAGGAGAGGCACT-3? was used. Reverse transcription
cDNA synthesis, with the reaction mixture primed with the antisense oligonu-
cleotide, was followed by PCR amplification utilizing the same antisense primer
together with the sense oligonucleotide labeled at its 5? end with32P (40). PCR
products were then separated in 6% denaturing polyacrylamide gel, visualized by
autoradiography, and subjected to PhosphorImager quantification. Expected
sizes for the hCS-A and hCS-B RT-PCR-amplified fragments were 153 bp and
157 bp, respectively.
Northern blot analysis. Total RNA was purified from cells or tissues using
TRIzol reagent (Invitrogen Life Technologies) according to the manufacturer’s
protocol. The RNA was electrophoresed on a formaldehyde-agarose gel and
transferred to a Zetabind membrane (Cuno Inc., Meriden, CT). The blot was
hybridized for 18 h with a32P-labeled probe at 42°C in 50% formamide, 5?
Denhardt’s solution, 5? SSPE (1? SSPE is 0.18 M NaCl, 10 mM NaH2PO4, and
1 mM EDTA [pH 7.7]), 1% sodium dodecyl sulfate, and 100 ?g/ml salmon sperm
DNA. The probes for hGH/hCS and rpL32 mRNAs were prepared by RT-PCR
as described above. The probes for human chorionic gonadotropin ? (hCG?) and
mouse placental lactogen II were also prepared by RT-PCR using specific primer
pairs (5?-CATCACCGTCAACACCACCAT-3? and 5?-TCACAGGTCAAGGG
GTGGTC-3? and 5?-ACTCCTCAGAGATGAAGCTG-3? and 5?-ACATCACG
ACACTTCAGGAC-3?, respectively). All the probes were labeled with the Kle-
now fragment by using a random-priming labeling kit (Roche Diagnostics,
Mannheim, Germany). The oligoprobe for 18S rRNA (5?-CGGCATGTATTA
GCTCTAGAATTACCACAG-3?) (21) was32P labeled with T4 polynucleotide
kinase. After hybridization, the membrane was washed at 50°C in 0.1% sodium
dodecyl sulfate–0.1? SSC (1? SSC is 0.15 M NaCl plus 0.015 M sodium citrate)
and exposed to Kodak Biomax film (Eastman Kodak, Rochester, NY).
HS region mapping. Nuclei from cultured cells and from human term
placentas were isolated and stored as described previously (32). HS regions
were detected as described previously (17, 26, 31). Briefly, 300 to 500 ?g of
nuclei were digested with 45 units of RNase-free DNase I (Invitrogen Life
Technologies) at 37°C in 1 ml of 50 mM Tris-HCl (pH 8.0), 0.1 M NaCl, 3 mM
MgCl2, 1 mM CaCl2, and 1 mM phenylmethylsulfonyl fluoride. At the time
points of 0, 2, 4, 8, and 15 min, 200-?l samples were transferred into new
tubes containing 20 ?l of 0.5 M EDTA. After the genomic DNA was purified,
each sample was digested with EcoRI and the signals were detected by
Southern blot hybridization.
ChIP assay of unfixed chromatin. A chromatin immunoprecipitation (ChIP)
assay with unfixed chromatin to determine the histone modification patterns was
performed as previously described (32). In brief, 300 to 600 ?g of nuclei was
digested with micrococcal nuclease and soluble chromatin was collected. Immu-
noprecipitation was performed with antibodies for modified histones, and DNA
was purified from the precipitated (bound) chromatin as well as from the fraction
before the precipitation (input). Antibodies used were anti-acetyl histone H3,
anti-acetyl histone H4, anti-dimethylated-histone H3 (Lys4) (Upstate Bio-
technology, Inc., Lake Placid, NY), and anti-trimethylated-histone H3 (Lys4)
(Abcam, Cambridge, United Kingdom). PCR was conducted with the serially
diluted DNA samples from the bound and input fractions. The histone modifi-
cation levels were calculated as ratios of the signal intensity of the bound fraction
to that of the input fraction. Those ratios were further normalized to the corre-
sponding ratio for the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) pro-
moter, considered 100. The amplimer of ?-globin was assessed to determine the
histone modification levels of an inactive gene. All the primers used in this study
were described previously (32), except for the ?-globin primers, whose sequences
were 5?-ATGTTCCTGTCCTTCCCCACC-3? and 5?-ATGGTGCTGTCTCCTG
CCGA-3?. It should also be noted that the background in the day 4 CTB cultures,
as monitored by modification at the ?-globin control locus, was unusually and
consistently high. We therefore further normalized the data set for this study to
the day 4/day 0 modification ratio at the ?-globin locus. For assessing the
modification level of each gene in the cluster, we coamplified the 5? region
(amplimer set 233GH/CS) and digested the products with four restriction en-
zymes (HinfI, DraIII, PstI, and MscI) as described previously (32).
Transgenic constructs and generation of transgenic mouse lines. A human
bacterial artificial chromosome (BAC) genomic clone encompassing the hGH
LCR and the entire hGH cluster was identified by a screen of a total genomic
library (CITB Human B&C Library; Invitrogen Life Technologies) with probes
corresponding to the TCAM1 and SCN4A genes (69). A 148.3-kb clone (clone
535D15) was identified, and a 123-kb insert was released by NotI digestion. To
construct the HSIII-V/Pa(CSA)Ea transgene, the 11.5-kb ClaI-EcoRI fragment
of cosmid clone GH5 (31) encompassing the placental LCR HSIII-HSV region
was ligated to a 7.5-kb fragment containing the hCS-A locus, including its con-
tiguous 5? P element and contiguous 3? putative enhancer. This hCS-A fragment
was generated by PCR using a specific primer set (5?-GGTACCTGGTCCATG
GTTGGCATGGTAACCCCTTAC-3? and 5?-GTCGACGCGGCCGCACCAC
AACTGCCATCTCCTTTTTCTCC-3?). The long-distance PCR was performed
with 50 ?l that included 100 to 200 ng of the BAC DNA template, 1? LA PCR
buffer (Takara Bio, Ohtsu, Japan), a 0.4 mM concentration of a deoxynucleoside
triphosphate mixture, a 0.2 ?M concentration of paired primers, 2.5 units of LA
Taq (Takara Bio), and 1.3 units of Pfu Ultra DNA polymerase (Stratagene, La
Jolla, CA) under the following conditions: 1 min at 94°C, 35 cycles of 20 s at 98°C
and 10 min at 66°C, and a final extension for 10 min at 72°C.
The hGH/BAC and HSIII-HSV/Pa(CSA)Ea DNAs were adjusted to a concentra-
male pronuclei of fertilized mouse eggs. All procedures involved in generating the
two sets of transgenic mouse lines were carried out by the University of Pennsylvania
Transgenic and Chimeric Mouse Core Facility (http://www.med.upenn.edu/genetics
/core-facs/tcmf/index.html). Positive founders were identified by dot blot analysis of
tail DNA using the HSIII probe (32).
VOL. 27, 2007PLACENTAL ACTIVATION OF THE hGH CLUSTER 6557
Activation of the hGH cluster in BeWo cells. A number of
human-placenta-derived cell lines have been used to model
CTB function (63). Two of the most intensively studied are the
choriocarcinoma lines JEG3 and BeWo. These two lines were
studied with the aim of defining the structure of the hGH
cluster prior to the CTB-to-STB transition. A human skin
fibroblast line, CCDsk-25, was used as a nonexpressing control.
Expression profiles of the JEG3, BeWo, and CCDsk-25 cell
lines were initially assessed by Northern blotting. hCG?
mRNA served as a marker of CTB gene expression. The hCG?
gene is not linked to the hGH cluster, and its expression pre-
cedes activation of the hCS genes during gestation and during
the CTB-to-STB differentiation (40). A map of the hGH-hCS
multigene cluster, closely linked genes, and the hGH LCR
region is shown (Fig. 1A). A single hybridization probe that
recognizes all five of the evolutionarily related hGH-hCS
mRNAs was used to monitor total output from the hGH clus-
ter (hGH/CS mRNA). The Northern blot revealed an hCG?
mRNA signal in the BeWo cells, with no evidence of coexisting
hCS expression (Fig. 1B). The same three mRNA samples
were reassessed at higher levels of sensitivity by RT-PCR (35
cycles) with primers that coamplify all five of the hGH and hCS
mRNAs (Fig. 1C, top panel). hGH/hCS mRNAs were detected
in BeWo cells and at lower levels in JEG3 cells. These tran-
scripts were absent in the CCDsk-25 fibroblasts. The output
from the hGH cluster was further defined using a validated
RT-PCR/endonuclease cleavage assay that distinguishes the
individual transcripts (see Materials and Methods). Expression
in JEG3 cells was limited to hCS-L. In contrast, BeWo cells
expressed hGH-V, hCS-A, hCS-B, and hCS-L but no hGH-N
mRNA. The expression of hCS mRNA in the BeWo cells,
which could be detected only by RT-PCR at high cycle num-
bers, constituted less than 0.01% of that in STBs (see below).
These trace levels of hCS and hGH-V mRNAs in the BeWo
cells might represent expression from a small number of the
cells that have spontaneously initiated STB differentiation or
may represent trace expression from the BeWo population as
a whole. In either case, analysis of hGH cluster chromatin in
the BeWo cell model should reveal the epigenetic structure at
the hGH-hCS locus prior to the robust activation characteristic
of STB differentiation.
The “competence” of the hGH cluster to be induced in the
BeWo cells was next assessed. Such an induction would attest
to the “poised” state of the cluster and possibly allow us to
follow the chromatin transition from CTB to STB. Attempts to
induce expression in the BeWo cells by treatment with the
histone deacetylase inhibitor TSA or with the adenylyl cyclase
activator FSK (see Materials and Methods) (65, 68) were in-
effective (Fig. 2A and B and data not shown). However, in
agreement with prior studies (68), FSK triggered a striking
enhancement of hCG? expression (compare Fig. 2A and B). A
third approach was based on the model that the transition from
CTB to STB is enhanced by cell-cell interactions. The BeWo
cells were grown to high density over a 32-day period to allow
the cells to become tightly packed and multiply layered. As
assessed by Northern blotting, the hCG? mRNA was rapidly
induced early in the study and reached a robust plateau be-
tween days 4 and 8 (Fig. 2C). This hCG? mRNA induction was
followed by a steady increase in expression from the hGH
cluster (hGH/CS mRNA) (Fig. 2C). Beyond 32 days, the high
cell density resulted in dislodgment from the plastic surface.
Analysis of the expression profile by the RT-PCR/endonucle-
ase cleavage assay confirmed that the activation of the hGH
cluster in the BeWo cells paralleled that seen in native STBs,
with selective induction of the hCS-A, hCS-B, and hGH-V
mRNAs (Fig. 2D). Consistent with this native pattern of in-
duction, hGH-N and hCS-L mRNAs were not detected. To
further optimize this approach, the long-term culture was re-
peated using a variety of cell adhesion substrates (see Materi-
als and Methods). The only significant improvement was ob-
tained with Matrigel. The cells grown on this matrix activated
the hGH cluster more rapidly and robustly than those grown
on noncoated plates (Fig. 2E). Importantly, the placenta-spe-
cific profile of mRNA accumulation was the same as on the
noncoated dishes; i.e., the induction of hGH-V, hCS-A, and
hCS-B mRNAs was selective (data not shown). These studies
demonstrated that hGH cluster gene expression can be in-
duced by growing BeWo cells to high density.
Activation of the hGH cluster in primary CTBs. A second
model for analysis of the CTB-to-STB transition is ex vivo
differentiation of primary placental CTBs. CTBs can be iso-
lated in a highly enriched state from normal human term
placentas (33). These can be induced to differentiate into STBs
by culturing them over a 4- to 6-day period in media supple-
mented with second-trimester maternal serum (5, 52). North-
ern blot analysis of freshly prepared primary CTBs revealed
the presence of hCG? mRNA in the absence of mRNAs orig-
inating from the hGH cluster (Fig. 3A, day 0 lane). The pri-
mary CTBs were then incubated in media containing second-
trimester maternal serum (52), and the cells were assessed on
a daily basis for gene expression. hCG? mRNA levels re-
mained fairly stable over the first 4 days of culture and then
demonstrated a low-level induction on days 5 and 6. In con-
trast, there was a steady induction of expression from the hGH
cluster throughout the 6-day culture period. This induction is
consistent with the results of prior reports using this same
protocol (5, 52). While days 0 and 1 were negative for hGH-
hCS mRNAs by Northern analysis and day 2 showed only trace
levels, the more sensitive RT-PCR/endonuclease assay of day 0
revealed trace levels of hCS mRNA. Furthermore, day 2
mRNA samples revealed trace levels of hCS and hGH-V
mRNAs in the absence of hGH-N and hCS-L (Fig. 3B). The
trace expression of these placental mRNAs from the hGH
cluster in the day 0 CTBs could reflect low-level contamination
of the CTB preparation with STBs or trace expression of the
cluster in the CTB population as a whole.
The preceding studies demonstrated that the hGH cluster is
minimally active in both BeWo cells and freshly harvested
CTBs. In both models, the cluster could be induced in culture
and the profile of mRNA expression from the induced cluster
corresponded to that of primary STBs. As a final assessment of
these two models, we compared the maximal levels of hCS
mRNA in the induced BeWo cells (at day 32 of cell culture)
and CTBs (at day 6 of culture) to that in a preparation of
primary term placental villi (Fig. 3C). In this comparison, the
level of expression from the hGH cluster in both models was
significantly below that in the term placenta. When normalized
to a ribosomal-protein mRNA (rpL32), the levels of hCS
6558 KIMURA ET AL.MOL. CELL. BIOL.
mRNA in day 32 BeWo cells and day 6 CTBs were 0.04% and
3.3% of those in primary placental villous tissue. Although this
may be an underestimation because the comparison is based
on steady-state mRNA levels and the hCS genes were in the
process of induction, it is clear that the numbers of cells gen-
erated in both induction models fell far below that seen in the
native STB population. This may limit the utility of these cell
culture models, once cells are induced, to accurately reflect the
final STB chromatin state. However, defining the structure of
the hGH chromatin locus in BeWo cells and CTBs in their
“poised” state, prior to induction (day 0 BeWo cells and day 0
CTBs), and comparing the epigenetic profiles with those of the
fully induced cluster in the primary STBs should establish the
endpoints of the CTB-to-STB differentiation process.
DNase I-HSs that mark the hGH LCR in STB chromatin are
present in uninduced BeWo cells and in primary CTBs. Of the
five DNase I-HSs that mark proven and/or potential elements
of the hGH LCR, HSIII and HSV are constitutive, HSI and
HSII assemble selectively in pituitary somatotrope cells, and
HSIV is STB specific. To determine whether the chromatin
structures characteristic of STB chromatin are already estab-
lished at the CTB stage of differentiation, nuclear chromatin
samples from BeWo cells and primary CTBs were DNase I
mapped for HSIII, HSIV, and HSV (Fig. 4). Parallel HS map-
ping studies were also carried out on day 32 BeWo cells to
determine whether the culture-based induction altered this
conformation. Analysis of day 6 CTBs could not be performed
because the amount of material after 6 days in culture was
inadequate for reliable DNase I/Southern blot analysis. These
DNase I mapping studies revealed that HSIII, HSIV, and HSV
were already present in BeWo cell and in day 0 CTB chroma-
tin. Of particular note, placenta-specific HSIV was present in
BeWo chromatin at both days 0 and 32, in CTB chromatin at
day 0, and in primary STB but not fibroblast chromatin. Con-
stitutive HSIII was present in all samples. HSV could not be
well visualized in this study and was not further pursued (17,
31). These results indicated that the placental (STB) chromatin
configuration, as reflected by the presence of HSIV, is estab-
lished during the CTB stage prior to transcriptional induction.
Histone modification patterns at the hGH locus in BeWo
cells before and after transcriptional activation. The forma-
tion of placenta-specific HSIV in uninduced BeWo cells and
freshly isolated CTBs suggests that chromatin structures at the
hGH cluster in these cells have been “primed” for subsequent
FIG. 2. Induction kinetics of the placental hGH-hCS genes in BeWo cells. (A) Histone deacetylase inhibition with TSA failed to induce BeWo
cell gene expression. The cells were treated with the indicated concentrations of TSA for 12 h. Total RNA was analyzed by Northern blotting using
ethanol as a vehicle control. Hybridization probes used are indicated to the right of each panel. (B) FSK-induced BeWo cell syncytialization
resulted in only a moderate increase in hGH-hCS expression. BeWo cells were induced to syncytialize by culturing them in the presence of 80 ?M
FSK for 4 days. Total RNA from the treated cells was used for Northern blot analyses. The syncytialization was confirmed by hCG? induction.
Dimethyl sulfoxide (DMSO) was the vehicle control. (C) Placental hGH-hCS expression was induced by long-term culture of BeWo cells. The cells
were plated at a high density and incubated without chemical additives for 32 days. Total RNA (20 ?g) was collected every 4 days for analysis by
Northern blotting. (D) The BeWo hGH-hCS transcript profile after 32 day of culture was qualitatively similar to that in term placentas. RT-PCR
analysis was performed as described for Fig. 1C but with total RNA (5 ?g) from BeWo cells at days 0 and 32 of culture and from human term
placental villi. hCS-A, hCS-B, and hGH-V mRNAs were detected in all three samples; no hGH-N or hCS-L signals were observed. (E) Matrigel
coating of the culture plates enhanced hGH-hCS activation in BeWo cells. The BeWo cells were cultured on Matrigel-coated dishes for 32 days,
and total RNA was collected every 8 days, as described for panel C above. The Northern blot analysis showed higher and faster induction of
hGH-hCS genes than that shown in panel C when cells were grown on Matrigel.
VOL. 27, 2007PLACENTAL ACTIVATION OF THE hGH CLUSTER6559
induction (Fig. 4). The structure of the hGH locus at the
preinduction stage was further assessed by ChIP analysis (16,
32). Four modifications of core histones H3 and H4 were
chosen for study based on their consistent linkage to gene
activation: histone H3 and H4 acetylation and histone H3K4
di- and trimethylation (41, 49). H3K4 methylation has been
linked to gene activation in a number of systems, and the
addition of two rather than three methyl groups at the critical
lysine may play distinct roles in activation pathways (35, 43,
56). Chromatin containing each of these modifications was
enriched by immunoprecipitation (see Materials and Meth-
ods), and DNA sequences in these samples were assayed for
enrichment over the starting material (“input”) at 14 sites
along the cluster. These sites span 98 kb and encompass the
entire hGH LCR and hGH cluster, along with flanking regions
(Fig. 5A). The analysis takes into account the fact that the five
genes within the cluster were generated via segmental dupli-
cations of an ancestral hGH-hCS precursor along with its
flanking sequences (7, 11). Due to this evolutionary history,
several amplimer sets detect three or four copies of the re-
peated regions within the cluster. A core placental-gene repeat
(PGR) unit, encompassing each placentally expressed gene
along with 3 kb of 5?-end-flanking regions, was evaluated at
high resolution with a set of four additional amplimers (Pe0,
Pe2, Pe6, and Pe9). This PGR includes the “P-element,” which
has been postulated to contribute to the specificity of placental
expression for these genes (17, 46). An element that enhances
hCS expression in cell transfection studies (“Enhancer”) is
located 3? of the three hCS genes (but not hGH-V) and is
separately detected by a specific amplimer (En) (66).
The day 0 and day 32 BeWo cells were studied for enrich-
ment of H3K4 methylation across the hGH locus (Fig. 5B and
C). The controls for both the H3K4-me2 and H3K4-me3 ChIP
studies revealed a 20-fold difference between the expressed
GAPDH locus and the repressed ?-globin locus. The H3K4-
me2 analysis in the BeWo day 0 cells revealed a well-defined
FIG. 3. hGH-hCS genes were induced in ex vivo cultures of primary human placental CTBs. (A) Expression of the hGH cluster genes in
primary placental CTB cell culture. The CTBs were cultured in the presence of second-trimester maternal serum for 6 days to induce
spontaneous differentiation into syncytial cells capable of expressing the hGH-hCS genes. Total RNA was prepared from daily aliquots and
analyzed by Northern blotting. Probes used are indicated to the right of the autoradiographs. (B) hCS-A, hCS-B, and hGH-V are expressed
by the induced CTBs. RT-PCR analysis was conducted with RNA from CTBs at days 0 and 2 of culture. The PCR products were digested
with TaqI and analyzed as described for Fig. 1C. The hCS-A and hCS-B mRNAs were detected at day 0, and their expression and that of
hGH-V were induced by day 2. (C) Quantitative comparison of hGH-hCS expression levels among the placental model systems used in this
study. Northern blot analysis was performed on total RNA from human term placenta, cultured primary CTBs at day 6 (CTB 6 d), and day
32 BeWo cells (BeWo 32 d). Relative hGH-hCS mRNA levels, estimated using the rpL32 signals as a loading control, are shown at the
FIG. 4. DNase I-HS mapping in the BeWo, CTB, and STB chro-
matins. Nuclei from the indicated cells were digested with DNase I,
genomic DNA was purified from each sample, and the DNAs were
digested with EcoRI and subjected to Southern blot hybridization
analysis. The positions corresponding to HSIII, HSIV, and HSV are
indicated by arrows at the left. An open arrowhead points to a
nonspecific band that was detected before the DNase I treatment
(time zero). At the bottom is a diagram of the EcoRI fragment
showing the probe position and the sizes of each fragment after
DNase I digestion. The results reveal the presence of HSIV and the
placenta-specific HS in day 0 and 32 BeWo cells as well as in
primary CTBs and STBs. HSIV is absent from the parallel analysis
of the human fibroblast line CCDsk-25, which does not express any
of the hGH cluster genes. A diagram of the relative fragment sizes
is at the bottom.
6560 KIMURA ET AL.MOL. CELL. BIOL.
island of H3K4-me2 modification that extended from the pla-
cental-gene promoters through the hCS enhancer element
(amplimers Pe9 through 3?CS1). The hGH-N promoter and its
flanking regions remained unmodified. This pattern of H3K4-
me2 modification was stable during the induction period in
that the same pattern was observed in the day 32 BeWo cells.
H3K4 trimethylation in the day 0 BeWo cells showed minimal
variation throughout the region, with the exception of a slight
relative enrichment at the 5? boundary of the LCR (HSV) and
over the level in structural genes. As in the case of the H3K4-
me2 analysis, the pattern and level of H3K4-me3 remained
stable over the induction period (day 32).
The BeWo cells were next analyzed for histone H3 and H4
acetylation. In the uninduced, day 0 BeWo cells, the levels of
H3 acetylation across the entire hGH locus were quite low,
with marginal elevations at HSV and in the immediate prox-
imity of the structural genes (Fig. 5D, gray bars). ChIP analysis
of H4 acetylation of the BeWo day 0 chromatin revealed
slightly higher levels but a similar flat profile (Fig. 5E, gray
bars). Analysis of the induced day 32 BeWo cells revealed a
robust enrichment in H3 and H4 acetylation across the entire
region (Fig. 5D and E, black bars). These levels of H3 and H4
acetylation equaled or exceeded those of the GAPDH-positive
control. The enrichment for H3 acetylation was highest at the
5? boundary of the LCR (HSV), with a progressive decrease at
HSIV and HSIII. Otherwise, the levels across the locus dif-
fered by less than twofold. The H4 acetylation pattern similarly
showed a robust increase of acetylation across the entire region
compared to that in the day 0 samples with evidence for se-
lective enrichment at HSV compared to that at HSIV and
In summary, at day 0 of BeWo culture, we observed a well-
FIG. 5. Histone modifications at the hGH cluster and its LCR in BeWo cells during hGH-hCS activation. (A) Positions of amplimers used in
the ChIP assays. The structure of the hGH cluster, linked genes, and the hGH LCR are depicted. P-elements and the putative hCS enhancers are
indicated (shaded and open ovals, respectively). Shaded and striped rectangles indicate extensive conserved segments among the PGR units and
the hGH-N gene. The amplimer names and their positions are shown below the map. The PGR unit is also shown in an expanded format. (B) The
histone H3K4 dimethylation pattern in BeWo cells was established prior to gene activation and did not change after long-term culture. Soluble
chromatin was collected from BeWo cells at days 0 and 32 of culture by digesting nuclei with micrococcal nuclease. The ChIP analysis was
conducted with anti-dimethylated-histone H3K4 antibody, and the resulting input and bound fractions were subjected to PCR amplification. The
bound/input signal ratio at each amplicon (A) was normalized to that at the GAPDH promoter, which was considered 100. The relative histone
modification levels in day 0 (shaded bars) and day 32 (black rectangles) BeWo cells are shown. Standard error values are indicated. The ?-globin
promoter was used as a negative control. (C) Histone H3K4 trimethylation was detected at low levels across the entire locus in BeWo cells at both
day 0 and day 32. Histone H3K4 trimethylation patterns, examined by ChIP analysis with anti-trimethylated-H3K4 antibody, were analyzed and
plotted as described for panel B. (D) Histone H3 acetylation was significantly induced across the locus during the 32 days of BeWo cell culture.
Histone H3 acetylation patterns, using anti-acetylated-histone H3 antibody, were studied as described for panel B. (E) Histone H4 acetylation was
significantly induced across the locus during BeWo cell culture. The ChIP assay was performed, analyzed, and plotted as described for panel B but
with anti-acetyl histone H4 antibody.
VOL. 27, 2007PLACENTAL ACTIVATION OF THE hGH CLUSTER6561
defined island of H3K4-me2 modification at the PGR units
that extended from the placental-gene promoters to their 3?
enhancer regions. During the subsequent induction period,
this segment of H3K4-me2 modification remained unaltered,
while there was a robust increase in H3 and H4 acetylation
across the entire locus. These data suggested that H3K4 di-
methylation and H3/H4 acetylation play distinct roles in the
process of placental-gene activation within the hGH cluster,
with methylation corresponding to the “primed” state and
acetylation tracking with the subsequent transcriptional induc-
Histone modification patterns in primary CTBs. We next
applied ChIP analysis to primary CTBs. Chromatin was iso-
lated from freshly prepared CTB nuclei (day 0 CTBs) or CTB
nuclei following 4 days in STB differentiation media (day 4
CTBs). Day 4 cells rather than day 6 cells were used for ChIP
analysis due to the more reliable and consistent harvest of
chromatin at this time point. The results of these assays are
displayed along with the previously reported profiles for pri-
mary STB chromatin (open bars) (32) to facilitate a direct
comparison (Fig. 6). Levels of H3K4-me2 are exceedingly low
in the primary CTBs; the most prominent modification was
noted at HSV, with more-moderate levels of modification im-
mediately 5? and 3? of the hGH/CS genes (Pe9 and Enhancer,
respectively) (Fig. 6A). In the transition to day 4, the levels of
H3K4-me2 modification at HSV show a sharp decrease, while
FIG. 6. Histone modification at the hGH cluster and its LCR in primary CTBs before and after 4 days of differentiation. (A) Histone H3K4
dimethylation was established after day 4 of culture in primary CTBs. The amplimer positions are the same as those in Fig. 5A. Primary CTBs were
isolated from human term placentas and cultured for 4 days in the presence of second-trimester maternal serum. The ChIP assay was performed
with anti-dimethylated-histone H3K4 antibody as described for Fig. 5B on chromatin from freshly prepared CTBs (day 0) and after culture (day
4). For comparison, corresponding values for modifications of chromatin isolated from primary placental STBs, previously reported by us (32), are
shown. (B) The histone H3K4 trimethylation pattern was largely preset in the freshly prepared, day 0 CTBs. The histone H3K4 trimethylation
pattern was examined by ChIP with anti-trimethylated-H3K4 antibody, analyzed, and labeled. (C) Histone H3 acetylation was induced at the hGH
locus during differentiation from CTBs to STBs. The histone H3 acetylation patterns were determined by ChIP with anti-acetylated-histone H3
antibody. (D) Histone H4 acetylation was induced at the hGH locus during differentiation of the CTBs. The histone H4 acetylation levels were
investigated by ChIP with anti-acetylated-histone H4 antibody. (E) All genes in the hGH cluster were similarly modified in the day 4 CTBs. Histone
modification levels at each cluster gene were examined by ChIP. The PCR analysis of the isolated DNA was performed with an amplimer set
common to the five genes; the amplified products were digested with four restriction enzymes to distinguish each signal.
6562 KIMURA ET AL.MOL. CELL. BIOL.
modification at the PGR units extends to encompass a contin-
uous domain extending from amplimers Pe6 to 3?CS1. Primary
STBs show a further decrease in HSV to background levels,
while modification within the PGR continues to demonstrate
an overall increase within the PGR (except at the 3? enhancer).
The analysis of H3K4-me3 in day 0 CTB chromatin revealed a
discrete and prominent peak at HSV and enrichment over that
in the PGR unit (Fig. 6B). In contrast to the H3K4-me2 data,
the substantial modifications at HSV are fully sustained in day
4 CTBs and STBs. The only further change in H3K4-me3
modification that appeared to take place during the CTB-to-
STB differentiation process was a twofold increase in modifi-
cation 3? of the hCS enhancer (amplimer 3?CS1) and 3? of the
two GH genes (amplimer 3?GH). Thus, H3K4 di- and tri-
methylation was limited to HSV and the PGR units and is
established early in the differentiation process.
Assessment of H3 acetylation by ChIP revealed very low
levels of modification across the hGH locus in the day 0 CTBs,
with no evidence of enrichment at any position (Fig. 6C,
shaded bars). The 4-day induction period boosted levels of the
H3 acetylation modification by two- to threefold at HSV and
HSIII as well as over that of the promoters in the PGR dupli-
cation unit and extending into the 3?-end-flanking enhancer
region (Fig. 6C, black bars). The increase in H3 acetylation
seen in CTBs from day 0 to day 4 was further exaggerated and
broadened in the fully induced STB chromatin locus (open
bars). The increase in H3 acetylation became more promi-
nently marked over that of HSV and HSIII. Thus, the levels of
H3 acetylation in the regions encompassing the placental genes
were dramatically increased and spread to flanking sequences.
H4 acetylation analysis (Fig. 6D) revealed that HSV was prom-
inently modified in day 0 cells and that this level of modifica-
tion remained stable in day 4 cells and in the STBs. Modifica-
tion in day 0 cells at the hCS promoter (Pe9) remained
constant in the day 4 CTBs and STBs, while levels of acetyla-
tion on both sides of this site increased incrementally in the day
4 and STB chromatin. As was the case with H3 acetylation, H4
acetylation in the STBs extended to encompass the entire clus-
ter, with the specific exclusion of the hGH-N promoter.
Modifications of each of the individual genes in the cluster
were analyzed in the day 4 CTBs using a previously described
coamplification/restriction endonuclease assay that is specifi-
cally designed to distinguish the 5? termini of the five genes
(32; see also Materials and Methods). (Modifications in this
region were insufficient for accurate analysis of the day 0
CTBs.) These modifications in the day 4 CTB samples, sum-
marized in Fig. 6E, revealed that all genes in the cluster were
equally modified within a twofold range of each other. These
relatively equivalent levels of modification among the five
paralogs agree with a similar analysis carried out with STBs
(32) and highlight the lack of a direct correlation between
modification at the individual paralogs in the cluster and their
corresponding steady-state mRNA levels.
In summary, the ChIP studies revealed distinct distributions
and dynamics of H3K4 methylation and H3/H4 acetylation
during the CTB-to-STB transition. These modifications are
summarized in Fig. 7. The initial H3K4 di- and trimethylation
modifications in the day 0 and 4 CTBs were localized to the
HSIII-HSV region and to the PGR units. The levels of H3K4-
me2 increased during the STB transition but maintained the
restricted pattern. The hGH-N gene appeared to be transiently
modified at day 4 (Fig. 6E), and this modification was subse-
quently lost in the STBs, while the modifications over the PGR
units were strengthened and extended. There appeared to be a
complete absence of H3 acetylation at the beginning of the
process (day 0 CTBs), and H4 acetylation at HSV and within
the cluster was marginal. Remarkably, the levels of both H3
and H4 acetylation underwent a dramatic increase in parallel
with the induction of hCS/hGH-V transcription in the day 4
CTB and primary STB populations. At the end of this differ-
FIG. 7. Summary of epigenetic modifications during the transition
of human placental CTBs to STBs. (Top) The histone modification
patterns in freshly prepared primary CTBs are diagrammed below the
map. Prior to terminal STB differentiation, all the placental LCR HSs
were already formed and the placental genes were expressed at trace
levels, whereas hGH-N was totally inactive. Moderate levels of histone
H3K4 dimethylation were observed at discrete regions, including HSV,
the sites of the promoters of the placental genes, and the sites of the
putative enhancers. Histone H3K4 trimethylation was also established
at this point within the PGR regions. In contrast, only minimum levels
of H3 acetylation were observed across the locus, and H4 acetylation
was limited to HSV and at the placental-gene promoters. (Middle)
After 4 days of culture, most of the CTBs fused to form a syncytium
and the expression of the placental genes, especially hCS-A and hCS-B,
was dramatically enhanced. Histone H3K4 di- and trimethylation pat-
terns were well established, and the histone H3/H4 acetylation levels
increased modestly at the LCR and at the cluster region in a block
encompassing the four placental genes. (Bottom) In the full-term
placental STBs, histone H3/H4 acetylation increased to maximum lev-
els and the epigenetic patterns necessary for full activation of the
placental genes were completed. The four placental genes are ex-
pressed at the highest levels, whereas hGH-N, lacking the epigenetic
modifications, remains inactive in STBs.
VOL. 27, 2007 PLACENTAL ACTIVATION OF THE hGH CLUSTER6563
entiation process, the 5? remote HS regions and the entire
placentally expressed portion of the hGH cluster were encom-
passed within a continuous domain of H3/H4 acetylation. Of
note, the region between HSIII and the cluster remained en-
tirely excluded from these histone modifications.
Testing of the HSIII-HSV region and an individual PGR
unit for sufficiency of gene activation. The initial modifications
at the hGH locus appeared to be limited to the LCR determi-
nants and to the individual placental-gene units (see above;
summarized in Fig. 7). These data suggest that the modifica-
tions of the chromatin at each PGR unit, in conjunction with
the remote 5? HS elements, may suffice at the critical first step
in the gene activation pathway. Whether these modifications
are sufficient to trigger subsequent events involved in robust
transcriptional activation is not clear. To test this possibility,
we assembled a minimal expression unit in which an 11.5-kb
genomic fragment encompassing HSIII to HSV was linked to
a 7.5-kb genomic fragment encompassing hCS-A and its con-
tiguous 5? P-element and 3? enhancer flanking sequences (Fig.
8A). Expression from this HSIII-V/Pa(CSA)Ea transgene was
compared to that of the 123-kb BAC transgene encompassing
the entire hGH locus and contiguous hGH LCR (hGH/BAC).
Four mouse lines carrying each of these two transgenes were
established. Transgenic males from each line were crossed with
wild-type females. In the case of the 1254D line, the transgene
had inserted into the X chromosome as determined by the
inheritance pattern. For this reason, it was necessary to cross a
transgenic female with a wild-type male in order to overcome
the paternal imprinting that occurs in extraembryonic tissues in
mice. Individual 18.5-day transgenic embryos generated by
these crosses were identified, and hCS-A gene expression was
assessed per transgene copy in the corresponding placentas
(Fig. 8B and C). As is the case for expression from the hGH
cluster in the human term placenta (40), the expression from
hGH/BAC in the transgenic mouse placenta was predomi-
nantly hCS-A mRNA (Fig. 8C). The hCS-A expression in all
four hGH/BAC lines was robust and copy number depen-
dent, indicating strong site-of-integration independence. In
contrast, expression from the HSIII-V/Pa(CSA)Ea transgene
by Northern analysis was evident only in the placenta after
prolonged exposure (data not shown). By the more sensitive
RT-PCR (Fig. 8C), hCS-A expression from the HSIII-V/
Pa(CSA)Ea transgene could be reliably detected in the pla-
centas of one of the four lines; a second line expressed at a
much lower level, and expression from the two remaining
lines could be detected only at trace levels by high-cycle
amplification (data not shown). Thus, the HSIII-V/Pa(CSA)Ea
transgene was poorly expressed in the placenta and its expression
failed to correlate with transgene copy number, indicating a
marked sensitivity to the site of transgene integration. The spec-
ificity of expression from the two transgenes was further assessed
in a panel of eight tissues (Fig. 8D). The expression of hCS from
the hGH/BAC was appropriately confined to the placenta, with
trace levels in the testes. In clear contrast, the expression of the
HSIII-V/Pa(CSA)Ea transgene was strongly expressed in the kid-
ney (Fig. 8D and data not shown).
In summary, the HSIII-V/Pa(CSA)Ea transgene was unable
to establish an autonomous chromatin domain or selectively
target gene expression to the placenta. The expression that did
occur in the placentas of the HSIII-V/Pa(CSA)Ea transgenic
embryos was well below that of the hGH/BAC transgene.
Thus, although early histone modifications within the cluster
were limited to the placental-gene units and the 5?, remote HS
determinants, additional sequences and/or genomic organiza-
tion are necessary for full and appropriate placental-gene ac-
The terminal differentiation of CTBs to STBs in placental
villi is paralleled by a well-defined shift in gene expression (5,
24). A robust example of this shift is the induction of hCS
expression from the hGH gene cluster. We report two dynamic
cell culture models of placental-cell differentiation that reca-
pitulate this shift in gene expression. These models reveal
distinct patterns and timing for two categories of activating
histone modifications within the locus—the methylation of
H3K4 and the acetylation of H3 and H4. The initial modifica-
tions that define the activated hGH cluster in both models are
localized to each of the conserved PGR units and to the 5?,
remote HS determinants. In vivo functional testing revealed
that these sites of initial chromatin modification are not in
themselves sufficient for full and appropriate activation of the
hGH cluster in the placenta. These data suggested that addi-
tional elements situated between the cluster and the remote 5?
HS determinants or involving the native overall configuration
of the multigene cluster are involved in this process.
Differentiation of the placental cells. Two cell culture mod-
els were found to reliably recapitulate transcriptional induction
of the placental genes from the hGH cluster. Clearly defined
increases in expression from the placental genes of the hGH
cluster were observed in BeWo cells during long-term culture.
However, the level of gene expression in this system, even at its
peak, was well below that of the primary placental STBs (Fig.
3C). This is not entirely surprising given that the BeWo cells,
even at high density, fail to show clear evidence for the syncy-
tialization characteristic of STB differentiation in vivo (unpub-
lished observations). The relevance of the BeWo cells to the in
vivo situation is, however, supported by molecular character-
izations of this cell culture model. Expression of a number of
genes playing important roles in STB formation, such as TEF-5
(28, 29, 30), PPAR? (6), Gcm-1 (4, 57), and HERV-W/syncytin
(9, 19, 44), can be detected, and syncytin mRNA levels are
induced threefold by day 16 of culture (our unpublished data).
Although it is not clear why growing the cells to high density
induced the placental-gene expression in the absence of syn-
cytialization, the high-density culture of BeWo cells reported
here represents a useful, although imperfect, model for pla-
The induction of gene expression from the hGH cluster in
the explanted CTBs was substantially more robust than that
observed in the BeWo cells. This model is also closer to the
“physiologic” system; primary CTBs freshly explanted from
term placenta form syncytia and recapitulate on a global scale
the STB gene expression profile (24). A weakness of this model
is the difficulty in obtaining sufficient numbers of cells for
detailed study. Importantly, we demonstrated in the present
study that both the day 0 BeWo cells and day 0 CTBs are
“poised” to express the placental genes. Thus, analysis of their
6564 KIMURA ET AL.MOL. CELL. BIOL.
chromatin structure can be informative in defining the initial
stage of this process.
Histone acetylation and methylation play different roles in
placental-gene activation within the hGH cluster. In a previous
report, we observed distinct patterns of histone acetylation and
methylation at the hGH locus in primary STBs (32). These
distinct patterns suggested that these modifications may be
independently targeted and may play different roles in the
control of gene expression. Our current data extend this ob-
servation by suggesting that localized histone H3K4 methyla-
tion and H4 acetylation precede global and robust H3 and H4
histone acetylation of the locus in both BeWo cell and primary
CTB cultures. In the BeWo cells, the histone H3K4 dimethyla-
tion pattern is present at day 0 and the subsequent induction of
placental-gene transcription is paralleled by generalized and
robust acetylation throughout the locus. Similarly, in the pri-
FIG. 8. HSIII-HSV is insufficient to activate hCS-A expression in the transgenic placenta. (A) Transgene constructs. Each of the two transgene
constructs is shown below the map of the hGH locus. Each of the four P-elements was named to correspond with its adjacent gene’s name (Pl, Pa,
Pv, and Pb), and the three enhancers were named similarly (El, Ea, and Eb). The hGH/BAC transgene encompasses the entire region. The
HSIII-V/Pa(CSA)Ea transgene is composed of an 11.5-kb fragment encompassing HSV through HSIII ligated directly to a 7.5-kb hCS-A gene
fragment that includes its contiguous 5? and 3? sequences, as shown. Sets of transgenic mouse lines were generated with each of these two
constructs. (B) Expression of the hGH cluster genes in the transgenic placentas. Four lines of mice carrying the HSIII-V/Pa(CSA)Ea transgene,
each with a unique transgene insertion site, were generated and studied (1196D, 1197C, 1199E, and 1200D). Similarly, four unique transgenic lines
carrying hGH/BAC were established and studied (1210B, 1252B, 1253E, 1254D). Total RNAs were purified from the transgenic placentas of the
indicated lines and subjected to Northern blot analysis. The transgene copy number in each line is indicated at the bottom of each lane. The probes
for 18S rRNA and mouse placental lactogen II (PLII) were used as controls. Although robust and copy-number-dependent hGH-hCS expression
was observed for all four hGH/BAC lines, no signals were detected for any of the HSIII-V/Pa(CSA)Ea lines at the level of Northern blot sensitivity.
(C) Comparison of the hCS-A expression levels in transgenic placentas. RT-PCR analysis was performed to amplify the hCS-A and hCS-B mRNAs
using total RNAs isolated from placentas of the indicated transgenic lines. The signal intensity for hCS-A was normalized to that of ?-actin as well
as to the transgene copy numbers noted in panel B. The calculated relative expression levels are plotted in the frame below. The hGH/BAC lines
showed strong site-of-integration-independent and copy-number-dependent expression of hCS-A in placenta. In contrast, expression in the
HSIII-V/Pa(CSA)Ea lines was quite low and showed marked position effects. (D) The HSIII-V/Pa(CSA) gene failed to retain placental specificity.
Northern blot analysis was performed with RNAs prepared from the indicated tissues. The tissues were isolated from the lines with the highest
transgene copy number for each construct. Line 1251F (hGH/BAC) had 35 copies of the transgene; line 1196C (HSIII-V/Pa(CSA)Ea) had 63 copies.
The oligoprobe for 18S rRNA was used as a loading control. hCS-A expression is predominantly placenta specific in the hGH/BAC line but is
ectopically expressed at high levels in kidneys of the HSIII-V/Pa(CSA)Ea mice.
VOL. 27, 2007 PLACENTAL ACTIVATION OF THE hGH CLUSTER 6565
mary CTBs at day 0, the histone H3K4 di- and trimethylation
within the LCR and within the placental-gene duplication units
appears to define the “poised” state of the cluster prior to
transcriptional activation. Significant levels of histone H4 acet-
ylation were also detected in the day 0 CTBs, primarily at HSV
and at the placental-gene promoters. By day 4 of culture, the
levels of methylation and acetylation increased and extended.
The patterns then became remarkably similar to, although not
as great as, those seen in primary STBs. Levels of histone
acetylation across the cluster appeared to track with levels of
gene transcription both in the BeWo cells and in the CTB
culture model. These results point to distinct roles for histone
H3K4 methylation in early gene activation and for H3 and H4
acetylation in subsequent transcriptional enhancement.
Distinct roles for histone acetylation and methylation during
cell differentiation or development in a number of systems
have been described. At the murine IgH locus, the HS4 regu-
latory determinant is H3K4 dimethylated in pro-B cells,
whereas histone acetylation is observed at later stages of B-cell
differentiation and during IgH induction (20). The mouse
Hoxb9 gene also undergoes H3K4 methylation prior to histone
H3 acetylation as embryonic stem cells are induced to differ-
entiate by retinoic acid (10). A similar schedule of modifica-
tions is observed at the human HNF4? promoter (25), the
human collagenase promoter (42), and the mouse IL-2 gene
regulatory region (1). These results are consistent with our
present data, suggesting that histone H3K4 methylation pre-
cedes acetylation at many gene loci and may serve specific
mechanistic functions in the early stages of gene activation.
However, this pattern is not absolute. Several reports have
indicated that histone acetylation and methylation may occur
concurrently or in the inverse temporal order (3, 37, 41, 64).
Thus, the epigenetic regulation of subgroups of genes can
reflect different temporal orders and functions of histone mod-
Activation of placental genes in the hGH cluster. The
“poised” configuration of the hGH cluster in day 0 BeWo cells
and CTBs is accompanied by the formation of placenta-specific
HSIV as well as constitutive HSIII and HSV (Fig. 4). There-
fore, the full sets of HSs that define the hGH LCR in the
placenta are already assembled at a time when the CS genes
were barely active. This finding is consistent with observations
at other loci that DNase I-HSs can form in differentiating cells
prior to transcriptional activation (3, 37, 64). The subsequent
acetylation of H3 and H4 at the placental LCR elements and
within the hGH cluster tracks with transcriptional induction
(Fig. 7). The coordinated increase in histone H3 acetylation at
the 5? remote HS and at the gene repeat units seen as CTBs
transit to STBs (day 4 CTBs to STBs) is consistent with a direct
mechanistic and possibly physical communication. A “looping”
model, previously proposed to explain the long-range interac-
tion of the LCR with the placental genes over a distance
exceeding 40 kb, is consistent with such interactions (15, 32).
The gene units in the hGH cluster are highly conserved in
structure, and the placental units appear to be individually
targeted by the initial H3K4 methylation activities. These ob-
servations suggest that the LCR and an individual PGR unit
might be sufficient for activating placental-gene expression. To
test this model, we compared the expression from the cluster
contained in a BAC transgene encompassing the entire cluster
and LCR with that in a more limited transgene containing the
HSIII-HSV region and the single most active PGR unit (hCS-
A). The BAC transgene demonstrated robust, placenta-spe-
cific, site-of-integration-independent, and copy-number-de-
pendent expression, indicating that the 123-kb sequence
included all the components needed to fully activate the pla-
cental genes in vivo (Fig. 8B and C). This is consistent with the
results of an analysis of mice carrying a somewhat shorter
hGH/P1 transgene (62). In contrast, expression of the HSIII-
V/Pa(CSA)Ea transgene was quite low in the placenta (Fig. 8B
and C), was sensitive to the site of integration, lacked placenta
specificity, and demonstrated marked ectopic expression in the
kidney (Fig. 8D). This inability of a transgene that links the
HSIII-HSV components directly to hCS-A to activate placental
expression contrasts markedly with the robust and pituitary-
specific expression of a parallel short transgene that links the
hGH-N gene directly to the pituitary-specific LCR unit (HSI,
HSII) (8, 26, 31, 58, 59). Thus, the current data highlight
substantial differences between the pathways of gene activation
of the pituitary and placental genes from the hGH cluster.
These data are in agreement with our prior model indicating
that activation of the hGH-N gene is mediated by direct ex-
tension of activating signals from its LCR elements, although
the active placental genes communicate with the correspond-
ing 5? remote determinants in a more complex manner (17, 26,
The present studies revealed that the HSIII-HSV region is
not sufficient to activate the hCS-A gene. It is useful in this
context to ask whether these remote 5? HSs that form in pla-
cental chromatin are in fact an LCR for placental-gene expres-
sion. Our studies have consistently demonstrated that the
HSIII-HSV region undergoes chromatin alterations that are
specific to the placenta; these sites form in placental chromatin
and include the specific formation of DNase I-HSIV (refer-
ence 31 and the present study). Furthermore, localized enrich-
ment for histone modifications is specific to the HSIII-HSV
region in the placenta (reference 32 and the present study). In
functional studies, we have shown that transgenes that retain
the HSIII-HSV region in its native spacing express hCS spe-
cifically and robustly in mouse placentas (62). Transgenes that
do not include the HSIII-HSV region do not express hCS in
mouse placentas (31). In addition, multiple transgenic lines in
which the HSIII-HSV region is intact but in which deletions
have removed limited determinants between HSIII and the
gene cluster failed to decrease or alter hCS transgene expres-
sion in mouse placentas (26). These combined data support a
model in which the HSIII-HSV region is a critical remote
regulatory region for placental-gene expression. However, the
present data demonstrating that the HSIII-HSV region is not
sufficient for hCS activation in the placenta differ from similar
studies that demonstrated the sufficiency of HSI and HSII
LCR elements to activate the hGH-N gene in the pituitary.
Thus, the question of whether the HSIII-HSV region is nec-
essary for this activity needs to be further addressed.
The present data lead us to conclude that the activation of
placental hGH-hCS genes involves the sequential establishment
of chromatin restructuring at the hGH LCR, discrete histone
H3K4 methylation targeted to the individual placental-gene units,
and subsequent and expansive histone H3 and H4 acetylation
during placental differentiation. The functional comparison of the
6566 KIMURA ET AL.MOL. CELL. BIOL.
hGH/BAC and HSIII-V/Pa(CSA)Ea transgenes points to the ex-
istence of elements in addition to the HSIII-HSV region and the
individual PGR unit that are indispensable for appropriate site-
of-integration-independent and copy-number-dependent expres-
sion of the hGH-hCS genes in placentas. These missing compo-
nents may be responsible for the higher-order organization of this
chromatin region and a corresponding looping configuration that
has been postulated to occur in the activation process (32). The
identification of such elements and structures, as well as the re-
construction of the entire process of placental hGH-hCS gene
activation, remains an important challenge for future studies.
We thank Michael Hubert for harvesting and culturing primary
CTBs. We thank Jean Richa, Pei Fu He, and Kathleen Moosbrugger
of the University of Pennsylvania Transgenic and Chimeric Mouse
Core Facility for their help in generating transgenic mice.
This work was supported by R01s HD46737, HD25147 (N.E.C. and
S.A.L.), and HD07447 (S.H.).
1. Adachi, S., and E. V. Rothenberg. 2005. Cell-type-specific epigenetic marking
of the IL2 gene at a distal cis-regulatory region in component, nontranscrib-
ing T-cells. Nucleic Acids Res. 33:3200–3210.
2. Alsat, E., J. Guibourdenche, D. Luton, F. Frankenne, and D. Evain-Brion.
1997. Human placental growth hormone. Am. J. Obstet. Gynecol. 177:1526–
3. Anguita, E., J. Hughes, C. Heyworth, G. A. Blobel, W. G. Wood, and D. R.
Higgs. 2004. Globin gene activation during haemopoiesis is driven by protein
complexes nucleated by GATA-1 and GATA-2. EMBO J. 23:2841–2852.
4. Anson-Cartwright, L., K. Dawson, D. Holmyard, S. J. Fisher, R. A. Lazzarini,
and J. C. Cross. 2000. The glial cells missing-1 protein is essential for branching
morphogenesis in the chorioallantoic placenta. Nat. Genet. 25:311–314.
5. Aronow, B. J., B. D. Richardson, and S. Handwerger. 2001. Microarray
analysis of trophoblast differentiation: gene expression reprogramming in
key function categories. Physiol. Genomics 6:105–116.
6. Barak, Y., M. C. Nelson, E. S. Ong, Y. Z. Jones, P. Ruiz-Lozano, K. R. Chien,
A. Koder, and R. M. Evans. 1999. PPAR? is required for placental, cardiac,
and adipose tissue development. Mol. Cell 4:585–595.
7. Barrera-Saldan ˜a, H. A. 1998. Growth hormone and placental lactogen: bi-
ology, medicine and biotechnology. Gene 211:11–18.
8. Bennani-Baiti, I. M., S. L. Asa, D. Song, R. Iratni, S. A. Liebhaber, and N. E.
Cooke. 1998. DNase I-hypersensitive sites I and II of the human growth
hormone locus control region are a major developmental activator of soma-
totrope gene expression. Proc. Natl. Acad. Sci. USA 95:10655–10660.
9. Blond, J.-L., F. Bese `me, L. Duret, O. Bouton, F. Bedin, H. Perron, B.
Mandrand, and F. Mallet. 1999. Molecular characterization and placental
expression of HERV-W, a new human endogenous retrovirus family. J. Vi-
10. Chambeyron, S., and W. A. Bickmore. 2004. Chromatin decondensation and
nuclear reorganization of the HoxB locus upon induction of transcription.
Genes Dev. 18:1119–1130.
11. Chen, E. Y., Y.-C. Liao, D. H. Smith, H. A. Barrera-Saldan ˜a, R. E. Gelinas,
and P. H. Seeburg. 1989. The human growth hormone locus: nucleotide
sequence, biology, and evolution. Genomics 4:479–497.
12. Cross, J. C., Z. Werb, and S. J. Fisher. 1994. Implantation and the placenta:
key pieces of the development puzzle. Science 266:1508–1518.
13. Cross, J. C., D. Baczyk, N. Dobric, M. Hemberger, M. Hughes, D. G. Simmons,
H. Yamamoto, and J. C. P. Kingdom. 2003. Genes, development and evolution
of the placenta. Placenta 24:123–130.
14. Cross, J. C. 2005. How to make a placenta: mechanisms of trophoblast cell
differentiation in mice. Placenta 26(Suppl. A):S3–S9.
15. Dean, A. 2006. On a chromosome far, far away: LCRs and gene expression.
Trends Genet. 22:38–45.
16. Elefant, F., N. E. Cooke, and S. A. Liebhaber. 2000. Targeted recruitment of
histone acetyltransferase activity to a locus control region. J. Biol. Chem.
17. Elefant, F., Y. Su, S. A. Liebhaber, and N. E. Cooke. 2000. Patterns of histone
acetylation suggest dual pathways for gene activation by a bifunctional locus
control region. EMBO J. 19:6814–6822.
18. Frankenne, F., J. Closset, F. Gomez, M. L. Scippo, J. Smal, and G. Hennen.
1988. The physiology of growth hormones (GHs) in pregnant women and
partial characterization of the placental GH variant. J. Clin. Endocrinol.
19. Frendo, J.-L., D. Olivier, V. Cheynet, J.-L. Blond, O. Bouton, M. Vidaud, M.
Rabreau, D. Evain-Brion, and F. Mallet. 2003. Direct involvement of
HERV-W Env glycoprotein in human trophoblast cell fusion and differen-
tiation. Mol. Cell. Biol. 23:3566–3574.
20. Garrett, F. E., A. V. Emelyanov, M. A. Sepulveda, P. Flanagan, S. Volpi, F.
Li, D. Loukinov, L. A. Eckhardt, V. V. Lobanenkov, and B. K. Birshtein.
2005. Chromatin architecture near a potential 3? end of the Igh locus in-
volves modular regulation of histone modifications during B-cell develop-
ment and in vivo occupancy at CTCF sites. Mol. Cell. Biol. 25:1511–1525.
21. Goldman, W. E., G. Goldberg, L. H. Bowman, D. Steinmetz, and D.
Schlessinger. 1983. Mouse rDNA: sequences and evolutionary analysis of
spacer and mature RNA regions. Mol. Cell. Biol. 3:1488–1500.
22. Goodman, H. M., L.-R. Tai, J. Ray, N. E. Cooke, and S. A. Liebhaber. 1991.
Human growth hormone variant produces insulin-like and lipolytic re-
sponses in rat adipose tissue. Endocrinology 129:1779–1783.
23. Handwerger, S. 1991. Clinical counterpoint: the physiology of placental
lactogen in human pregnancy. Endocr. Rev. 12:329–336.
24. Handwerger, S., and B. Aronow. 2003. Dynamic changes in gene expression
during human trophoblast differentiation. Recent Prog. Horm. Res. 58:263–
25. Hatzis, P., and I. Talianidis. 2002. Dynamics of enhancer-promoter commu-
nication during differentiation-induced gene activation. Mol. Cell 10:1467–
26. Ho, Y., F. Elefant, N. Cooke, and S. Liebhaber. 2002. A defined locus control
region determinant links chromatin domain acetylation with long-range gene
activation. Mol. Cell 9:291–302.
27. Ho, Y., F. Elefant, S. A. Liebhaber, and N. E. Cooke. 2006. Locus control
region transcription plays an active role in long-range gene activation. Mol.
28. Jacquemin, P., J. A. Martial, and I. Davidson. 1997. Human TEF-5 is
preferentially expressed in placenta and binds to multiple functional ele-
ments of the human chorionic somatomammotropin-B gene enhancer.
J. Biol. Chem. 272:12928–12937.
29. Jacquemin, P., V. Sapin, E. Alsat, D. Evain-Brion, P. Dolle, and I. Davidson.
1998. Differential expression of the TEF family of transcription factors in the
murine placenta and during differentiation of primary human trophoblasts in
vitro. Dev. Dyn. 212:423–436.
30. Jiang, S.-W., K. Wu, and N. L. Eberhardt. 1999. Human placental TEF-5
transactivates the human chorionic somatomammotropin gene enhancer.
Mol. Endocrinol. 13:879–889.
31. Jones, B. M., B. R. Monks, S. A. Liebhaber, and N. E. Cooke. 1995. The
human growth hormone gene is regulated by a multicomponent locus control
region. Mol. Cell. Biol. 15:7010–7021.
32. Kimura, A. P., S. A. Liebhaber, and N. E. Cooke. 2004. Epigenetic modifi-
cations at the human growth hormone locus predict distinct roles for histone
acetylation and methylation in placental gene activation. Mol. Endocrinol.
33. Kliman, H. J., J. E. Nestler, E. Sermasi, J. M. Sanger, and J. F. Strauss III.
1986. Purification, characterization, and in vitro differentiation of cytotro-
phoblast from human term placentae. Endocrinology 118:1567–1582.
34. Kno ¨fler, M., R. Vasicek, and M. Schreiber. 2001. Key regulatory transcrip-
tion factors involved in placental trophoblast development. Placenta
35. Lachner, M., R. J. O’Sullivan, and T. Jenuwein. 2003. An epigenetic road
map for histone lysine methylation. J. Cell Sci. 116:2117–2124.
36. Lacroix, M. C., J. Guibourdenche, J. L. Frendo, F. Muller, and D. Evain-
Brion. 2002. Human placental growth hormone. Placenta 23(Suppl. A):S87–
37. Litt, M. D., M. Simpson, M. Gaszner, C. D. Allis, and G. Felsenfeld. 2001.
Correlation between histone lysine methylation and developmental changes
at the chicken ?-globin locus. Science 293:2453–2455.
38. Loregger, T., J. Pollheimer, and M. Kno ¨fler. 2003. Regulatory transcription
factors controlling function and differentiation of human trophoblast—a
review. Placenta 24(Suppl. A):S104–S110.
39. MacLeod, J. N., I. Worsley, J. Ray, H. G. Friesen, S. A. Liebhaber, and N. E.
Cooke. 1991. Human growth hormone-variant is a biologically active somato-
gen and lactogen. Endocrinology 128:1298–1302.
40. MacLeod, J. N., A. K. Lee, S. A. Liebhaber, and N. E. Cooke. 1992. Devel-
opmental control and alternative splicing of the placentally expressed tran-
scripts from the human growth hormone gene cluster. J. Biol. Chem. 267:
41. Margueron, R., P. Trojer, and D. Reinberg. 2005. The key to development:
interpreting the histone code? Curr. Opin. Genet. Dev. 15:163–176.
42. Martens, J. H. A., M. Verlaan, E. Kalkhoven, and A. Zantema. 2003. Cas-
cade of distinct histone modifications during collagenase gene activation.
Mol. Cell. Biol. 23:1808–1816.
43. Martin, C., and Y. Zhang. 2005. The diverse functions of histone lysine
methylation. Nat. Rev. Mol. Cell Biol. 6:838–849.
44. Mi, S., X. Lee, X.-P. Li, G. M. Veldman, H. Finnerty, L. Racie, E. LaVallie,
X.-Y. Tang, P. Edouard, S. Howes, J. C. Keith, Jr., and J. M. McCoy. 2000.
Syncytin is a captive retroviral envelope protein involved in human placental
morphogenesis. Nature 403:785–789.
45. Misra-Press, A., N. E. Cooke, and S. A. Liebhaber. 1994. Complex alterna-
VOL. 27, 2007PLACENTAL ACTIVATION OF THE hGH CLUSTER 6567
tive splicing partially inactivates the human chorionic somatomammotropin- Download full-text
like (hCS-L) gene. J. Biol. Chem. 269:23220–23229.
46. Nachtigal, M. W., B. E. Nickel, and P. A. Cattini. 1993. Pituitary-specific
repression of placental members of the human growth hormone gene family:
a possible mechanism for locus regulation. J. Biol. Chem. 268:8473–8479.
47. Norwitz, E. R., D. J. Schust, and S. J. Fisher. 2001. Implantation and the
survival of early pregnancy. N. Engl. J. Med. 345:1400–1408.
48. Ohlsson, C., B. A. Bengtsson, O. G. Isaksson, T. T. Andreassen, and M. C.
Slootweg. 1998. Growth hormone and bone. Endocr. Rev. 19:55–79.
49. Peterson, C. L., and M.-A. Laniel. 2004. Histones and histone modifications.
Curr. Biol. 14:R546–R551.
50. Po ¨tgens, A. J. G., U. Schmitz, P. Bose, A. Versmold, P. Kaufmann, and H.-G.
Frank. 2002. Mechanisms of syncytial fusion. Placenta 23(Suppl. A):S107–
51. Red-Horse, K., Y. Zhou, O. Genbacev, A. Prakobphol, R. Foulk, M. McMaster,
and S. J. Fisher. 2004. Trophoblast differentiation during embryo implantation
and formation of the maternal-fetal interface. J. Clin. Investig. 114:744–754.
52. Richards, R. G., S. M. Hartman, and S. Handwerger. 1994. Human cytotro-
phoblast cells cultured in maternal serum progress to a differentiated syn-
cytial phenotype expressing both human chorionic gonadotropin and human
placental lactogen. Endocrinology 135:321–329.
53. Roberts, R. M., T. Ezashi, and P. Das. 2004. Trophoblast gene expression:
transcription factors in the specification of early trophoblast. Reprod. Biol.
54. Rossant, J., and J. C. Cross. 2001. Placental development: lessons from
mouse mutants. Nat. Rev. Genet. 2:538–548.
55. Rote, N. S., S. Chakrabarti, and B. P. Stetzer. 2004. The role of human
endogenous retroviruses in trophoblast differentiation and placental devel-
opment. Placenta 25:673–683.
56. Santos-Rosa, H., R. Schneider, A. J. Bannister, J. Sherriff, B. E. Bernstein,
N. C. T. Emre, S. L. Schreiber, J. Mellor, and T. Kouzarides. 2002. Active
genes are tri-methylated at K4 of histone H3. Nature 419:407–411.
57. Schreiber, J., E. Riethmacher-Sonnenberg, D. Riethmacher, E. E. Tuerk, J.
Enderich, M. R. Bo ¨sl, and M. Wegner. 2000. Placental failure in mice lacking
the mammalian homolog of glial cells missing, GCMa. Mol. Cell. Biol.
58. Shewchuk, B. M., S. L. Asa, N. E. Cooke, and S. A. Liebhaber. 1999. Pit-1
binding sites at the somatotrope-specific DNase I hypersensitive sites I, II of
the human growth hormone locus control region are essential for in vivo
hGH-N gene activation. J. Biol. Chem. 274:35725–35733.
59. Shewchuk, B. M., S. A. Liebhaber, and N. E. Cooke. 2002. Specification of
unique Pit-1 activity in the hGH locus control region. Proc. Natl. Acad. Sci.
60. Simmons, D. G., and J. C. Cross. 2005. Determinants of trophoblast lineage
and cell subtype specification in the mouse placenta. Dev. Biol. 284:12–24.
61. Simpson, E. R., and P. C. MacDonald. 1981. Endocrine physiology of the
placenta. Annu. Rev. Physiol. 43:163–188.
62. Su, Y., S. A. Liebhaber, and N. E. Cooke. 2000. The human growth hormone
gene cluster locus control region supports position-independent pituitary-
and placenta-specific expression in the transgenic mouse. J. Biol. Chem.
63. Sullivan, M. H. F. 2004. Endocrine cell lines from the placenta. Mol. Cell.
64. Szutorisz, H., C. Canzonetta, A. Georgiou, C.-M.Chow, L. Tora, and N.
Dillon. 2005. Formation of an active tissue-specific chromatin domain initi-
ated by epigenetic marking at the embryonic stem cell stage. Mol. Cell. Biol.
65. Taylor, R. N., E. D. Newman, and S. Chen. 1991. Forskolin and methotrexate
induce an intermediate trophoblast phenotype in cultured human choriocar-
cinoma cells. Am. J. Obstet. Gynecol. 164:204–210.
66. Walker, W. H., S. L. Fitzpatrick, and G. F. Saunders. 1990. Human placental
lactogen transcriptional enhancer: tissue specificity and binding with specific
proteins. J. Biol. Chem. 265:12940–12948.
67. Watson, E. D., and J. C. Cross. 2005. Development of structures and trans-
port functions in the mouse placenta. Physiology 20:180–193.
68. Xu, B., L. Lin, and N. S. Rote. 1999. Identification of a stress-induced protein
during human trophoblast differentiation by differential display analysis.
Biol. Reprod. 61:681–686.
69. Yoo, E. J., I. Cajiao, J.-S. Kim, A. P. Kimura, A. Zhang, N. E. Cooke, and
S. A. Liebhaber. 2006. Tissue-specific chromatin modifications at a multigene
locus generate asymmetric transcriptional interactions. Mol. Cell. Biol. 26:
70. Young, J. A. T., and J. Trowsdale. 1985. A processed pseudogene in an intron
of the HLA-DP?1 chain gene is a member of the ribosomal protein L32 gene
family. Nucleic Acids Res. 13:8883–8891.
6568 KIMURA ET AL.MOL. CELL. BIOL.