Review

Gestational trophoblastic neoplasia—pathogenesis and

potential therapeutic targets
Ie-Ming Shih

Lancet Oncol 2007; 8: 642–50 Gestational trophoblastic neoplasia comprises a unique group of human neoplastic diseases that derive from fetal
Departments of Pathology, trophoblastic tissues and represent semiallografts in patients. This group is composed of choriocarcinoma, placental-
Oncology, and Gynecology and site trophoblastic tumour, and epithelioid trophoblastic tumour, and many forms are derived from the precursor
Obstetrics, Johns Hopkins
lesions, hydatidiform moles. Although most patients with gestational trophoblastic neoplasia are cured by
Medical Institutions,
Baltimore, MD, USA chemotherapy and tumour resection, some patients suffer from metastatic diseases that are refractory to conventional
(I-M Shih MD) chemotherapy. Therefore, new therapeutic regimens are needed to reduce the toxic effects associated with current
Correspondence to: chemotherapy and to salvage the occasional non-operable patients with recurrent and chemoresistant disease. Until
Dr Ie-Ming Shih the fundamental biology of gestational trophoblastic neoplasia becomes more clearly understood, development of a
shihie@yahoo.com
new treatment will remain empirical. This review will briefly summarise the recent advances in understanding the
molecular aetiology of this group of diseases and highlight the molecules that can be potentially used for therapeutic
targets to treat metastatic gestational trophoblastic neoplasia.

Introduction aberrant placental derivatives; and true gestational

Gestational trophoblastic diseases represent a spectrum
of related disorders including benign trophoblastic
lesions, premalignant hydatidiform moles, clinically
malignant invasive hydatidiform moles, and neoplastic
diseases. Hydatidiform moles, especially the invasive
moles, are sometimes clinically considered as gestational
trophoblastic neoplasia because they can locally invade
and distantly metastasise, but, biologically, they represent
abnormally formed placental tissues rather than true
neoplasia. Therefore, from the perspectives of patho-
biology, gestational trophoblastic diseases can be broadly
divided into three groups (panel), modified from the
2003 WHO classification1: benign trophoblastic lesions
(placental-site nodule and exaggerated placental reaction),
which are non-neoplastic lesions; hydatidiform moles
(complete, partial, and invasive moles), which are
Figure 1: ETT (arrow) in a hysterectomy specimen of a 32-year-old woman
ETT was first diagnosed based on endometrial curettage and patient subsequently underwent total hysterectomy.
1 year after surgery, metastatic ETT developed in the lung and was treated with chemotherapy and lung resection.
A second recurrent tumour, resistant to prior chemotherapy, developed 1 year later.
trophoblastic neoplasia.2,3
The term gestational trophoblastic neoplasia encom-
passes a group of interrelated but distinct tumours that
include choriocarcinoma, placental-site trophoblastic
tumour (PSTT), and epithelioid trophoblastic tumour
(ETT; figure 1). Gestational trophoblastic neoplasms are
unique among human neoplastic diseases because they
are genetically related to fetal tissues (ie, trophoblastic
cells from placentas) and, therefore, represent semi-
allografts in patients.
Although there has been substantial interest in
exploring the pathogenesis of these neoplasms, studies
on their molecular aspects are challenging for several
reasons. First, gestational trophoblastic neoplasms are
rare diseases and the cumulative incidence rate for
gestational choriocarcinoma, for example, continues to
decline. As a result, tissue specimens, especially those
from fresh tumours, are scarce for molecular analysis.
Second, many metastatic gestational trophoblastic neo-
plasms, which are clinically classified under the rubric of
persistent gestational trophoblastic diseases are not
usually obtained for pathological classification and
characterisation. Therefore, their natures remain intrig-
uing, because whether persistent gestational tropho-
blastic disease represents an invasive hydatidiform mole,
a metastatic choriocarcinoma, PSTT, or ETT is not
known. An absence of such clinicopathological correlation
compromises attempts to further understand the
biological nature of gestational trophoblastic neoplasia.
Third, because of the above reasons, reagents, such as
cell lines and animal models, are limited for mechanistic
studies. For example, only a few choriocarcinoma cell
lines and one PSTT cell line are currently available for
molecular and cell-biology studies.
Despite these challenges, progress has been made in
recent years towards a better understanding of the
molecular aetiology underlying the development of these
neoplasms. This review will briefly summarise recent

642 http://oncology.thelancet.com Vol 8 July 2007


Panel: Gestational trophoblastic diseases
Benign trophoblastic lesions
Exaggerated placental reaction
Placental-site nodule
Hydatidiform moles (abnormally formed placentas)
Complete mole
Partial mole
Invasive mole
Trophoblastic neoplasia
Choriocarcinoma
Placental-site trophoblastic tumour
Epithelioid tophoblastic tumour
advances and highlight the molecules that could
potentially be used for therapeutic targets to treat meta-
static gestational trophoblastic neoplasia.
Pathogenesis of trophoblastic tumours
Previous clinical, pathological, and molecular studies
have provided fundamental insights into the pathogenesis
of hydatidiform moles,4 but the molecular and cellular
basis in the development of true gestational trophoblastic
neoplasia remain poorly understood.
Gestational trophoblastic neoplasia has long been
thought to be a homogeneous group of diseases arising
from neoplastic transformation of trophoblastic cells.
However, recent clinicopathological studies have provided
new evidence that there are at least three distinct types of
gestational trophoblastic neoplasm including the most
common type, choriocarcinoma, and the less common
ones, PSTT and ETT. Molecular analysis of gestational
trophoblastic neoplasia is largely based on the
characterisation of gene-expression profiles in various
types of these neoplasms and the reference of their
unique gene-expression patterns to different trophoblastic
subpopulations in healthy early placentas.2 The main
conclusion from these studies is that, after neoplastic
transformation of trophoblastic stem cells, presumably
cytotrophoblast stem cells, specific differentiation progra-
mmes dictate the type of trophoblastic tumour that
develops (figure 2).5
Of interest, is the fact that these patterns of
differentiation in gestational trophoblastic neoplasia
recapitulate the stages of early placental development.2,3,5–7
Choriocarcinoma is composed of variable amounts of
neoplastic cytotrophoblast, syncytiotrophoblast, and extra-
villous (intermediate) trophoblast and resembles the
previllous blastocyst, which is composed of a similar
mixture of trophoblastic subpopulations. By contrast, the
neoplastic cytotrophoblast in PSTT differentiates mainly
into extravillous (intermediate) trophoblastic cells in an
implantation site whereas the neoplastic cytotrophoblast
in ETT differentiates into chorionic-type extravillous
(intermediate) trophoblastic cells in the chorion laeve.5
http://oncology.thelancet.com Vol 8 July 2007
Review
According to this model, choriocarcinoma is the most
primitive trophoblastic tumour whereas PSTT and ETT
are relatively more differentiated. This hypothesis
explains the existence of gestational trophoblastic
neoplasia with a mixed histological feature, including
choriocarcinoma, PSTT, and ETT. This new model might
also help to explain the findings previously reported by
Mazur8 who described ETT that occurred after intense
chemotherapy of metastatic choriocarcinomas in the
lung. In the patients described by Mazur, it is plausible
that chemotherapeutic agents allow choriocarcinoma
cells to differentiate into an ETT phenotype, which is
more refractory to chemotherapy than the original chorio-
carcinoma counterpart. As a result, ETT predominated
the choriocarcinoma after chemotherapy.
Alongside studies of the histogenesis of gestational
trophoblastic neoplasia, several studies have focused on
the biology of these neoplasms by characterising the
expressions of genes that are important in tumorigenesis.
The pathogenesis of each type of neoplasm will now be
summarised. Furthermore, advances in the pathogenesis
of hydatidiform moles will be briefly highlighted because
hydatidiform moles represent the precursor lesions in
some gestational trophoblastic neoplasms and invasive
hydatidiform moles are, sometimes, clinically considered
as gestational trophoblastic neoplasia because they can
locally invade and distantly metastasize.
Choriocarcinoma
Gestational choriocarcinoma is a highly malignant
epithelial tumour that can be associated with any type of
gestational event, most often a complete hydatidiform
mole.4 Several molecular studies have been done to
establish the expression of tumour-associated proteins in
choriocarcinoma. In the P53 pathway, overexpression of
Cytotrophoblast
Syncytiotrophoblast
Normal
trophoblast
Extravillous trophoblast Extravillous trophoblast
(intermediate trophoblast) (intermediate trophoblast)
at implantation site at chorion laeve
Neoplastic
transformation
Trophoblastic
neoplasia
Placental-site Epithelial trophoblastic
trophoblastic tumour
Choriocarcinoma
tumour
Figure 2: Proposed model of pathogenesis for gestational trophoblastic neoplasia
643
 Review
the P53 protein has been detected in choriocarcinoma,9
but mutational analysis of TP53 has failed to show
somatic mutations.10–12 The P53-associated protein,
MDM2, has also been seen to be overexpressed in chorio-
carcinomas.9,13 Because upregulated MDM2 protein is
associated with various human cancers, overexpression
of MDM2 in choriocarcinomas might overcome growth
suppression by wild-type P53 proteins and, therefore,
contribute to the development of choriocarcinoma. To
show the biological role of MDM2 in choriocarcinoma,
Wang and co-workers14 used an MDM2 antisense
oligonucleotide and showed a synergistic antitumour
effect with DNA-damaging agents in JAR choriocarcinoma
xenografts in mice.
By use of a PCR-based subtracted fragmentary cDNA
library between normal chorionic villi and a chorio-
carcinoma cell line, Asanoma and colleagues15,16 identified
a novel homeobox gene designated NECC1 (located on
4q11-q12) that might be involved in the development of
choriocarcinoma. In this study, the expression of NECC1
was lost in all choriocarcinoma cell lines and most
choriocarcinoma tissues, but healthy adult tissues,
including trophoblastic cells in healthy placentas,
expressed abundant NECC1. The investigators further
showed that engineered expression of NECC1 in
choriocarcinoma cell lines suppressed tumorigenicity
and induced terminal differentiation.15,16
Other genes that are potentially involved in the
development of choriocarcinoma include epidermal
growth factor receptor (EGFR),17 DOC-2/hDab2
(a candidate tumour suppressor gene),18,19 putative
tumour suppressor loci at chromosomes 7p12-7q11.2320
and 8p12-p21,21 and the gene for ras GTPase activating
protein.22 Synergistic upregulation of c-MYC, c-ERB-2,
c-FMS, and BCL-2 oncoproteins have also been
suggested to have an important role in the pathogenesis
of choriocarcinoma.10 Conversely, promoter hyper-
methylation of E-cadherin, HIC-1, P16, and TIMP3 has
been frequently seen in choriocarcinomas, suggesting
that downregulation of these genes by promoter
methylation could be involved in the development of
choriocarcinomas.23 In addition to the candidate
approaches described above, Vegh and colleagues24
compared the gene-expression pattern between healthy
trophoblast and choriocarcinoma cell lines using cDNA
microarrays. In their study, one of the downregulated
genes in choriocarcinoma cells, HSP-27, was selected
for characterisation because the HSP-27 protein has
been implicated in the development of several human
neoplastic diseases.25 The investigators noted that
choriocarcinoma tissues had a lower level of HSP-27
expression compared with trophoblastic cells in early
placentas. This downregulation of HSP-27 might
contribute to the high sensitivity of choriocarcinomas to
chemotherapeutic agents because HSP-27 proteins have
been shown to contribute to drug resistance in
malignant cells.26,27
644
The development of a choriocarcinoma might also be
related to tumour microenvironment—eg, immune
regulation and stromal tissues. HLA-G, a non-classic
MHC class I molecule, has been shown to participate in
immune regulation and response. The membrane-bound
and secreted HLA-G molecules are able to inactivate the
local immune response and, therefore, help tumour cells
escape from immune surveillance. Trophoblastic cells in
normal placentas and gestational trophoblastic neoplasms
have been shown to upregulate HLA-G expression.28,29 In
fact, in the human cancers analysed, gestational tropho-
blastic neoplasms, including choriocarcinoma, contain
the highest levels of HLA-G expression.28 Therefore,
HLA-G expressed in choriocarcinomas might assist
tumour cells to escape from the host immune recognition
and promote tumour growth.29
Another explanation for tumour growth could be
matrix metalloproteinases (MMP) and tissue inhibitors
of metalloproteinases (TIMPs), which have a key role in
tumour invasion. MMP is a zinc-dependent proteinase
that degrades and remodels the extracellular matrix
surrounding tumour cells. Choriocarcinomas have been
shown to express increased concentrations of MMP and
decreased concentrations of TIMP to help with tumour
invasion and metastasis,30 and this finding could explain
why choriocarcinoma is highly metastatic if left
untreated.
PSTT and ETT
PSTT and ETT are rare gestational trophoblastic
neoplasms and, so far, have not been well studied at the
molecular level, although both tumours have been
described in the published work on diagnostic pathology.2
Both forms of these neoplasms can be derived from any
type of gestational event, including complete and partial
hydatidiform moles.7,31–33 The origin of both tumours has
puzzled investigators since their discovery because,
unlike choriocarcinomas, PSTTs and ETTs can develop
long after the prior gestational events. The trophoblastic
nature of PSTT and ETT has been shown by molecular
genetic analysis, which shows that they contain new
(paternal) alleles not present in adjacent healthy uterine
tissue.34–36 Additionally, both tumours express trophoblast-
associated markers, including HLA-G and HSD3B1.29 On
the basis of published work, most cases of PSTT and ETT
are thought to be benign, especially for those tumours
that are confined to the uterus, but about 15–25% of cases
are malignant and present with local invasion and distant
metastasis.37–40 In some cases, recurrent or metastatic
PSTT and ETT can occur in patients long after the initial
treatment but, unfortunately, analysis of histopathological
features and molecular changes has not provided reliable
markers to predict the long-term clinical behaviour of
PSTT and ETT.41
By contrast to the normal implantation site where
invasion of the extravillous (intermediate) trophoblast is
tightly regulated and confined to the inner third of the
http://oncology.thelancet.com Vol 8 July 2007

myometrium, tumour cells of PSTT and ETT are highly
invasive and infiltrate deeply into the myometrium.
Although PSTT and ETT share similar clinical features,
careful examination of tumour histology and gene-
expression patterns shows that PSTT and ETT are
composed of different extravillous (intermediate) tropho-
blastic cells.3,7,38 During neoplastic transformation, the
tumour cells in PSTT and ETT are likely to assume
unique differentiation programmes toward the extra-
villous (intermediate) trophoblast in the implantation
site and the extravillous (intermediate) trophoblast in the
chorion laeve, respectively (figure 2).
Molecularly, PSTT is associated with abnormal
expression of cell-cycle regulatory genes, including those
for cyclins, cyclin-dependent kinases, and P53.42 Most
PSTTs are diploid, based on flow cytometric DNA
analysis, but PSTTs might also be triploid and, thus, be
derived from a partial hydatidiform mole.31 ETT expresses
markers of epithelial cells, including cytokeratin,
epithelial membrane antigen, E-cadherin, and EGFR,
and this finding is compatible with their characteristic
epithelioid growth pattern.
The P63 gene, a transcription factor belonging to the
P53 family, is expressed in ETT but not in PSTT. P63 has
various isoforms that are classified into two groups
designated as TAP63 and ∆NP63 isoforms.38 The
cytotrophoblast on chorionic villi has been shown to
express the ∆NP63 isoform, whereas extravillous
(intermediate) trophoblastic cells in the chorion laeve
and ETT express the TAP63.38 Further studies are needed
to delineate the functional role of TAP63 in ETT.
Expression of cyclin E is found in ETT but not in the
extravillous (intermediate) trophoblastic cells of the
chorion laeve, the normal counterpart of ETT.43 This
finding suggests that cyclin E probably plays a part in
neoplastic transformation of ETT because an oncogenic
role of cyclin E has been shown in other neoplastic
diseases.
Hydatidiform moles
Hydatidiform moles include complete, partial, and
invasive moles, the latter representing a hydatidiform
mole that becomes invasive or is exported to a distant
anatomical site. Although hydatidiform moles are not
classed as true gestational trophoblastic neoplasia,
postmolar gestational trophoblastic neoplasia can be
seen in many patients with hydatidiform moles,
especially complete moles. This finding indicates that
hydatidiform moles are the precursor lesions of some
gestational trophoblastic neoplasms. Because of the
close relationship of these moles with gestational
trophoblastic neoplasia, their pathogenesis will be briefly
highlighted.
Advances in cytogenetics, and in the molecular and
immunohistochemical characterisation of hydatidiform
moles, have shown a distinct pathogenesis in complete
and partial moles. Most complete moles are diploid and
http://oncology.thelancet.com Vol 8 July 2007
Review
usually have the karyotype 46,XX with both sets of
chromosomes originating from the paternal complement
as a result of haploid genome (23X) duplication. This
chromosomal composition is referred to as androgenetic,
meaning that genetic material is derived exclusively
from paternal DNA. Partial moles, however, are generally
triploid (XXY) with two sets of chromosomes from the
paternal complement and a haploid set from the
maternal genetic content. This triploid chromosomal
composition is known as diandric or biparental. Under-
standing the cytogenetic constituents that characterise
complete and partial moles is fundamental for the
future study of how maternally or paternally imprinted
genes contribute to different types of molar pregnancy.
Despite the well-established cytogenetic findings
associated with different hydatidiform moles, the
molecular aetiology underlying the development of
molar pregnancy remains elusive. One of the most
exciting studies showed that MALP7 is mutated in
patients with familial and recurrent biparental moles.44
MALP7 encodes a member of the CATERPILLER protein
family that is potentially involved in inflammatory and
apoptotic processes. More work is needed to elucidate
the possible mechanisms of mutant NALP7 proteins in
the pathogenesis of hydatidiform moles, such as during
oogenesis and implantation.
Additionally, several oncogenes and tumour-suppressor
genes have been studied in complete moles. Similar to
choriocarcinomas, synergistic upregulation of c-MYC,
c-ERB-2, c-FMS, and BCL-2 oncoproteins has been
suggested in the pathogenesis of complete moles.10
Mutational analysis of K-ras and P53 did not show
mutations in complete moles, but overexpression of P53
and MDM2 proteins has been recorded.9 Furthermore,
the role of promoter methylation has been studied in
hydatidiform moles and gestational trophoblastic
neoplasia. Frequent promoter hypermethylation and
decreased expression of PTEN, CDH1, HIC-1, and
CDKN2A have been found in hydatidiform moles
compared with healthy placentas,23,45 suggesting that
inactivating genes that might participate in tumour
suppression by promoter methylation could be one of the
mechanisms in the development of molar pregnancy.
Furthermore, promoter hypermethylation of CDKN2A
alone, or combined with CDH1, was significantly
correlated with subsequent development of gestational
trophoblastic neoplasia.23
New molecular targets with therapeutic potential
Choriocarcinoma is among the few human cancers that are
highly responsive to chemotherapy and, unlike other
malignant tumours, metastatic choriocarcinomas are
potentially curable by combined chemotherapy and
adjuvant surgical procedures.46,47 Despite this fact, a small
but significant proportion of choriocarcinoma patients
develop recurrent diseases after primary treatment.48,49 By
contrast with choriocarcinoma, PSTT and ETT are generally
645
 Review
more refractory to conventional chemotherapy. Patients
with PSTT and ETT might develop metastatic diseases and
some patients will have terminal disease.7,37 Therefore,
new therapeutic regimens are needed to reduce the toxic
effects associated with current chemotherapy—especially
the combination regimens, ie, etoposide, methotrexate,
dactinomycin; etoposide, methotrexate, dactinomycin,
cyclophosphamide, vincristine; and etoposide, cisplatin,
etoposide, methotrexate, and dactinomycin—and to salvage
the occasional non-operable patients with metastatic
chemoresistant diseases.
Target-based treatment designed to inactivate molecular
pathways that are essential for tumour-cell growth and
survival has shown clinical promise. Unlike standard
chemotherapy, which affects most proliferating cells,
inhibitors that target specific pathways selectively eliminate
tumour cells to achieve maximum therapeutic effects and
minimum adverse side-effects. These new anticancer
agents include gefitinib, a small kinase inhibitor that
targets EGFR, and trastuzumab, which is a humanised
antibody that targets ERBB2 receptors. Because the merits
of molecular targeting have been shown by clinical studies,
targeted treatment might also be applicable to treat
metastatic gestational trophoblastic neoplasms. In the
following sections, several molecular targets that are
expressed in gestational trophoblastic neoplasia tissues
and cell lines are discussed (table 1). However, there are
many other inhibitors that target emerging cancer-
associated molecules and new therapeutic approaches,
such as anti-human chorionic gonadotropin antibody
treatment and immunotherapy, might also prove useful
as new therapeutics in gestational trophoblastic neoplasia
in future.
c-MYC proto-oncogene
The activation and overexpression of the c-MYC proto-
oncogene is one of the most frequent molecular changes
in human cancers.50 In trophoblastic lesions, high levels
of c-MYC expression have been detected in chorio-
carcinomas and complete hydatidiform moles, and an
increased expression of c-MYC proteins might suppress
differentiation and enhance cellular proliferation in
gestational trophoblastic neoplasia.10 Direct evidence of
tumour dependence on c-MYC for growth and survival
comes from elegant studies showing that induction of
c-MYC expression results in tumour formation, and that
the blocking c-MYC activation generally leads to tumour
stasis or regression.51,52
In recent years, the targeting c-MYC oncoproteins has
been proposed as a new anticancer regimen.53,54 For
example, several strategies that aim to inactivate c-MYC
function and expression are being developed. These
approaches use: inhibitors that block c-MYC expression,
eg, triple-helix-forming oligonucleotides; molecules that
disrupt MYC-MAX interaction; agents that block the
functions of c-MYC-regulated genes, eg, cationic
porphyrins and ornithine decarboxylase; and antisense
646
Molecular targets Possible drugs
c-MYC Antisense oligonucleotides
EGFR Cetuximab, gefitinib, erlotinib
MAPK CI-1040, PD59089
MMP Marimastat
mTOR AP23573, RAD001, CCI-779
MAPK=mitogen-activated protein kinase. mTOR=mammalian target of
rapamycin.
Table 1: Molecular targets and examples of possible clinical drugs for
gestational trophoblastic neoplasia
oligonucleotides and siRNA that silence c-MYC
expression.54,55 Some of the antisense oligonucleotides
have successfully completed phase I clinical trials and
are at an advanced stage of drug development. Therefore,
future patients with metastatic gestational trophoblastic
neoplasia could benefit from anti-MYC agents in
combination with conventional treatment.
EGFR
EGFR is a transmembrane-receptor tyrosine kinase that
belongs to the ERBB-related kinase family. Many
epithelial tumours overexpress EGFR and harbour
somatic mutations, leading to constitutive activation of
the EGFR pathway.56 High levels of EGFR expression
have been detected in choriocarcinomas, PSTT, and
ETT.7,17,57 Additionally, the expression levels of EGFR in
choriocarcinomas and complete hydatidiform moles
were significantly higher than those of trophoblastic cells
in healthy placentas and partial hydratidiform moles.17
Further analysis showed that intense EGFR immuno-
reactivity in the extravillous (intermediate) trophoblastic
cells of a complete mole was seen to be significantly
correlated with the development of a persistent postmolar
gestational trophoblastic neoplasm.17 Expression of
ERBB-3 and ERBB-4, however, did not differ between
gestational trophoblastic neoplasms and healthy placental
tissues.17
The anti-EGFR regimen has become the prototype of
target-based treatment for cancer. Clinical and survival
benefits of anti-EGFR treatment have been shown in
patients with non-small-cell lung cancer and other
neoplastic diseases. The repertoire of cancer types that
might respond to anti-EGFR treatment is expanding,
including head and neck and pancreatic carcinomas.
Drugs that target EGFR, including cetuximab, gefitinib,
and erlotinib, have emerged as effective treatments
against various malignant diseases. Therefore, patients
with metastatic gestational trophoblastic neoplasia
might also benefit from anti-EGFR treatment in
combination with other anticancer agents.
Mitogen-activated protein kinase
Mitogen-activated protein kinase (MAPK) has a role in
signal transduction, and activation (phosphorylation) of
http://oncology.thelancet.com Vol 8 July 2007

MAPK results in the regulation of downstream effectors
to coordinate various cellular activities, including
proliferation, differentiation, apoptosis, angiogenesis,
and migration. Aberrant molecular changes that result
in the constitutive activation of MAPK contribute to
tumour development, with activation being caused by
either a genetic or epigenetic event. For example,
activating mutations in KRAS and BRAF, the upstream
genes of MAPK, have been found in various human
carcinomas.58,59 Additionally, MAPK activation can result
from growth-factor stimulation in the tumour micro-
environment.60
Kobel and co-workers61 studied MAPK activation in
PSTT tissues using immunohistochemistry. They
showed that the activated (phosphorylated) form of
MAPK was expressed in 84% of cases, but not in healthy
extravillous (intermediate) trophoblastic cells, the
normal counterparts of PSTT.61 However, because the
total concentration of MAPK was similar between PSTT
and healthy extravillous (intermediate) trophoblastic
cells, the investigators proposed that hyperactive MAPK
was not likely to be the result of MAPK-gene upreg-
ulation, but rather a result of other epigenetic events
that lead to its activation. To characterise the biological
role of MAPK activation in PSTT, they established the
first PSTT cell line, implantation-site trophoblastic cells,
in which the cells were highly motile and invasive in
vitro. Treatment with the mitogene inhibitors, CI-1040
and PD59089, which prevent activation of MAPK,
significantly reduced the motility and invasion of IST-2
cells in vitro.61 These findings suggest a functional role
of MAPK activation in the motility and invasion of PSTT
cells. Because CI-1040 and other mitogen inhibitors that
prevent MAPK activation have already been tested in
clinical trials,62–64 these inhibitors could also be assessed
for treating patients with metastatic PSTT who have not
responded to conventional treatment. Whether or not
inhibitors of the MAPK pathway can reduce the invasive-
ness and metastatic potential of PSTT would be intere-
sting to learn.
Mammalian target-of-rapamycin (mTOR)
mTOR is a member of the phosphatidyl inositol 3´ kinase
(PI3K) family and is one of the key molecules that
participate in the cellular signalling network. On
stimulation by miscellaneous growth signals, a cascade
of signalling occurs from PI3K, Akt, and mTOR to P70S6
kinase, resulting in a variety of cellular responses,
including cellular migration and proliferation. Osteo-
pontin, a cell-surface phosphorylated glycoprotein, repre-
sents one of the signals that trigger activation of the
PI3K/mTOR/P70S6 pathway and has been known to
have an important role during implantation by aiding
attachment of the trophectoderm of a blastocyst to the
surface epithelium of the endometrium.65 In gestational
trophoblastic neoplasia, osteopontin is highly expressed
in choriocarcinoma tissues and the derived cell lines.66,67
http://oncology.thelancet.com Vol 8 July 2007
Review
This glycoprotein is co-localised with an adhesion
molecule, called carcinoembryonic antigen-related cell
adhesion molecule 1 (CEACAM1), in the extravillous
(intermediate) trophoblast of a healthy placenta and also,
probably, in choriocarcinoma cells.68 The expression of
osteopontin has been shown to enhance the invasion of
CAECAM1-expressing trophoblastic cells.68 Likewise,
expression of highly phosphorylated osteopontin activates
the PI3K/mTOR/P70S6 kinase pathway and triggers
migration of CAECAM1-expressing trophoblastic cells in
choriocarcinoma cells grown in culture. Therefore, this
tumour phenotype could be suppressed by the
compounds LY294002 and rapamycin, which inhibit
PI3K and mTOR, respectively.66
Over the past years, several mTOR inhibitors have been
developed and tested in preclinical tumour models,69,70
and, more importantly, several of these inhibitors,
including AP23573, RAD001, and CCI-779, have been
tested in clinical trials as single agents or in combination
with other drugs in patients with solid or haematological
neoplastic diseases.71,72 Ongoing clinical studies involving
mTOR inhibitors will, hopefully, identify the efficacy of
these drugs in treating patients with cancer and provide
the foundation for future use of mTOR inhibitors in
patients with gestational trophoblastic neoplasia.
Matrix metalloproteinase (MMP)
As previously discussed, the upregulation of MMP and
downregulation of TIMP in choriocarcinomas suggest
that patients with metastatic, chemoresistant gestational
trophoblastic neoplasia might benefit from MMP
inhibitors. Marimastat is an MMP inhibitor and
preclinical studies in animal-tumour models have shown
that MMP inhibitors restrict the growth of solid tumours,
inhibit metastatic potential, and block tumour angio-
genesis. This inhibitor is an orally bioavailable drug that
has also been assessed in clinical trials of patients with
advanced non-small-cell lung cancer, relapsed prostate
cancer, glioblastoma multiforme, and several other types
of cancer. In general, modest toxic effects have been
reported, on the basis of several phase I clinical trials,73
and marimastat has been shown to delay progression in
patients with biochemically relapsed prostate cancer.74
Further studies need to be done to establish the efficacy
of marimastat and other new MMP inhibitors in treating
metastatic choriocarcinomas, and to find out if the
clinical outcome can be improved in patients with
choriocarcinoma who do not respond to conventional
chemotherapy.
Promises and challenges of target-based
treatment
The molecules discussed above have been shown to be
expressed in neoplastic trophoblastic cells and, thus, their
inhibitors represent attractive new therapeutics in
chemoresistant gestational trophoblastic neoplasia.
However, there are several challenges ahead before
647
 Review
molecular target-based treatment becomes a reality for
this group of diseases. First, the expression frequency of
the target genes in a large series of neoplastic tissues is
not yet known, because such clinical samples are rare for
study. Furthermore, studies of gestational trophoblastic
diseases have primarily focused on hydatidiform moles
and choriocarcinoma. PSTT and ETT have rarely been
studied so knowledge of these forms of gestational
trophoblastic neoplasia is limited. Second, it is crucial to
question whether the anticancer effects seen in preclinical
models using gestational trophoblastic neoplasia cell
lines can be translated to treatment efficacy in vivo.
Third, the clinical usefulness of applying molecular
inhibitors in gestational trophoblastic neoplasia depends
on careful correlation studies between clinical response
and the genetic and molecular changes of genes and the
pathways they control, respectively. For example,
mutational analysis of EGFR and members of the PI3K
and MAPK pathways would be important, in addition to
the assessment of their gene-expression levels. The
results of such genotype–phenotype correlation studies
would be essential for the development of tailored
treatment in order to keep potential side effects to a
minimum and optimise therapeutic effects. Fourth, given
the fact that gestational trophoblastic neoplasia is rare
and most patients can be cured by the current treatment,
recruitment of patients with this disease in clinical trials
can be challenging because only a limited number of
patients qualify and are available for such studies.
Nevertheless, drugs that inhibit the expression or function
of the molecular targets have been assessed in both
preclinical and clinical trials of various cancer types.
With results from ongoing studies soon to emerge and
new inhibitors in the development pipeline, future
patients with gestational trophoblastic neoplasia should
benefit from new waves of target-based treatment that
have been primarily developed to treat other types of
cancer. Effective multicentre collaborative networks
should be developed or better orchestrated to aid with
these clinical studies and to share valuable clinical
specimens, in order to characterise emerging targets
with therapeutic potential in gestational trophoblastic
neoplasia. With this continuous effort, we will be well-
positioned to further establish the pathogenesis of this
rare but intriguing disease and offer a better treatment
alternative for patients.
Conclusion
Recent advances in molecular studies of trophoblastic
cells in healthy placentas and in gestational trophoblastic
neoplasms have indicated that these neoplasms recap-
itulate normal trophoblast differentiation seen in early
developing placentas. Relating gestational trophoblastic
neoplasia with normal trophoblastic differentiation
programmes highlights the fundamental nature of this
group of diseases and provides the framework for future
molecular studies. These new findings not only shed
648
Search strategy and selection criteria
Published and unpublished data for this review were
identified by searching PubMed. Key words used to identify
papers included, “trophoblast”, “gestation”, “neoplasm”,
“metastasis”, “target”, “therapy”, “pathogenesis”,
“choriocarcinoma”, “placenta”, “mole”, and “epithelioid”. The
search strategy was not limited by year of publication, but
only articles published in English were selected.
light on the pathogenesis of gestational trophoblastic
neoplasia but also provide an array of candidate molecules
with therapeutic potential. Characterisation of the
molecular changes that have been targeted to treat several
other human cancers in gestational trophoblastic neo-
plasia would open an exciting avenue for target-based
treatment of metastatic forms of this disease that are
refractory to conventional treatment.
Conflicts of interest
The author declared no conflict of interest.
References
1 Genest DR, Berkowitz RS, Fisher RA. Gestational trophoblastic
disease. Lyon: IARC press, 2003.
2 Shih IeM, Kurman RJ. Molecular basis of gestational trophoblastic
diseases. Curr Mol Med 2002; 2: 1–12.
3 Shih IM, Kurman RJ. The pathology of intermediate trophoblastic
tumors and tumor-like lesions. Int J Gynecol Pathol 2001; 20: 31–47.
4 Soper JT. Gestational trophoblastic disease. Obstet Gynecol 2006;
108: 176–87.
5 Mao TL, Kurman RJ, Huang CC, Lin MC, Shih IM.
Immunohistochemistry of choriocarcinoma: an aid in differential
diagnosis and in elucidating pathogenesis. Am J Surg Pathol 2007;
(in press).
6 Shih IM, Seidman JD, Kurman RJ. Placental site nodule and
characterization of distinctive types of intermediate trophoblast.
Hum Pathol 1999; 30: 687–94.
7 Shih IM, Kurman RJ. Epithelioid trophoblastic tumor: a neoplasm
distinct from choriocarcinoma and placental site trophoblastic
tumor simulating carcinoma. Am J Surg Pathol 1998;
22: 1393–403.
8 Mazur MT. Metastatic gestational choriocarcinoma. Unusual
pathologic variant following therapy. Cancer 1989; 63: 1370–77.
9 Fulop V, Mok SC, Genest DR, Gati I, Doszpod J, Berkowitz RS. p53,
p21, Rb and mdm2 oncoproteins. Expression in normal placenta,
partial and complete mole, and choriocarcinoma. J Reprod Med
1998; 43: 119–27.
10 Fulop V, Mok SC, Genest DR, Szigetvari I, Cseh I, Berkowitz RS.
c-myc, c-erbB-2, c-fms and bcl-2 oncoproteins. Expression in
normal placenta, partial and complete mole, and choriocarcinoma.
J Reprod Med 1998; 43: 101–10.
11 Chen CA, Chen YH, Chen TM, et al. Infrequent mutation in tumor
suppressor gene p53 in gestational trophoblastic neoplasia.
Carcinogenesis 1994; 15: 2221–23.
12 Shi Y-F, Xie X, Zhao C-L, et al. Lack of mutation in tumour-
suppressor gene p53 in gestational trophoblastic tumours.
Br J Cancer 1996; 73: 1216–19.
13 Landers JE, Haines DS, Strauss JF 3rd, George DL. Enhanced
translation: a novel mechanism of mdm2 oncogene overexpression
identified in human tumor cells. Oncogene 1994; 9: 2745–50.
14 Wang H, Zeng X, Oliver P, et al. MDM2 oncogene as a target for
cancer therapy: an antisense approach. Int J Oncol 1999;
15: 653–60.
15 Asanoma K, Kato H, Inoue T, Matsuda T, Wake N. Analysis of a
candidate gene associated with growth suppression of
choriocarcinoma and differentiation of trophoblasts. J Reprod Med
2004; 49: 617–26.
http://oncology.thelancet.com Vol 8 July 2007

16 Asanoma K, Matsuda T, Kondo H, et al. NECC1, a candidate
choriocarcinoma suppressor gene that encodes a homeodomain
consensus motif. Genomics 2003; 81: 15–25.
17 Tuncer ZS, Vegh GL, Fulop V, Genest DR, Mok SC, Berkowitz RS.
Expression of epidermal growth factor receptor-related family
products in gestational trophoblastic diseases and normal placenta
and its relationship with development of postmolar tumor.
Gynecol Oncol 2000; 77: 389–93.
18 Fulop V, Colitti CV, Genest D, et al. DOC-2/hDab2, a candidate
tumor suppressor gene involved in the development of gestational
trophoblastic diseases. Oncogene 1998; 17: 419–24.
19 Fulop V, Mok SC, Berkowitz RS. Molecular biology of gestational
trophoblastic neoplasia: a review. J Reprod Med 2004; 49: 415–22.
20 Matsuda T, Sasaki M, Kato H, et al. Human chromosome 7 carries
a putative tumor suppressor gene(s) involved in choriocarcinoma.
Oncogene 1997; 15: 2773–81.
21 Burke B, Sebire NJ, Moss J, et al. Evaluation of deletions in 7q11.2
and 8p12-p21 as prognostic indicators of tumour development
following molar pregnancy. Gynecol Oncol 2006; 103: 642–48.
22 Stahle-Backdhal M, Inoue M, Zedenius J, et al. Decreased
expression of Ras GTPase activating protein in human trophoblastic
tumors. Am J Pathol 1995; 146: 1073–78.
23 Xue WC, Chan KY, Feng HC, et al. Promoter hypermethylation of
multiple genes in hydatidiform mole and choriocarcinoma.
J Mol Diagn 2004; 6: 326–34.
24 Vegh GL, Fulop V, Liu Y, et al. Differential gene expression pattern
between normal human trophoblast and choriocarcinoma cell lines:
downregulation of heat shock protein-27 in choriocarcinoma in
vitro and in vivo. Gynecol Oncol 1999; 75: 391–96.
25 Ciocca DR, Oesterreich S, Chamness GC, McGuire WL,
Fuqua SA. Biological and clinical implications of heat shock
protein 27,000 (Hsp27): a review. J Natl Cancer Inst 1993;
85: 1558–70.
26 Richards EH, Hickey E, Weber L, Master JR. Effect of
overexpression of the small heat shock protein HSP27 on the heat
and drug sensitivities of human testis tumor cells. Cancer Res 1996;
56: 2446–51.
27 Garrido C, Ottavi P, Fromentin A, et al. HSP27 as a mediator of
confluence-dependent resistance to cell death induced by anticancer
drugs. Cancer Res 1997; 57: 2661–67.
28 Shih IM. Application of human leukocyte antigen-G expression in
the diagnosis of human cancer. Hum Immunol 2007; 68: 272–76.
29 Singer G, Kurman RJ, McMaster M, Shih IeM. HLA-G
immunoreactivity is specific for intermediate trophoblast in
gestational trophoblastic disease and can serve as a useful marker
in differential diagnosis. Am J Surg Pathol 2002; 26: 914–20.
30 Vegh GL, Selcuk Tuncer Z, Fulop V, Genest DR, Mok SC,
Berkowitz RS. Matrix metalloproteinases and their inhibitors in
gestational trophoblastic diseases and normal placenta.
Gynecol Oncol 1999; 75: 248–53.
31 Palmieri C, Fisher RA, Sebire NJ, et al. Placental site trophoblastic
tumour arising from a partial hydatidiform mole. Lancet 2005;
366: 688.
32 Felemban AA, Bakri YN, Alkharif HA, Altuwaijri SM, Shalhoub J,
Berkowitz RS. Complete molar pregnancy. Clinical trends at King
Fahad Hospital, Riyadh, Kingdom of Saudi Arabia. J Reprod Med
1998; 43: 11–13.
33 Moore-Maxwell CA, Robboy SJ. Placental site trophoblastic tumor
arising from antecedent molar pregnancy. Gynecol Oncol 2004;
92: 708–12.
34 Oldt RJ 3rd, Kurman RJ, Shih IeM. Molecular genetic analysis of
placental site trophoblastic tumors and epithelioid trophoblastic
tumors confirms their trophoblastic origin. Am J Pathol 2002;
161: 1033–37.
35 Fisher RA, Paradinas FJ, Newlands ES, Boxer GM. Genetic evidence
that placental site trophoblastic tumours can originate from a
hydatidiform mole or a normal conceptus. Br J Cancer 1992;
65: 355–58.
36 Arima T, Imamura T, Sakuragi N, et al. Malignant trophoblastic
neoplasms with different modes of origin. Cancer Genet Cytogenet
1995; 85: 5–15.
37 Chang YL, Chang TC, Hsueh S, et al. Prognostic factors and
treatment for placental site trophoblastic tumor- report of 3 cases
and analysis of 88 cases. Gynecol Oncol 1999; 73: 216–22.
http://oncology.thelancet.com Vol 8 July 2007
Review
38 Shih IM, Kurman RJ. p63 expression is useful in the distinction of
epithelioid trophoblastic and placental site trophoblastic tumors by
profiling trophoblastic subpopulations. Am J Surg Pathol 2004;
28: 1177–83.
39 Shih IeM, Kurman RJ. Molecular basis of gestational trophoblastic
diseases. Curr Mol Med 2002; 2: 1–12.
40 Shih I-M, Mazur MT, Kurman RJ. Gestational trophoblastic disease.
In: Kurman RJ, ed. Blaustein’s pathology of the female genital tract,
5th edn. New York: Springer-Verlag, 2002: 1193–247.
41 Shih IM, Kurman RJ. Placental site trophoblastic tumor--past as
prologue. Gynecol Oncol 2001; 82: 413–14.
42 Ichikawa N, Zhai YL, Shiozawa T, et al. Immunohistochemical
analysis of cell cycle regulatory gene products in normal trophoblast
and placental site trophoblastic tumor. Int J Gynecol Pathol 1998;
17: 235–40.
43 Mao TL, Seidman JD, Kurman RJ, Shih IeM. Cyclin E and p16
immunoreactivity in epithelioid trophoblastic tumor–an aid in
differential diagnosis. Am J Surg Pathol 2006; 30: 1105–10.
44 Murdoch S, Djuric U, Mazhar B, et al. Mutations in NALP7 cause
recurrent hydatidiform moles and reproductive wastage in humans.
Nat Genet 2006; 38: 300–02.
45 Chen H, Ye D, Xie X, Lu W, Zhu C, Chen X. PTEN promoter
methylation and protein expression in normal early placentas and
hydatidiform moles. J Soc Gynecol Investig 2005; 12: 214–17.
46 Cagayan MS, Lu-Lasala LR. Management of gestational
trophoblastic neoplasia with metastasis to the central nervous
system: a 12-year review at the Philippine General Hospital.
J Reprod Med 2006; 51: 785–92.
47 Lurain JR, Singh DK, Schink JC. Primary treatment of metastatic
high-risk gestational trophoblastic neoplasia with EMA-CO
chemotherapy. J Reprod Med 2006; 51: 767–72.
48 Ngan HY, Tam KF, Lam KW, Chan KK. Relapsed gestational
trophoblastic neoplasia: a 20-year experience. J Reprod Med 2006;
51: 829–34.
49 Newlands ES. The management of recurrent and drug-resistant
gestational trophoblastic neoplasia (GTN). Best Pract Res Clin Obstet
Gynaecol 2003; 17: 905–23.
50 Dang CV, O’Donnell KA, Zeller KI, Nguyen T, Osthus RC, Li F.
The c-Myc target gene network. Semin Cancer Biol 2006;
16: 253–64.
51 Jonkers J, Berns A. Oncogene addiction: sometimes a temporary
slavery. Cancer Cell 2004; 6: 535–38.
52 Yu D, Dews M, Park A, Tobias JW, Thomas-Tikhonenko A.
Inactivation of Myc in murine two-hit B lymphomas causes
dormancy with elevated levels of interleukin 10 receptor and
CD20: implications for adjuvant therapies. Cancer Res 2005;
65: 5454–61.
53 Vita M, Henriksson M. The Myc oncoprotein as a therapeutic target
for human cancer. Semin Cancer Biol 2006; 16: 318–30.
54 Ponzielli R, Katz S, Barsyte-Lovejoy D, Penn LZ. Cancer
therapeutics: targeting the dark side of Myc. Eur J Cancer 2005;
41: 2485–501.
55 Nilsson JA, Keller UB, Baudino TA, et al. Targeting ornithine
decarboxylase in Myc-induced lymphomagenesis prevents tumor
formation. Cancer Cell 2005; 7: 433–44.
56 Mendelsohn J, Baselga J. Epidermal growth factor receptor
targeting in cancer. Semin Oncol 2006; 33: 369–85.
57 Muller-Hocker J, Obernitz N, Johannes A, Lohrs U. P53 gene
product and EGF-receptor are highly expressed in placental site
trophoblastic tumor. Hum Pathol 1997; 28: 1302–06.
58 Singer G, Oldt R 3rd, Cohen Y, et al. Mutations in BRAF and KRAS
characterize the development of low-grade ovarian serous
carcinoma. J Natl Cancer Inst 2003; 95: 484–86.
59 Cohen Y, Xing M, Mambo E, et al. BRAF mutation in papillary
thyroid carcinoma. J Natl Cancer Inst 2003; 95: 625–27.
60 Satyamoorthy K, Li G, Gerrero MR, et al. Constitutive mitogen-
activated protein kinase activation in melanoma is mediated by both
BRAF mutations and autocrine growth factor stimulation.
Cancer Res 2003; 63: 756–59.
61 Kobel M, Pohl G, Schmitt WD, Hauptmann S, Wang TL, Shih IeM.
Activation of mitogen-activated protein kinase is required for
migration and invasion of placental site trophoblastic tumor.
Am J Pathol 2005; 167: 879–85.
649
 Review
62 Rinehart J, Adjei AA, Lorusso PM, et al. Multicenter phase II study
of the oral MEK inhibitor, CI-1040, in patients with advanced non-
small-cell lung, breast, colon, and pancreatic cancer. J Clin Oncol
2004; 22: 4456–62.
63 Lorusso PM, Adjei AA, Varterasian M, et al. Phase I and
pharmacodynamic study of the oral MEK inhibitor CI-1040 in
patients with advanced malignancies. J Clin Oncol 2005;
23: 5281–93.
64 Thompson N, Lyons J. Recent progress in targeting the Raf/MEK/
ERK pathway with inhibitors in cancer drug discovery.
Curr Opin Pharmacol 2005; 5: 350–56.
65 Martin PM, Sutherland AE. Exogenous amino acids regulate
trophectoderm differentiation in the mouse blastocyst through an
mTOR-dependent pathway. Dev Biol 2001; 240: 182–93.
66 Al-Shami R, Sorensen ES, Ek-Rylander B, Andersson G, Carson
DD, Farach-Carson MC. Phosphorylated osteopontin promotes
migration of human choriocarcinoma cells via a p70 S6 kinase-
dependent pathway. J Cell Biochem 2005; 94: 1218–33.
67 Briese J, Oberndorfer M, Schulte HM, Loning T, Bamberger AM.
Osteopontin expression in gestational trophoblastic diseases:
correlation with expression of the adhesion molecule, CEACAM1.
Int J Gynecol Pathol 2005; 24: 271–76.
68 Briese J, Oberndorfer M, Patschenik C, et al. Osteopontin is
colocalized with the adhesion molecule CEACAM1 in the
extravillous trophoblast of the human placenta and enhances
invasion of CEACAM1-expressing placental cells. J Clin Endocrinol
Metab 2005; 90: 5407–13.
650
69 Lane HA, Lebwohl D. Future directions in the treatment of
hormone-sensitive advanced breast cancer: the RAD001
(Everolimus)-letrozole clinical program. Semin Oncol 2006;
33 (2 suppl 7): S18–25.
70 Nathan CA, Amirghahari N, Rong X, et al. Mammalian target of
rapamycin inhibitors as possible adjuvant therapy for microscopic
residual disease in head and neck squamous cell cancer. Cancer Res
2007; 67: 2160–68.
71 Smolewski P. Recent developments in targeting the mammalian
target of rapamycin (mTOR) kinase pathway. Anticancer Drugs 2006;
17: 487–94.
72 Okuno S. Mammalian target of rapamycin inhibitors in sarcomas.
Curr Opin Oncol 2006; 18: 360–62.
73 Goffin JR, Anderson IC, Supko JG, et al. Phase I trial of the matrix
metalloproteinase inhibitor marimastat combined with carboplatin
and paclitaxel in patients with advanced non-small cell lung cancer.
Clin Cancer Res 2005; 11: 3417–24.
74 Rosenbaum E, Zahurak M, Sinibaldi V, et al. Marimastat in the
treatment of patients with biochemically relapsed prostate cancer:
a prospective randomized, double-blind, phase I/II trial. Clin Cancer
Res 2005; 11: 4437–43.
http://oncology.thelancet.com Vol 8 July 2007