The genetics of gestational trophoblastic disease: a rare complication of pregnancy

Cancer Genetics 205 (2012) 63e77

REVIEW

The genetics of gestational trophoblastic disease: a rare complication of pregnancy a, *, Urvashi Surti a,b Lori Hoffner a

Magee-Womens Research Institute and Foundation, Pittsburgh, PA, USA; b Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA Gestational choriocarcinoma is usually a rapidly spreading fatal disease, but it is curable if diagnosed early and treated. It is a unique malignancy that is a partial or complete allograft with a genotype that is not the same as the host genotype. It is most often preceded by an abnormal molar pregnancy. The surprising and unique androgenetic origin of complete hydatidiform molar pregnancies was first revealed by Kajii and Ohama in 1977. We describe the current understanding of the morphology, epidemiology and genetics of gestational trophoblastic disease that followed the milestone findings by Kajii and Ohama. Keywords Choriocarcinoma, gestational trophoblastic disease, androgenetic hydatidiform mole, NLRP7, genomic imprinting ª 2012 Elsevier Inc. All rights reserved.

Gestational trophoblastic disease (GTD) encompasses a heterogeneous family of diseases, which includes hydatidiform moles, invasive mole, choriocarcinoma, placental site trophoblastic tumor, and epithelioid trophoblastic tumor. These proliferations arise from placental villous trophoblast and vary in propensity for local invasion and metastasis. The benign forms of GTD include complete and partial hydatidiform molar pregnancies. The malignant forms, which include invasive mole, choriocarcinoma, placental site trophoblastic tumor, and epithelioid trophoblastic tumor, can progress, metastasize, and lead to death if not treated. However, because of advances in medicine, most cases of malignant GTD can now be successfully treated and cured. Approximately 10% of complete hydatidiform moles transform into one of the malignant forms of GTD (1). And although it is true that most women with malignant GTD can expect to be successfully treated with modern chemotherapy, more than 10% of these patients in some parts of the world still die. This is usually due to delayed diagnosis and inadequate follow-up care (2). This spectrum of pregnancy-related disorders is unique and interesting. First, hydatidiform moles and gestational choriocarcinoma have rarely been recognized in any other species. Additionally, gestational trophoblastic tumors are naturally occurring allografts, arising not from the patient’s

Received October 11, 2011; received in revised form December 15, 2011; accepted January 10, 2012. * Corresponding author. E-mail address: [email protected] 2210-7762/$ - see front matter ª 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.cancergen.2012.01.004

own tissue but from a conceptus with a different genotype. It is thought that this genetic difference facilitates rejection of the tumor cells and contributes to their highly favorable response to cytotoxic chemotherapeutic drugs. Also, because all gestational trophoblastic tumors produce human chorionic gonadotropin (hCG), it can be used as a unique biochemical marker and aid in the early detection, diagnosis, and follow-up of these tumors and possible metastatic disease. Blood serum levels can be monitored for the presence of hCG (especially in postmolar cases), leading to early diagnosis and successful treatment. Untreated choriocarcinoma can rapidly spread and is fatal. This is exemplified by rare but unfortunate reports of dissemination of choriocarcinoma from organ transplant donors to recipients. Organs, including lung, kidney, and liver, from patients with undiagnosed choriocarcinoma have been inadvertently transplanted into recipients, in some cases tragically leading to rapid death of the recipients (3,4). Cases such as these emphasize the need for complete medical history of the donor, along with a thorough physical examination. Caution is needed when screening a female donor of childbearing age who has a history of menstrual irregularities, especially following a pregnancy or abortion. Appropriate screening tests such as b-hCG should be performed. The androgenetic origin of complete hydatidiform moles (CHMs) was first revealed by Kajii and Ohama in 1977, based on chromosomal polymorphisms (5). This was a very surprising finding at that time because androgenetic conceptions with exclusively paternal chromosomes were not described earlier in humans or any other animal species.

64 This result was rapidly confirmed by other groups (6e11). Elegant nuclear transplantation experiments performed in mice were reported in the 1980s that clearly described the fate of androgenetic as well as parthenogenetic diploid embryos and provided the proof that both the paternal and maternal genetic contributions are necessary for normal development (12e14). In this review, we describe the current understanding of the genetics of trophoblastic disease that followed the milestone findings by Kajii and Ohama.

Hydatidiform mole: complete and partial Pathology and genetics of hydatidiform moles Hydatidiform mole is the most common form of GTD, with an incidence of 1e3 out of every 1000 pregnancies (1). These abnormal conceptions have excessive placental development and little or no fetal development. Detailed morphology and genetic analysis in the late 1970s delineated two major types, namely, complete hydatidiform mole (CHM) and partial hydatidiform mole (PHM). These two differ clinically with distinctive pathologic and genetic features (15,16). In general, CHM is characterized by rapid hydatidiform change affecting the entire placenta. Histologically, there is marked villous hydrops and extensive circumferential villous trophoblastic hyperplasia, central villous cistern formation with compression of the villous stroma and random karyorrhexis, and little or no embryonic development. Genetically, CHMs are usually diploid and androgenetic in origin, with either a 46,XX or 46,XY karyotype (17). In our laboratory, we have collected data on 433 cases of CHMs since 1978. Of the androgenetic CHMs with cytogenetic analysis, we found 247 cases of diploid XX and 29 cases of diploid XY. In addition, there were three 92,XXXX cases, one 92,XXXY case, and one 92,XXYY case. Approximately 80% of androgenetic CHMs arise from the duplication of the haploid genome of a single sperm, whereas the other 20% appear to arise from dispermy, the fertilization of the egg by two sperm (18e20). Despite the androgenetic nature of the nuclear genome in CHM, the mitochondrial DNA has been shown to be maternally derived (21,22). The fate of the maternal genome in a CHM is still unclear. Originally, it was assumed that an error in meiosis II produced an anucleate egg, which was then fertilized by one or two sperm. More recently, arguments against the existence of “anucleate” eggs have resulted in an alternative theory, the postzygotic diploidization of a triploid conception (23). This theory originates with the formation of a diandric triploid conception usually by fertilization of a single oocyte (M) by two sperm (P1 and P2) or by a diploid sperm (P1P2). Triploidy is a relatively frequent fertilization error and occurs in approximately 1e2% of all human conceptions (23). The triploid zygote with three pronuclei (M, P1, and P2) can maintain the triploid state and produce a diandric triploid zygote (MP1P2) or the tripronuclear zygote could undergo abnormal cleavage resulting in 1n, 2n, and 3n derivatives. Some of these 1n and 2n cells may develop into hydatidiform moles or mosaicism involving an androgenetic cell line (Figure 2). Because the first zygotic division is orchestrated by the paternal centrosomes, cleavage errors often occur in cases of dispermic triploid zygotes due to the presence of two active centrioles in one

L. Hoffner, U. Surti ovum. During the first cleavage, the tripronuclear zygote can undergo abnormal division resulting in a 2n biparental derivative (MP1) and a 1n derivative (P2) that can undergo endoreduplication (P2P2). One sperm (P1) contributes to a biparental genome and the other sperm (P2) gives rise to a homozygous androgenetic mole. Abnormal division of the tripronuclear zygote could also result in the elimination of the maternal pronucleus (M) and the formation of a 2n heterozygous androgenetic mole (P1P2). Postzygotic diploidization of triploids provides a natural explanation for 2n homozygous and heterozygous androgenetic moles and may also explain unusual cases of chimerism, 2n/3n mosaicism, 2n/2n androgenetic/biparental conceptions, and 2n/2n molar/twin conceptions. In addition, this theory does not require the fertilization of an “empty egg”. Evidence for the natural occurrence of anuclear oocytes has never been observed. Examples of cases that could have originated from the diploidization of triploids and support this theory have been reported in several studies (24e30). However, not all features of the genetics of CHMs can be explained by the postzygotic diploidization of triploids theory, namely, the 4:1 frequency of homozygous versus heterozygous androgenetic CHMs. Partial hydatidiform moles (PHMs) demonstrate slow hydatidiform change, affecting only some of the placental villi. Microscopic examination of the villi shows focal trophoblastic hyperplasia, trophoblastic pseudoinclusions, and occasional cistern formation. Some fetal development is possible. The most common karyotype for PHM is triploid (69,XXX, 69,XXY, or 69,XYY), with the extra haploid genome being paternal in origin (8,9). The majority of these diandric triploid PHMs have been shown to arise from fertilization with two sperm, and less frequently by a diploid sperm (31). The existence of diploid PHMs is under debate, as some reported diploid PHMs may actually represent misdiagnosed diploid hydropic abortions, twin pregnancies, undetected mosaic cases, or early CHMs.

Diagnostic techniques for hydatidiform moles Improved ultrasound techniques have led to earlier termination of molar pregnancies, and their classic morphologic and histologic features may not be as evident earlier in gestation. Updated criteria for the histologic features of first-trimester moles are helpful (32); however, additional diagnostic techniques may be necessary to properly diagnose CHM versus PHM versus hydropic abortion. In addition, mosaic conceptions involving a diploid androgenetic cell line or diploid/triploid mosaic cases can pose even greater diagnostic confusion. Since CHMs and PHMs are genetically distinct, techniques that make use of those unique differences can be useful. Immunostaining techniques for maternally expressed genes such as p57KIP2 have proven to be valuable diagnostic tools for discriminating CHMs from other hydropic conceptions with a suspicious questionable molar phenotype, because most CHMs contain only paternal genes (33,34). In addition, p57KIP2 analysis can also detect mosaicism involving androgenetic and biparental cell lines and can further assess the distribution of these cells within the placental villus structure. However, p57KIP2 staining cannot distinguish PHMs from nonmolar hydropic abortions. Occasionally, p57KIP2 expression patterns can be erroneously interpreted because of a loss

The genetics of gestational trophoblastic disease of the maternal chromosome in a PHM (35,36) or because of retention of the maternal chromosome 11 in a CHM (37,38). Cytogenetic analysis, DNA flow cytometry, chromosome in situ hybridization (CISH), and fluorescence in situ hybridization (FISH) are useful in determining the ploidy of suspected molar tissues (34,39e41). Cytogenetic analysis by karyotype provides an accurate picture of the chromosomes and can also identify structural chromosomal abnormalities; however, it requires fresh tissue for cell culture and analysis of metaphase spreads and can be time and labor intensive. Techniques such as DNA flow cytometry and FISH/CISH can be applied to fixed paraffin-embedded tissues, thus allowing for retrospective studies. FISH is accurate in determining ploidy and can also differentiate between XX and XY conceptions. FISH is also very useful in detecting mosaicism as cell-to-cell analysis is possible. However, these ploidy techniques lack the ability to discern the exact maternal and paternal chromosomal contributions and therefore are not able to differentiate between diploid androgenetic moles and diploid hydropic abortions. Molecular genetic techniques that make use of the unique DNA polymorphisms in each individual can reliably determine the presence or absence of a maternal contribution and therefore identify the parental origin of the fetal genome. These techniques, such as microsatellite analysis, have proven very useful in the differential diagnosis of hydropic abortions, CHMs, and PHMs (30,40e44). However, some molecular results can be misleading, especially in mosaic cases; a case involving two diploid cell lines sharing a paternal genome could appear, on analysis, to be a triploid. The differential diagnosis of hydatidiform moles from nonmolar entities and the accurate classification of suspected molar conceptions is often difficult, as some of the classical morphological features may overlap or be difficult to recognize, especially in earlier gestations. This has led to interobserver diagnostic variability and occasional misdiagnosis. Accurate diagnosis is important both for clinical purposes and for investigational studies. The diagnosis of CHM or the detection of androgenetic cells within products of conception both carry a significant risk to the patient for developing persistent GTD. Therefore, for all products of conception exhibiting morphology consistent with a hydatidiform mole, we suggest p57KIP2 immunostaining. If the p57KIP2 results are negative and correlate with the morphology, a diagnosis of CHM can be made. If the p57KIP2 results are positive or equivocal (i.e., discordant results between cytotrophoblast and villous stromal cells), we suggest using another ancillary technique. This may depend on which techniques a laboratory routinely performs or has the capability to perform. Laboratories lacking access to certain techniques can consult with a reference lab that offers additional testing. In some very unusual cases, accurate diagnosis may depend on correlating the results of several ancillary tests; however, this may be necessary to accurately assess the risk of persistent gestational disease and for appropriate clinical management.

Epidemiology of hydatidiform moles Molar pregnancies can occur at any reproductive age; however, the risk increases significantly at the extremes of the reproductive age range. Females under 16 years of age

65 have a six times greater risk than women aged 16e40 years, and women over the age of 50 have a one-in-three chance of having a CHM (1). In our laboratory, of the 433 CHM cases that we have collected since 1978, we have age information on 365 cases. The patients with CHM ranged in age from 12 years to 53 years: 75 cases aged 12e19 years (31 of these were
16 y); 163 cases aged 20e29 years; 107 cases aged 30e39 years; and 20 cases aged 40e53 years (6 cases were 50 y). Most molar pregnancies still occur in the patients’ mid-20s and 30s, reflecting that most pregnancies in general occur in this age range. The reported incidence of molar pregnancy varies in different regions of the world. In Japan, the incidence is 2 per 1,000 pregnancies, which is three times greater than the incidence in North America and Europe (0.6e1.1 per 1,000) (45). A recent report on the incidence of CHM in Morocco demonstrated an incidence of 4.3 per 1,000 (46), and during the years 2002e08 at the University of PhilippinesePhilippine General Hospital, the prevalence rate of hydatidiform mole was 14 per 1,000 pregnancies (47). These variations may result from differences in reporting, and in addition, the high incidence historically reported in some areas may also be related to socioeconomic, nutritional, or genetic factors. The epidemiologic characteristics of CHM and PHM differ significantly. There does not appear to be a consistent association between maternal age and PHM. PHM appears to be more common in women with a history of irregular menses and the use of oral contraceptives for more that 4 years (45). The incidence of triploid conceptions is very similar worldwide. A healthy co-twin can develop alongside a CHM or PHM and occurs in 1 per 20,000e100,000 pregnancies (48). A diploid CHM and normal co-twin can be mistakenly reported as a molar pregnancy with a fetus and therefore given the diagnosis of diploid PHM. The trophoblast from the CHM in the twin pregnancy can pose a significant increased risk for persistent GTD, so an accurate diagnosis is critical.

Unusual cases of hydatidiform mole Although it is true that most hydatidiform moles are either androgenetic diploids or diandric triploids, there are some cases of moles that are genetically unusual. Tetraploid cells in a diploid CHM are frequently identified, especially in the extravillous trophoblast. Interphase FISH analysis on paraffin sections is useful in evaluating ploidy in cytotrophoblast, mesenchymal stroma, and extravillous trophoblast. This low level tetraploidy is relatively common in most placentas, including nonmolar conceptions. Rarely, however, complete tetraploidy has been documented in CHM, and triploid CHMs have also been reported. In most of these cases, the chromosomes are still all paternal in origin (49). Rare tetraploid PHMs have also been reported, often resulting from three paternal and one maternal genome, confirming the association of the molar phenotype with the excess of the paternal genome (20,49e51). Other numerical and structural abnormalities have been detected in CHMs including monosomies, trisomies, and tetrasomiesdall occurring in an androgenetic background. Reported abnormalities include monosomy X, trisomy 8, trisomy 10, trisomy 17, trisomy 21, trisomy 20 (34), trisomy 13 (52), and an additional derivative 7 (30). Our

66 unpublished data include additional cases of trisomy for chromosomes 5, 10, 20, and 21; monosomy for 21; and tetrasomy for 14. We also observed one case that was 50, XX,þX,þ5,þ7,þ21. There are a few rare reports of CHMs retaining a chromosome of maternal origin leading to trisomy. In one case, the CHM retained a maternal chromosome 11, and in another case, the CHM retained maternal chromosomes 6 and 11 (37,38). Some hydropic placentas with suspected molar characteristics have been shown to contain two genetically distinct cell lines and represent mosaic or chimeric conceptions. Several of these mosaic/chimeric conceptions with hydropic placental changes have been shown to result from the presence of two different diploid cell lines, one androgenetic and one biparental. We have identified at least seven cases where both biparental and androgenetic cells are admixed in the same placenta (27,34,35,53, and unpublished data). Cases of mosaicism/chimerism with an androgenetic and a biparental cell population have been observed in hydatidiform moles (24e26,30,34,53,54), in placentas displaying placental mesenchymal dysplasia (27e29,55e57), and in fetus/children with malformations or growth abnormalities mimicking the BeckwitheWiedemann phenotype (55e63). Placental mesenchymal dysplasia is the term used for rare cases in which marked hydropic swelling and cistern formation are present, but minimal or no trophoblastic hyperplasia is observed. Many of these cases were subsequently found to be androgenetic/biparental mosaic conceptions. The phenotype seems to correlate with the distribution of the androgenetic cells. The androgenetic cells can be confined to the placenta (confined placental mosaicism) or may be present in the fetus as well. In the cases presenting with malformations in a fetus/child, the androgenetic cells were observed in the fetus/child, and in placental mesenchymal dysplasia, the androgenetic cells were observed predominantly in the placental vessels, chorion, and mesenchymal cells. We have also noted that the distribution of the androgenetic cells in the placenta of a CHM may affect the histologic phenotype of the placental villi. In one CHM case where the villous stomal cells were biparental and the cytotrophoblast contained the androgenetic cells, the trophoblast hyperplasia was marked. This was in contrast to another mosaic CHM case where the stromal cells were androgenetic and the cytotrophoblast cells were biparental. This latter case resulted in very minimal trophoblastic hyperplasia (35). Clinically, it is important to recognize the presence of androgenetic cells because of their increased propensity for progressing to persistent gestational trophoblastic disease (PGTD). Recently, there have been reports of PGTD resulting from androgenetic/biparental chimeric and mosaic conceptions (30,53). Occasionally, a mosaic triploid cell line can be confined to the placenta while the associated fetus has a normal diploid biparental karyotype (64e67). We recently analyzed one unusual mosaic case that involved a diploid androgenetic cell line confined to the villous cytotrophoblast and a triploid biparental cell line confined to the mesenchymal stromal cells (unpublished data). A number of cases of CHM and placental mesenchymal dysplasia with biparental/androgenetic mosaicism or chimerism have now been described. Possibly, the crucial factor for what presents as a CHM and what presents as placental mesenchymal dysplasia is not the absolute presence or

L. Hoffner, U. Surti absence of androgenetic cells within certain parts of the placenta, but the relative frequency and location of androgenetic and biparental cells.

Diploid biparental CHM In 1982, Jacobs et al. reported on a series of CHMs and described a single case that was not androgenetic but was biparental in origin (67). Since then occasional diploid biparental moles have been reported (19,68,69). Occasionally these diploid biparental CHMs can be challenging to diagnose (70); most likely due to a more moderate level of trophoblastic proliferation that may be seen in biparental CHM (71). Diploid CHM of biparental origin represent a rare but unique class of hydatidiform mole. Most CHMs are sporadic and usually not recurrent. The risk of a subsequent molar pregnancy following a CHM is 1e6% (33,72,73). However, several familial and isolated cases of women with highly recurrent complete molar pregnancies have been reported (73e82). These CHMs are not androgenetic (AnCHM), but show normal biparental inheritance (BiCHM) (77,83,84) Cases of familial recurrent hydatidiform moles (FRHMs) are often from kindreds with evidence of consanguinity in the parents of the affected women, and they demonstrate an autosomal recessive inheritance pattern for the affected women (76,77,85). These rare, highly recurrent, familial CHMs that are pathologically indistinguishable from AnCHM, have a normal diploid biparental inheritance (BiCHM), but show abnormal patterns of expression and methylation status for imprinted genes (86e88). Studies have shown that a number of maternally imprinted genes in BiCHM have a paternal methylation pattern instead of a maternal methylation pattern (86,89e91). Thus, the molar phenotype may be the result of paternal methylation patterns on maternal chromosomes, resulting in a functional overexpression of the paternal genome. Whereas AnCHM and BiCHM differ genotypically, one being androgenetic and the other biparental, they share the same epigenotype due to aberrant methylation patterns in the germ line of women with BiCHM. This results in a shared phenotype, as both AnCHM and BiCHM display the cardinal feature of trophoblast proliferation and absence of fetal development. For women with an inherited predisposition for hydatidiform mole, the risk of a future CHM is 75% and other reproductive wastage is also common (92,93). Most of the molar gestations of the affected women are biparental CHMs, although they also have PHMs, spontaneous abortions, and stillbirths. It has also been speculated that in some cases, these various forms of reproductive wastage may share the same underlying etiology and are a continuous spectrum of the same condition (71). Classical genetic mapping studies demonstrated that almost all of the familial cases mapped to a gene-rich region of chromosome 19q13.4 (94). Linkage analysis, homozygosity mapping, and subsequent screening of candidate genes led to the identification of the NLRP7 gene, which has been shown to have a causal role in recurrent hydatidiform mole and possibly other reproductive wastage (95e97). NLRP7 is a cytoplasmic protein belonging to a group of proteins made up of an N-terminal pyrin (PYD) domain, a NACHT domain, and a C-terminal leucine-rich repeat

The genetics of gestational trophoblastic disease (LRR) domain. Little is known about the function of NLRP7; however, some members of the NLRP family have been implicated in the inflammatory processes and apoptosis (98). In an in vitro study, the wild-type NLRP7 protein acted as an inhibitor of interleukin-1b, a proinflammatory cytokine abundantly expressed in the female reproductive tract. Mutations in NLRP7 resulting in impaired cytokine secretion could impair the implantation and development of the embryo by altering inflammatory pathways in the uterus. The impaired inflammatory response of patients with NLRP7 mutations could make them tolerant to and delay the rejection of abnormal molar conceptions. Mutations in NLRP7 have been reported by various groups and in patients from several populations, supporting the evidence that it is a major gene for this condition (71,90e92,97,99e105). These mutations, which are listed on Infevers, an online database for autoinflammatory mutations (106), have been detected in more than 50 families with BiCHM and include deletions, insertions, duplications, and amino acid substitutions (90,91,93,97). Interestingly, many of the mutations cluster in the leucine-rich region of the protein, suggesting that this region may have a crucial functional role. Patients with FRHMs have two defective NLRP7 alleles that can be homozygous for the same mutation or a compound heterozygote for two different mutations. Two defective alleles are associated with BiCHM and more severe reproductive outcomes than one defective allele. To date, approximately 20 patients with a single defective allele in NLRP7 have been reported (71,99,100,104,105). These women with a single mutant allele have had reproductive loss due to triploid PHMs (100,104,107), spontaneous abortions (71,99,100), and stillbirths (71,99,100). In addition, two women with a single defective allele have also had AnCHMs (100,105). There has been no report of patients carrying two defective NLRP7 mutations who have had AnCHMs. However, there have been rare reports of recurrent cases of AnCHMs, and these cases have not been linked to NLRP7 mutations (103,108). Furthermore, not all cases of familial recurrent biparental CHMs have shown linkage to NLRP7 (82,109), so the roles of additional genes need to be evaluated. The exact function of NLRP7 in pregnancy and the mechanism by which mutations in NLRP7 result in the characteristic imprinting defects and abnormal development seen in CHM have yet to be determined; however, much work is currently being done to better understand the role of different NALP7 mutations in reproductive wastage. It appears that the protein-truncating NALP7 mutations are more severe than the missense mutations. In women with single missense mutations, a number of placental anomalies were found, all of which are known risk factors for perinatal morbidity. The spectrum of reproductive outcome in individuals with missense mutations must be further investigated and compared with that of individuals without any NALP7 mutations in a much larger population.

PGTD following hydatidiform mole When GTD is not cured by initial evacuation, it is referred to as persistent. PGTD occurs when the hydatidiform mole has grown from the surface layer of the uterus into the myometrium. Most patients with postmolar GTD will have

67 nonmetastatic molar proliferation or invasive moles, but gestational choriocarcinomas, placental site trophoblastic tumors, or epithelioid trophoblastic tumors can also develop. PGTD develops in patients usually following evacuation of CHMs and less frequently following other types of pregnancies. It is diagnosed when there is evidence of trophoblastic activity after the evacuation, as demonstrated by stationary or rising hCG levels. Approximately 10% of CHMs transform into one of the malignant forms of GTD (1); however, there have not been many indices that have proven useful in predicting which CHM will progress to PGTD. Essentially, the hCG regression pattern remains the most specific prognostic indicator (110,111). Whether the heterozygous CHM has a higher malignancy potential than its homozygous counterpart remains controversial (Figures 1A and 1B). Some studies have suggested that dispermic heterozygous moles (XX and XY) might have a more malignant potential than monospermic homozygous moles (112e114). A recent study found that the risk was two times greater for developing postmolar GTD when the causative pregnancy was a dispermic androgenetic mole compared with a monospermic androgenetic mole (115). Other studies have failed to support this theory (116e119). The risk of PGTD following a biparental CHM, in women with a maternal effect mutation in NLRP7, appears to be consistent with the risk following an androgenetic CHM (70) (Figure 1G). A recent report (115) on the incidence of postmolar GTD suggests that morphological features may be predictive in determining the risk of developing persistent trophoblastic disease following an androgenetic mole. This report suggests that the measurement of the shortest diameter of the largest hydropic villi may be used to estimate the risk of postmolar GTD. Cases with hydropic villi less than 2 mm in their shortest diameter did not develop postmolar GTD, whereas approximately 20 % of patients with hydropic villi greater than 2 mm in their shortest diameter had postmolar GTD. Of course, gestational age of the molar conception should be considered when using villus diameter measurements for risk assessment, because, in general, the villus size increases with gestational age. In general, the literature is conflicting regarding the risk of PGTD following PHM. On the basis of well-defined cases in the literature, the risk of PGTD after a PHM is estimated to be 0.2e5.6% (120e125). However, some studies have suggested that confirmed dispermic triploid PHMs simply do not develop PGTD (115,125e127) or it may be that the risk following a PHM is the same as the risk following a nonmolar conception (Figure 1C). Two studies provide well-defined cases of persistent disease following triploid PHMs. Seckl et al. describe three cases of PHM that transformed into choriocarcinoma, the most malignant form of GTD (120), and Cheung et al. describe two cases of PHM that metastasized to the lung (128). However, a review of the literature shows that, for many cases of PHM where PGTD followed, the histologically diagnosed PHM was diploid, and possibly a misdiagnosed CHM (115,119,125). In one study, of the patients classified as PHM, 5% developed PGTD; however, the genome was diploid in all of these “PHM” cases. None of the 131 patients in this study with triploid PHM encountered PGTD (125). It has been theorized that some cases of CHM with small hydropic villi may resemble PHM in terms of their morphological and histological features and, therefore, may

68

L. Hoffner, U. Surti

Figure 1 Genetic origins of molar conceptions. (A) Monospermic origin of CHM with duplication of paternal genome and loss of maternal genome. Resultant androgenetic conception is diploid XX and has a 1,000  greater risk of developing choriocarcinoma than a nonmolar conception. (B) Dispermic origin of CHM; two sperm fertilize an “empty” egg or two sperm fertilize a single egg resulting in a tri-pronuclear zygote and subsequent diploidization with loss of the maternal genome by abnormal division of the zygote. Resultant androgenetic conception can be diploid XX or XY and has a 1,000 greater risk of developing choriocarcinoma than a nonmolar conception. (C) Dispermic origin of PHM by fertilization of a single egg by two sperm. Resultant diandric triploid conception can be 69,XXX, 69,XXY, or 69,XYY. The potential increased risk for subsequent PGTD (including choriocarcinoma) is controversial. (D) Twin conception involving a normal biparental twin and a CHM. The CHM carries the same increased risk of developing PGTD (including choriocarcinoma) as described for (A). (Figure Continues)

The genetics of gestational trophoblastic disease

69

Figure 1 (Continued) (E) Androgenetic/biparental chimeric conception resulting from the fertilization of two eggs by two sperm. In one zygote, there is endoreduplication of the paternal pronucleus and loss of the maternal genome. The other zygote undergoes normal biparental fertilization. Following fusion of the two zygotes, the resultant chimeric conception can histologically appear as a CHM or a fetus with placental mesenchymal dysplasia, depending on the percentage and distribution of the androgenetic and biparental cell lines. The androgenetic cell line carries the same risk of PGTD as described in (A). (F) Androgenetic/biparental mosaic conception resulting from the fertilization of one egg by two sperm. Endoreduplication of one paternal pronucleus gives rise to the androgenetic cell line and the other paternal pronucleus contributes to the biparental cell line. The resultant mosaic conception can histologically appear as a CHM or a fetus with PMD, depending on the percentage and distribution of the androgenetic and biparental cell lines. The androgenetic cell line carries the same risk of PGTD as described in (A). (G) Biparental origin of a CHM in a female with a maternal-effect mutation of NLRP7. Because of aberrant methylation patterns, the resulting BiCHM has the same epigenotype as an androgenetic CHM and confers the same risk of PGTD (including choriocarcinoma). Blue color, paternal genome; pink color, maternal genome; purple color, biparental; PMD, placental mesenchymal dysplasia.

be incorrectly diagnosed as diploid PHM (125,128). Another potential explanation for triploid PHMs developing PGTD may be the finding that some morphologically appearing PHMs may in fact be a mosaic diploid androgenetic/diploid biparental conception. This could incorrectly suggest a triploid genotype by molecular analysis by showing the presence of more than two alleles. The androgenetic cell line in this conception could be precursor to the PGTD (Figures 1E and 1F). CHM and healthy co-twin pregnancies can also pose an increased risk of PGTD from the androgenetic molar pregnancy (129) (Figure 1D). The conclusion that PHM progresses to PGTD remains controversial. Many of the studies that provide a series of PHM cases with follow-up data do not provide convincing documentation that the cases are accurately classified as PHM; however, there are a few well-documented cases of

persistent trophoblastic disease originating from triploid PHM (120,128). We do caution that even in these studies, unrecognized mosaic androgenetic cell lines could be overlooked, and additional cases are needed. These rare cases should be analyzed with all means possible to ensure complete accuracy. Because of the uncertainty of the progression of PHM to persistent disease, routine blood tests to monitor hCG levels should be performed following the evacuation of any molar pregnancy, complete or partial, to prevent undetected persistent disease.

Gestational trophoblastic tumors Gestational trophoblastic tumors (GTTs) include the persistent forms of GTD, namely invasive mole, choriocarcinoma,

70

L. Hoffner, U. Surti

Figure 2 Genetic origins of homozygous and heterozygous complete moles following the postzygotic diploidization of a tripronuclear zygote. The triploid zygote with three pronuclei (M, P1, and P2) can undergo abnormal cleavage resulting in 1n and 2n derivatives. Some of these 1n and 2n cells may develop into hydatidiform moles or mosaicism involving an androgenetic cell line. In the first scenario, one of the male pronuclei (P1) of the tripronuclear zygote undergoes replication prior to the first zygotic cleavage. During the second cleavage, one of the P1 genomes along with the maternal genome (M) gives rise to a biparental cell line, whereas the other P1 genome along with the P2 genome produces a heterozygous androgenetic complete mole [Examples can be found in Robinson et al., 2007 (29); and Sunde et al., 2011 (30)]. In the second scenario, abnormal division of the tripronuclear zygote can result in the elimination of the maternal pronucleus (M) and the formation of a 2n heterozygous androgenetic complete mole (P1P2). In the third scenario, the tripronuclear zygote undergoes abnormal division resulting in a 2n biparental derivative (MP1) and a 1n derivative (P2) that can undergo endoreduplication (P2P2), giving rise to a homozygous androgenetic complete mole [Examples can be found in Ford et al., 1986 (24); Weaver et al., 2000 (25); Surti et al., 2005 (27); Robinson et al., 2007 (29); Sunde et al., 2011 (30)]. The fourth scenario starts with monospermic fertilization and precocious division of the male pronucleus. This results in a “temporary” tripronuclear zygote that can then undergo a zygotic division forming a 2n biparental derivative (MP1) and a 1n derivative (P1). A homozygous androgenetic complete mole can form following endoreduplication of this P1 genome [Examples of this can be found in Makrydimas et al., 2002 (26); Kaiser-Rogers et al., 2005 (28); and Sunde et al., 2011 (30)]. Blue color, paternal genome; pink color, maternal genome; purple color, biparental; M, maternal pronucleus; P1 and P2, male pronuclei.

placental site trophoblastic tumor, and epithelioid trophoblastic tumor. It is known that CHM is more likely to progress to GTT than any other type of pregnancy, with at least half of all GTTs originating from hydatidiform mole (130). In reality, this statistic is probably higher because some tumors arise not from the clinically antecedent pregnancy but from a previous, often unrecognized, CHM (131). Genetically, GTTs are poorly characterized because patients with these tumors are generally successfully treated with chemotherapy without surgical intervention, and tumor tissue is therefore rarely available for analysis. In addition, many patients are given prophylactic chemotherapy following a molar pregnancy thus preventing the occurrence of GTTs

entirely. This is common in the regions where patients fail to show up for the follow-up visits.

Invasive mole Invasive mole is a hydatidiform mole in which villous trophoblasts invade the myometrium or blood vessels or even metastasize to extrauterine sites. An invasive mole can develop from either a CHM or a PHM, but CHMs become invasive much more often than PHMs. In approximately 15% of cases, invasive moles can spread to tissues outside the uterus, most commonly to the lungs (132). The pathologic

The genetics of gestational trophoblastic disease diagnosis of invasive mole is rarely made because most cases are treated with chemotherapy a rarely by hysterectomy; however, histologic diagnosis of invasive mole is based on the proliferation of cytotrophoblast and syncytiotrophoblast cells and presence of villi scattered among the trophoblast. Cytogenetic studies on invasive moles have revealed a diploid karyotype in most cases. There also appears to be a high percentage of tetraploid cells (133). Wake et al. (113) published data on four invasive moles: three were 46,XX and one was 46,XY. Interestingly, they all appeared to be dispermic in origin. In one of these cases, the antecedent pregnancy was also analyzed and it was genetically identical to the invasive mole, confirming that invasive mole represents a molar pregnancy with aggressive trophoblast and deep invasion resulting often in persistent viable residual tissue and positive hCG.

Choriocarcinoma Choriocarcinoma is a highly malignant trophoblastic neoplasm that can occur after a pregnancy, as a component of a germ cell tumor, or in association with a poorly differentiated somatic carcinoma (134). Most commonly, choriocarcinoma is derived from pregnancies, including molar pregnancies, induced and spontaneous abortions, ectopic pregnancies and term or preterm deliveries. These are called gestational choriocarcinoma. Although gestational choriocarcinoma can follow any type of pregnancy, the overall risk is more than 1,000 times greater after a CHM than after a nonmolar conception, and greater than 50% of all gestational choriocarcinomas are preceded by a complete mole (135). Cases of gestational choriocarcinoma usually present with highly elevated levels hCG; histologically, they are diagnosed based on the presence of purely trophoblastic proliferation, and chorionic villi are absent. There are alternating areas of cytotrophoblast, syncytiotrophoblast, and intermediate trophoblast. Gestational choriocarcinoma can grow rapidly and metastasize widely. Occasionally, choriocarcinoma does not originate following a pregnancy. These cases of choriocarcinoma that are independent of pregnancy are called nongestational choriocarcinoma. These can be found in areas other than the uterus and can occur in both men and women (136). Although gestational and nongestational choriocarcinoma are pathologically and morphologically similar, they differ in genetic origin, immunogenicity, sensitivity to chemotherapy, and prognosis, with gestational choriocarcinoma having a better prognosis than nongestational choriocarcinoma (134). Genetically, nongestational choriocarcinomas are like other cancers that originate entirely from the patient and therefore, they have poor immunogenicity resulting in lowered sensitivity to chemotherapy. Gestational choriocarcinoma contains some or all paternal genetic material and is commonly considered to be a partial (originating from a biparental pregnancy) or a complete allograft (originating from an androgenetic molar pregnancy); therefore, it is more immunogenic and reacts well to chemotherapy, giving the patient a better prognosis. In cases of gestational choriocarcinoma, the prognosis of those secondary to molar pregnancy is much better than those secondary to term or nonmolar abortion (137). Therefore, it is important to

71 determine the causative pregnancy. Studies have demonstrated that the antecedent pregnancy is not always the causative pregnancy of gestational choriocarcinoma; therefore the causative pregnancy should be identified by it genetic origin (131). Cytogenetic analysis of choriocarcinoma tissue is very rare. Because of the vascular nature of the tumor, most cases are treated with chemotherapy and are not surgically removed; therefore, fresh tissue for genetic analysis is rarely available. Most of the genetic information on choriocarcinoma comes from older established cell lines and occasionally from primary tumor samples. The cytogenetic makeup of choriocarcinoma is largely determined by the causative pregnancy. Tumors that result from term pregnancies, nonmolar abortions, or PHMs will have both maternal and paternal chromosomes, whereas those derived from CHMs will be androgenetic in origin. Cytogenetic analyses of choriocarcinoma cell lines and tumor tissue usually show an aneuploid karyotype with modes in the hyperdiploid and hypotetraploid range and chromosomal alterations involving almost every chromosome. Although no consistent abnormality has been identified, karyotypic analyses of choriocarcinoma show a range of abnormalities, including chromosomal gains, losses, and rearrangements (138e145). These chromosomal gains, losses, and rearrangements can enhance the predisposition of trophoblasts to malignant transformation through activation of oncogenes and inactivation or loss of tumor suppressor genes. Recent molecular genetic studies have identified certain consistently altered chromosomal regions. The most significant are deletions of 7p12-q11.2 (146), amplification of 7q21-q31, and loss of 8p12-p21 (145). Allelic loss of 7p12-q11.2 and 8p12-p21 suggest the possible sites of tumor suppressor genes in these two regions that may be involved in the development of choriocarcinoma (145,147,148). However, the exact loci and identities of specific tumor suppressor genes have not been identified. One tumor suppressor gene that does appear to play a role in gestational trophoblastic tumors may be NECC1 (not expressed in choriocarcinoma clone 1) located on chromosome 4q11-q12 (149). It is abundantly expressed in normal placental villi and absent in all choriocarcinoma cell lines examined and most choriocarcinoma tissue samples. The amplification of 7q21-q31 observed in a series of choriocarcinoma suggests a role of oncogenes possibly located in this region. In a review article, Alifrangis and Seckl list the expression pattern of some tumor suppressor genes and oncogenes that may be upregulated or downregulated in the development of gestational trophoblastic neoplasia (150).

Placental-site trophoblastic tumor and epithelioid trophoblastic tumor Placental-site trophoblastic tumor (PSTT) is a rare form of PGTD. Patients with PSTT often present with amenorrhea or irregular bleeding months or possibly years after an antecedent pregnancy. Most follow a normal pregnancy or nonmolar pregnancy loss but can follow a complete or partial mole (151e155). PSTT grows more slowly, metastasizes later (months or years after origination), more commonly involves lymph nodes, and produces less hCG than choriocarcinoma (152,153). Patients with PSTT often have

72 persistent low and variable levels of serum hCG; however, serum blood analysis shows that hCG-free b-subunit is the predominant hCG form in PSTT and, therefore, is a reliable marker in diagnosis that can be used to discriminate PSTT from choriocarcinoma (156). Histologically, PSTT is composed predominantly of intermediate trophoblast, and chorionic villi are generally not seen. The trophoblastic tumor cells infiltrate the myometrium, and there is prominent vascular and lymphatic invasion. The majority of PSTTs are slow-growing malignant tumors, but approximately 10e15% behave in a clinically aggressive fashion (157). Genetic data on PSTT are very rare; however, of the few cases examined, comparative genomic hybridization and FISH studies have shown the majority to be diploid and occasionally tetraploid (158). Interestingly, in a genetic study of 20 cases of PSTT, all of the tumors were Y-chromosome negative (159). In tumors following nonmolar pregnancies, loss of heterozygosity was detected for chromosomes 7p11.2 and 8p12-21 (148). In another small study, two of the four cases that were analyzed identified chromosomal gains of chromosome 21q (160,161). Epithelioid trophoblastic tumor (ETT) is also a rare but distinctive gestational trophoblastic tumor (162). Microscopically, it is composed of intermediate trophoblasts. ETT grows in a more nodular fashion than the more infiltrating pattern seen in PSTT. Frequently, the tumor involves the uterine cervix (163); however, epithelioid trophoblastic tumors of the lungs have been reported following hydatidiform mole, invasive mole and term pregnancy (164). A recent study evaluated the chromosomal profile of five cases of ETT by comparative genomic hybridization. The analysis was successful in three cases and revealed a balanced chromosomal profile without detectable gains or losses (165).

Genomic imprinting in GTD A small number of genes are remarkable in that they are transcribed only from the maternally or paternally inherited allele, while the allele that is inherited from the other parent is imprinted or silent. This genomic imprinting seems to have an important relationship to the characteristic pathological features shared by CHM and PHM, namely, trophoblastic proliferation and abnormal or absent embryonic development. When only paternal chromosomes present, trophoblastic hyperplasia is increased, there is no fetal development, and a CHM develops. In PHM, the presence of a maternal genome results in more moderate trophoblastic hyperplasia and some fetal development. Moreover, digynic triploids, which have two maternal contributions and only one paternal, are not associated with a molar phenotype and have an abnormally small placenta. Thus, both overexpression of paternally transcribed genes and loss of maternally transcribed genes are likely to play a role in molar development. Several imprinted genes have been studied in CHM. Because of their lack of a maternal genome, CHMs fail to express imprinted genes that are normally transcribed from the maternally inherited allele. Both p57KIP2 and IPL, the products of the paternally imprinted maternally expressed genes CDKN1C (cyclin-dependent kinase inhibitor 1C) and IPL/TSSC3 (imprinted in placentas and liver) show high

L. Hoffner, U. Surti levels of expression in the villous cytotrophoblast of the normal human placenta but are absent from these cells in CHM (166e173). Whereas p57KIP2, but not IPL, is also expressed in the villous stroma of normal placenta, expression of p57KIP2 is lost from the villous stromal cells of CHMs. Thus, immunostaining for p57KIP2 can provide useful diagnostic information for discriminating CHM from PHM and other hydropic abortions (33,34). Dysregulation of the normal methylation patterns of imprinted genes may also play a role in the malignant potential of CHMs and tumor development (70,174). CHM’s high propensity to malignancy may be related to the abnormal expression of imprinted genes that occurs as a consequence of its androgenetic origin. This is further supported by the fact that women with diploid biparental CHMs display the same aberrant patterns of expression and methylation status for imprinted genes as androgenetic CHMs (33) and seem to carry the same risk of PGTD as androgenetic diploid CHMs. This indicates that it is not the double dosage of a recessive mutation in the paternal genome but unbalanced expression of imprinted genes that predisposes to malignant transformation (87). The inherent growth-promoting role of paternal genes, in the absence of growth-inhibiting maternal genes, may contribute to the malignant potential of CHMs, because the normal control mechanism over unrestrained cell growth is lost.

Future perspectives Genetic diagnosis has proven important in the management of patients with trophoblastic tumors. Distinctions among CHM, PHM, and hydropic abortion are clinically important because of the difference in the risk for PGTD. An algorithmic approach can be used to correctly diagnose them by using morphology and additional molecular techniques (44). The detection of unusual mosaic and chimeric cases of molar pregnancies is difficult, and the use of multiple technologies is essential to figure out the genetic contribution of each parent in multiple cell lines. Traditional DNA analysis is of limited use, but in situ analyses such as FISH and immunohistochemistry using pertinent antibodies such as p57KIP2 allow analysis of individual cells. A recent report of FISH using common deletion polymorphisms detected by analysis of DNA copy number variants promises to be very useful for this purpose (175). Such analysis will most likely confirm the suspicion that mosaicism is much more common than reported and can help us to understand the mechanism of origin of these fascinating conceptions. In addition, genetic analysis has proven very useful in discriminating between GTTs and nongestational trophoblastic tumors for which there is a very different prognosis. Genetic studies on gestational choriocarcinomas have identified a number of chromosome deletions and amplification hot spots. Further studies should be performed to identify putative tumor suppressor genes and oncogenes in such loci to enhance our understanding of the specific genes involved in the pathogenesis of GTTs and develop novel markers of malignancy in molar pregnancies. Genetic markers that could be used to identify hydatidiform moles that will progress to PGTD would greatly facilitate the appropriate management of women with molar pregnancies.

The genetics of gestational trophoblastic disease Research on the NLRP7 gene has shed light on a possible connection between impaired inflammatory response in some patients and recurrent molar pregnancies and other reproductive wastage. Initial studies focussing on NLRP7 mutations implicated homozygous mutations as having a causal role in biparental CHMs. However, after further investigation, it appears that NLRP7 may play a larger role, as women with heterozygous mutations also have experienced various forms of reproductive wastage. Further large-scale studies including patients with a complete spectrum of trophoblastic disease as well as reproductive failures and normal populations are required to understand the plasticity of various NALP7 mutations and variants in the presence or absence of a normal allele. In addition, cases of recurrent biparental complete moles that have not shown linkage to NLRP7 need to be further investigated for the probable involvement of other genes. Recently, Parry et al. analyzed 15 patients with recurrent biparental complete moles in which NLRP7 mutations had not been identified (176). They reported biallelic mutation of C6orf221 in three patients. C6orf221 is a member of a reproduction-related gene cluster (KHDC1/DPPA5/ C6orf221(ECAT1)/OOEO) located on human chromosome 6. Patients with biallelic mutations of C6orf221 and of NLRP7 have identical phenotypes, indicating that mutations of these two genes probably have similar causal roles in recurrent biparental complete moles. Future studies are needed to evaluate the roles of NLRP7, C6orf221, and possibly other genes in the origin of recurrent complete hydatidiform moles as well as to understand the role of these genes in conferring genetic susceptibility to other forms of reproductive wastage. Increased understanding of the biology of these genes and their role in normal oogenesis also promises to help patients with recurrent familial complete moles in achieving a normal pregnancy outcome as demonstrated by a recent report (177).

Acknowledgment The authors would like to thank Annah Hoffner for providing the illustration for Figures 1 and 2.

References 1. Sebire NJ, Seckl MJ. Gestational trophoblastic disease: current management of hydatidiform mole. BMJ 2008;337:1193. 2. Bower M, Newlands ES, Holden L, et al. EMA/CO for high-risk gestational trophoblastic tumors: results from a cohort of 272 patients. J Clin Oncol 1997;15:2636e2643. 3. Marsh JW Jr, Esquivel CO, Makowka L, et al. Accidental transplantation of malignant tumor from a donor to multiple recipients. Transplantation 1987;44:449e450. 4. Penn I. Transmission of cancer from organ donors. Nefrologia 1995;15:205e207. 5. Kajii T, Ohama K. Androgenetic origin of hydatidiform mole. Nature 1977;268:633e634. 6. Vassilakos P, Riotton G, Kajii T. Hydatidiform mole: two entities. A morphologic and cytogenetic study with some clinical consideration. Am J Obstet Gynecol 1977;127:167e170. 7. Wake N, Takagi N, Sasaki M. Androgenesis as a cause of hydatidiform mole. J Natl Cancer Inst 1978;60:51e57. 8. Jacobs PA, Hassold TJ, Matsuyama AM, et al. Chromosome constitution of gestational trophoblastic disease. Lancet 1978;2:49. 9. Lawler SD, Pickthall VJ, Fisher RA, et al. Genetic studies of complete and partial hydatidiform moles. Lancet 1979;2:580.

73 10. Jacobs PA, Wilson CM, Sprenkle JA, et al. Mechanism of origin of complete hydatidiform moles. Nature 1980;286:714e716. 11. Lawler SD, Povey S, Fisher RA, et al. Genetic studies on hydatidiform moles. II. The origin of complete moles. Ann Hum Genet 1982;46(Pt 3):209e222. 12. Surani MA, Barton SC, Norris ML. Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis. Nature 1984;308:548e550. 13. Barton SC, Surani MA, Norris ML. Role of paternal and maternal genomes in mouse development. Nature 1984;331:374e376. 14. McGrath J, Solter D. Completion of mouse embryogenesis requires both the maternal and paternal genomes. Cell 1984; 37:179e183. 15. Szulman A, Surti U. The syndromes of hydatidiform mole. I. Cytogenetic and morphological correlations. Am J Obstet Gynaecol 1978;13:665e671. 16. Szulman AE, Surti U. The syndromes of hydatidiform mole. II. Morphologic evolution of the complete and partial mole. Am J Obstet Gynecol 1978;132:20e27. 17. Surti U, Szulman AE, O’Brien S. Complete (classic) hydatidiform mole with 46, XY karyotype of paternal origin. Hum Genet 1979;51:153e155. 18. Fisher RA, Povey S, Jeffreys AJ, et al. Frequency of heterozygous complete hydatidiform moles, estimated by locusspecific minisatellite and Y chromosome-specific probes. Hum Genet 1989;82:259e263. 19. Kovacs BW, Shahbahrami B, Tast DE, et al. Molecular genetic analysis of complete hydatidiform moles. Cancer Genet Cytogenet 1991;54:143e152. 20. Lawler SD, Fisher RA, Dent J. A prospective genetic study of complete and partial hydatidiform moles. Am J Obstet Gynecol 1991;164:1270e1277. 21. Wallace DC, Surti U, Adams CW, et al. Moles have paternal chromosomes but maternal mitochondrial DNA. Hum Genet 1982;61:145e147. 22. Edwards YH, Jeremiah SJ, McMillan SL, et al. Complete hydatidiform moles combine maternal mitochondria with a paternal nuclear genome. Ann Hum Genet 1984;48(Pt 2):119e127. 23. Golubovsky MD. Postzygotic diploidization of triploids as a source of unusual cases of mosaicism, chimerism and twinning. Hum Reprod 2003;18:236e242. 24. Ford JH, Brown JK, Lew WY, et al. Diploid complete hydatidiform mole, mosaic for normally fertilized cells and androgenetic homozygous cells. Case report. Br J Obstet Gynaecol 1986;93: 1181e1186. 25. Weaver DT, Fisher RA, Newlands ES, et al. Amniotic tissue in complete hydatidiform moles can be androgenetic. J Pathol 2000;191:67e70. 26. Makrydimas G, Sebire NJ, Thornton SE, et al. Complete hydatidiform mole and normal live birth: a novel case of confined placental mosaicism: case report. Hum Reprod 2002; 17:2459e2463. 27. Surti U, Hill LM, Dunn J, et al. Twin pregnancy with a chimeric androgenetic and biparental placenta in one twin displaying placental mesenchymal dysplasia phenotype. Prenat Diagn 2005;25:1048e1056. 28. Kaiser-Rogers KA, McFadden DE, Livasy CA, et al. Androgenetic/biparental mosaicism causes placental mesenchymal dysplasia. J Med Genet 2006;43:187e192. 29. Robinson WP, Lauzon JL, Innes AM, et al. Origin and outcome of pregnancies affected by androgenetic/biparental chimerism. Hum Reprod 2007;22:1114e1122. 30. Sunde L, Niemann I, Hansen ES, et al. Mosaics and moles. Eur J Hum Genet 2011;19:1026e1031. 31. Zaragoza MV, Surti U, Redline RW, et al. Parental origin and phenotype of triploidy in spontaneous abortions: predominance of diandry and association with the partial hydatidiform mole. Am J Hum Genet 2000;66:1807e1820.

74 32. Sebire NJ, Fisher RA, Rees HC. Histopathological diagnosis of partial and complete hydatidiform mole in the first trimester of pregnancy. Pediatr Dev Pathol 2003;6:69e77. 33. Fisher RA, Hodges MD, Rees HC, et al. The maternally transcribed gene p57(KIP2) (CDNK1C) is abnormally expressed in both androgenetic and biparental complete hydatidiform moles. Hum Mol Genet 2002;11:3267e3272. 34. Hoffner L, Dunn J, Esposito N, et al. P57KIP2 immunostaining and molecular cytogenetics: combined approach aids in diagnosis of morphologically challenging cases with molar phenotype and in detecting androgenetic cell lines in mosaic/chimeric conceptions. Hum Pathol 2008;39:63e72. 35. Hoffner L, Parks WT, Swerdlow SH, et al. Simultaneous detection of imprinted gene expression (p57(KIP2)) and molecular cytogenetics (FICTION) in the evaluation of molar pregnancies. J Reprod Med 2010;55:219e228. 36. DeScipio C, Haley L, Beierl K, et al. Diandric triploid hydatidiform mole with loss of maternal chromosome 11. Am J Surg Pathol 2011 Oct;35(10):1586e1591. 37. Fisher RA, Nucci MR, Thaker HM, et al. Complete hydatidiform mole retaining a chromosome 11 of maternal origin: molecular genetic analysis of a case. Mod Pathol 2004;17:1155e1160. 38. McConnell TG, Norris-Kirby A, Hagenkord JM, et al. Complete hydatidiform mole with retained maternal chromosomes 6 and 11. Am J Surg Pathol 2009;33:1409e1415. 39. Sunde L, Mogensen B, Olsen S, et al. Flow cytometric DNA analyses of 105 fresh hydatidiform moles, with correlations to prognosis. Anal Cell Pathol 1996;12:99e114. 40. Genest DR, Dorfman DM, Castrillon DH. Ploidy and imprinting in hydatidiform moles. Complementary use of flow cytometry and immunohistochemistry of the imprinted gene product p57KIP2 to assist molar classification. J Reprod Med 2002;47: 342e346. 41. Lai CY, Chan KY, Khoo US, et al. Analysis of gestational trophoblastic disease by genotyping and chromosome in situ hybridization. Mod Pathol 2004;17:40e48. 42. Bifulco C, Johnson C, Hao L, et al. Genotypic analysis of hydatidiform mole: an accurate and practical method of diagnosis. Am J Surg Pathol 2008;32:445e451. 43. Murphy KM, McConnell TG, Hafez MJ, et al. Molecular genotyping of hydatidiform moles: analytic validation of a multiplex short tandem repeat assay. J Mol Diagn 2009;11:598e605. 44. Ronnett BM, DeScipio C, Murphy KM. Hydatidiform moles: ancillary techniques to refine diagnosis. Int J Gynecol Pathol 2011;30:101e116. 45. Berkowitz RS, Goldstein DP. Gestational trophoblastic disease: presentations from the XLVTH World Congress on Gestational Trophoblastic Diseases. J Reprod Med 2008;53:455e456. ^les hydati46. Boufettal H, Coulin P, Mahdaouia S, et al. Les mo  tes au Maroc: etude e pide miologique et clinique formes comple [Complete hydatiforme mole in Morocco: epidemiological and clinical study]. J Gyn Obst et Biologie Reprod 2011;40:419e429. 47. Cagayan MS. Changing trends in the management of gestational trophoblastic diseases in the Philippines. J Reprod Med 2010;55:267e272. 48. Seckl MJ, Sebire NJ, Berkowitz RS. Gestational trophoblastic disease. Lancet 2010;376:717e729. 49. Vejerslev LO, Dissing J, Hansen HE, et al. Hydatidiform mole: genetic origin in polyploid conceptuses. Hum Genet 1987;76: 11e19. 50. Sheppard DM, Fisher RA, Lawler SD, et al. Tetraploid conceptus with three paternal contributions. Hum Genet 1982; 62:371e374. 51. Surti U, Szulman AE, Wagner K, et al. Tetraploid partial hydatidiform moles: two cases with a triple paternal contribution and a 92, XXXY karyotype. Hum Genet 1986;72:15e21.  V, Chun K, Osborne R, et al. Androgenetic complete 52. Dube mole with trisomy 13: report of a case with microsatellite

L. Hoffner, U. Surti

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

genotyping and review of the literature. Pathol Res Pract 2010; 206:776e781. Surti U, Hoffner L, Kolthoff M, et al. Persistent gestational trophoblastic disease after an androgenetic/biparental fetal chimera: a case report and review. Int J Gynecol Pathol 2006; 25:366e372. Shiina H, Oka K, Okane M, et al. Coexisting true hermaphroditism and partial hydatidiform mole developing metastatic gestational trophoblastic tumors. A case report. Virchows Arch 2002;441:514e518. Reed RC, Beischel L, Schoof J, et al. Androgenetic/biparental mosaicism in an infant with hepatic mesenchymal hamartoma and placental mesenchymal dysplasia. Pediatr Dev Pathol 2008;11:377e383. H’mida D, Gribaa M, Yacoubi T, et al. Placental mesenchymal dysplasia with Beckwith-Wiedemann syndrome fetus in the context of biparental and androgenic cell lines. Placenta 2008; 29:454e460. Morales C, Soler A, Badenas C, et al. Reproductive consequences of genome-wide paternal uniparental disomy mosaicism: description of two cases with different mechanisms of origin and pregnancy outcomes. Fertil Steril 2009;92:393.e5e393.e9. Hoban PR, Heighway J, White GR, et al. Genome-wide loss of maternal alleles in a nephrogenic rest and Wilms’ tumour from a BWS patient. Hum Genet 1995;95:651e656. Bryke CR, Garber AT, Israel J. Evaluation of a complex phenotype in a unique patient with a paternal uniparental disomy for every chromosome cell line and a normal biparental inheritance cell line. In: Proceedings from the American Society of Human Genetics 54th Annual Meeting; October 26-30, 2004; Toronto, Canada. Abstract 823: 165. Giurgea I, Sanlaville D, Fournet JC, et al. Congenital hyperinsulinism and mosaic abnormalities of the ploidy. J Med Genet 2006;43:248e254. Wilson M, Peters G, Bennetts B, et al. The clinical phenotype of mosaicism for genomewide paternal uniparental disomy: two new reports. Am J Med Genet A 2008;146A:137e148. Romanelli V, Nevado J, Fraga M, et al. Constitutional mosaic genome-wide uniparental disomy due to diploidisation: an unusual cancer-predisposing mechanism. J Med Genet 2011; 48:212e216. Yamazawa K, Nakabayashi K, Matsuoka K, et al. Androgenetic/biparental mosaicism in a girl with Beckwith-Wiedemann syndrome-like and upd(14)pat-like phenotypes. J Hum Genet 2011;56:91e93. Sarno AP Jr, Moorman AJ, Kalousek DK. Partial molar pregnancy with fetal survival: an unusual example of confined placental mosaicism. Obstet Gynecol 1993;82:716e719. Ikeda Y, Jinno Y, Masuzaki H, et al. A partial hydatidiform mole with 2N/3N mosaicism identified by molecular analysis. J Assist Reprod Genet 1996;13:739e744. Zhang P, McGinniss MJ, Sawai S, et al. Diploid/triploid mosaic placenta with fetus. Towards A better understanding of ‘partial moles’. Early HumDev 2000;60:1e11. Jacobs PA, Hunt PA, Matsuura JS, et al. Complete and partial hydatidiform mole in Hawaii: cytogenetics, morphology and epidemiology. Br J Obstet Gynaecol 1982;89:258e266. Vejerslev LO, Fisher RA, Surti U, et al. Hydatidiform mole: cytogenetically unusual cases and their implications for the present classification. Am J Obstet Gynecol 1987;157:180e184. Ko TM, Hsieh CY, Ho HN, et al. Restriction fragment length polymorphism analysis to study the genetic origin of complete hydatidiform mole. Am J Obstet Gynecol 1991;164:901e906. Fisher RA, Hodges MD. Genomic imprinting in gestational trophoblastic diseaseea review. Placenta 2003;24: S111eS118. Messaed C, Chebaro W, Roberto RB, et al. NLRP7 in the spectrum of reproductive wastage: rare non-synonymous

The genetics of gestational trophoblastic disease

72.

73.

74.

75.

76.

77.

78.

79.

80. 81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

variants confer genetic susceptibility to recurrent reproductive wastage. J Med Genet 2011;48:540e548. Berkowitz RS, Im SS, Bernstein MR, et al. Gestational trophoblastic disease. Subsequent pregnancy outcome, including repeat molar pregnancy. J Reprod Med 1998 Jan;43:81e86. Horn LC, Kowalzik J, Bilek K, et al. Clinicopathologic characteristics and subsequent pregnancy outcome in 139 complete hydatidiform moles. Eur J Obstet Gynecol Reprod Biol 2006; 128:10e14. Ambani LM, Vaidya RA, Rao CS, et al. Familial occurrence of trophoblastic disease - report of recurrent molar pregnancies in sisters in three families. Clin Genet 1980;18:27e29. Parazzini F, La Vecchia C, Franceschi S, et al. Familial trophoblastic disease: case report. Am J Obstet Gynecol 1984; 149:382e383. Seoud M, Khalil A, Frangieh A, et al. Recurrent molar pregnancies in a family with extensive intermarriage: report of a family and review of the literature. Obstet Gynecol 1995;86:692e695. Helwani MN, Seoud M, Zahed L, et al. A familial case of recurrent hydatidiform molar pregnancies with biparental genomic contribution. Hum Genet 1999;105:112e115. Sensi A, Gualandi F, Pittalis MC, et al. Mole maker phenotype: possible narrowing of the candidate region. Eur J Hum Genet 2000;8:641e644. Fisher RA, Khatoon R, Paradinas FJ, et al. Repetitive complete hydatidiform mole can be biparental in origin and either male or female. Hum Reprod 2000;15:594e598. Fallahian M. Familial gestational trophoblastic disease. Placenta 2003;24:797e799. Agarwal P, Bagga R, Jain V, et al. Familial recurrent molar pregnancy: a case report. Acta Obstet Gynecol Scand 2004;83: 213e214. re JB, et al. Evidence of a genetic Slim R, Fallahian M, Rivie heterogeneity of familial hydatidiform moles. Placenta 2005;26: 5e9. Vejerslev L, Sunde L, Hansen BF, et al. Hydatidiform mole and fetus with normal karyotype: support of a separate entity. Obstet Gynecol 1991;77:868e874. Sunde L, Vejerslev LO, Jensen MP, et al. Genetic analysis of repeated, biparental, diploid, hydatidiform moles. Cancer Genet. Cytogenet 1993;66:16e22. Al-Hussaini TK, Abd el-Aal DM, Van den Veyver IB. Recurrent pregnancy loss due to familial and non-familial habitual molar pregnancy. Int J Gynaecol Obstet 2003 Nov;83:179e186. Judson H, Hayward BE, Sheridan E, et al. A global disorder of imprinting in the human female germ line. Nature 2002;416: 539e542. Van den Veyver IB, Al-Hussaini TK. Biparental hydatidiform moles: a maternal effect mutation affecting imprinting in the offspring. Hum Reprod Update 2006;12:233e242. Djuric U, El-Maarri O, Lamb B, et al. Familial molar tissues due to mutations in the inflammatory gene, NALP7, have normal postzygotic DNA methylation. Hum Genet 2006;120:390e395. El-Maarri O, Seoud M, Coullin P, et al. Maternal alleles acquiring paternal methylation patterns in biparental complete hydatidiform moles. Hum Mol Genet 2003;12:1405e1413. Kou YC, Shao L, Peng HH, et al. A recurrent intragenic genomic duplication, other novel mutations in NLRP7 and imprinting defects in recurrent biparental hydatidiform moles. Mol Hum Reprod 2008;14:33e40. Hayward BE, De Vos M, Talati N, et al. Genetic and epigenetic analysis of recurrent hydatidiform mole. Hum Mutat 2009;30: E629eE639. Fisher RA, Hodges MD, Newlands ES. Familial recurrent hydatidiform mole: a review. J Reprod Med 2004;49: 595e601. Wang CM, Dixon PH, Decordova S, et al. Identification of 13 novel NLRP7 mutations in 20 families with recurrent

75

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106. 107.

108.

109.

110.

111. 112.

113.

hydatidiform mole; missense mutations cluster in the leucinerich region. J Med Genet 2009;46:569e575. Moglabey YB, Kircheisen R, Seoud M, et al. Genetic mapping of a maternal locus responsible for familial hydatidiform moles. Hum Mol Genet 1999;8:667e671. Hodges MD, Rees HC, Seckl MJ, et al. Genetic refinement and physical mapping of a biparental complete hydatidiform mole locus on chromosome 19q13.4. J Med Genet 2003; 40:e95. Panichkul PC, Al-Hussaini TK, Sierra R, et al. Recurrent biparental hydatidiform mole: additional evidence for a 1.1-Mb locus in 19q13.4 and candidate gene analysis. J Soc Gynecol Investig 2005;12:376e383. Murdoch S, Djuric U, Mazhar B, et al. Mutations in NALP7 cause recurrent hydatidiform moles and reproductive wastage in humans. Nat Genet 2006;38:300e302. trilli V, et al. NALP inflammasomes: Martinon F, Gaide O, Pe a central role in innate immunity. Semin Immunopathol 2007; 29:213e229. Qian J, Deveault C, Bagga R, et al. Women heterozygous for NALP7/NLRP7 mutations are at risk for reproductive wastage: report of two novel mutations. Hum Mutat 2007;28: 741. Deveault C, Qian JH, Chebaro W, et al. NLRP7 mutations in women with diploid androgenetic and triploid moles: a proposed mechanism for mole formation. Hum Mol Genet 2009;18:888e897. Puechberty J, Rittore C, Philibert L, et al. Homozygous NLRP7 mutations in a Moroccan woman with recurrent reproductive failure. Clin Genet 2009;75:298e300. Slim R, Bagga R, Chebaro W, et al. A strong founder effect for two NLRP7 mutations in the Indian population: an intriguing observation. Clin Genet 2009 Sep;76:292e295. Buyukkurt S, Fisher RA, Vardar MA, et al. Heterogeneity in recurrent complete hydatidiform mole: presentation of two new Turkish families with different genetic characteristics. Placenta 2010;31:1023e1025. Muhlstein J, Golfier F, Rittore C, et al. The spectrum of NLRP7 mutations in French patients with recurrent hydatidiform mole. Eur J Obstet Gynecol Reprod Biol 2011;157:197e199. Qian J, Cheng Q, Murdoch S, et al. The genetics of recurrent hydatidiform moles in China: correlations between NLRP7 mutations, molar genotypes, and reproductive outcomes. Mol Hum Reprod 2011;17:612e619. Infevers. Available at: http://fmf.igh.cnrs.fr/ISSAID/infevers. Accessed June 9, 2011. Slim R, Ao A, Surti U, et al. Recurrent triploid and dispermic conceptions in patients with NLRP7 mutations. Placenta 2011; 32:409e412. van der Smagt JJ, Scheenjes E, Kremer JA, et al. Heterogeneity in the origin of recurrent complete hydatidiform moles: not all women with multiple molar pregnancies have biparental moles. BJOG 2006;113:725e728. Zhao J, Moss J, Sebire NJ, et al. Analysis of the chromosomal region 19q13.4 in two Chinese families with recurrent hydatidiform mole. Hum Reprod 2006;21:536e541. Bagshawe KD, Wilson H, Dublon P, et al. Follow-up after hydatidiform mole: studies using radioimmunoassay for urinary human chorionic gonadotrophin (HCG). J Obstet Gynaecol Br Commonw 1973;80:461e468. Bagshawe KD, Dent J, Webb J. Hydatidiform mole in England and Wales 1973-83. Lancet 1986;2:673e677. Kajii T, Kurashige H, Ohama K, et al. XY and XX complete moles: clinical and morphologic correlations. Am J Obstet Gynecol 1984;150:57e64. Wake N, Seki T, Fujita H, et al. Malignant potential of homozygous and heterozygous complete moles. Cancer Res 1984; 44:1226e1230.

76 114. Wake N, Fujino T, Hoshi S, et al. The propensity to malignancy of dispermic heterozygous moles. Placenta 1987;8:319e326. 115. Kaneki E, Kobayashi H, Hirakawa T, et al. Incidence of postmolar gestational trophoblastic disease in androgenetic moles and the morphological features associated with low risk postmolar gestational trophoblastic disease. Cancer Sci 2010;101: 1717e1721. 116. Fisher RA, Lawler SD. Heterozygous complete hydatidiform moles: do they have a worse prognosis than homozygous complete moles? Lancet 1984;2:51. 117. Mutter GL, Pomponio RJ, Berkowitz RS, et al. Sex chromosome composition of complete hydatidiform moles: relationship to metastasis. Am J Obstet Gynecol 1993;168:1547e1551. 118. Cheung AN, Sit AS, Chung LP, et al. Detection of heterozygous XY complete hydatidiform mole by chromosome in situ hybridization. Gynecol Oncol 1994;55:386e392. 119. van de Kaa CA, Schijf CP, de Wilde PC, et al. Persistent gestational trophoblastic disease: DNA image cytometry and interphase cytogenetics have limited predictive value. Mod Pathol 1996;9:1007e1014. 120. Seckl MJ, Fisher RA, Salerno GA, et al. Choriocarcinoma and partial hydatidiform moles. Lancet 2000;356:36e39. 121. Sebire NJ, Makrydimas G, Agnantis NJ, et al. Updated diagnostic criteria for partial and complete hydatidiform moles in early pregnancy. Anticancer Res 2003;23:1723e1728. 122. Feltmate CM, Growdon WB, Wolfberg AJ, et al. Clinical characteristics of persistent gestational trophoblastic neoplasia after partial hydatidiform molar pregnancy. J Reprod Med 2006; 51:902e906. 123. Hancock BW, Nazir K, Everard JE. Persistent gestational trophoblastic neoplasia after partial hydatidiform mole incidence and outcome. J Reprod Med 2006;51:764e766. 124. Wielsma S, Kerkmeijer L, Bekkers R, et al. Persistent trophoblast disease following partial molar pregnancy. Aust N Z J Obstet Gynaecol 2006;46:119e123. 125. Niemann I, Hansen ES, Sunde L. The risk of persistent trophoblastic disease after hydatidiform mole classified by morphology and ploidy. Gynecol Oncol 2007;104:411e415. 126. Genest DR, Ruiz RE, Weremowicz S, et al. Do nontriploid partial hydatidiform moles exist? A histologic and flow cytometric reevaluation of nontriploid specimens. J Reprod Med 2002;47:363e368. 127. Jeffers MD, O’Dwyer P, Curran B, et al. Partial hydatidiform mole: a common but underdiagnosed condition. A 3-year retrospective clinicopathological and DNA flow cytometric analysis. Int J Gynecol Pathol 1993;12:315e323. 128. Cheung AN, Khoo US, Lai CY, et al. Metastatic trophoblastic disease after an initial diagnosis of partial hydatidiform mole: genotyping and chromosome in situ hybridization analysis. Cancer 2004;100:1411e1417. 129. Sebire NJ, Foskett M, Paradinas FJ, et al. Outcome of twin pregnancies with complete hydatidiform mole and healthy cotwin. Lancet 2002;359:2165e2166. 130. World Health Organization Scientific Group. Gestational trophoblastic disease. World Health Organ Techl Rep Ser 1983;692:7e81. 131. Zhao J, Xiang Y, Wan XR, et al. Molecular genetic analyses of choriocarcinoma. Placenta 2009;30:816e820. 132. Lurain JR. Gestational trophoblastic disease II: classification and management of gestational trophoblastic neoplasia. Am J Obstet Gynecol 2011;204:11e18. 133. Yang YH, Kwak HM, Park TK, et al. Comparative cytogenetic and clinicopathologic studies on gestational trophoblastic neoplasia, especially hydatidiform mole. Yonsei Med J 1986; 27:250e260. 134. Cheung AN, Zhang HJ, Xue WC, et al. Pathogenesis of choriocarcinoma: clinical, genetic and stem cell perspectives. Future Oncol 2009;5:217e231.

L. Hoffner, U. Surti 135. Palmer JR. Advances in the epidemiology of gestational trophoblastic disease. J Reprod Med 1994;39:155e162. 136. Fisher RA, Newlands ES, Jeffreys AJ, et al. Gestational and nongestational trophoblastic tumors distinguished by DNA analysis. Cancer 1992;69:839e845. 137. Kohorn EI. Negotiating a staging and risk factor scoring system for gestational trophoblastic neoplasia. A progress report. J Reprod Med 2002;47:445e450. 138. Wake N, Tanaka K, Chapman V, et al. Chromosome and cellular origin of choriocarcinoma. Cancer Res 1981;41:3137e3143. 139. Sasaki S, Katayama PK, Roesler M, et al. Cytogenetic analysis of choriocarcinoma cell lines. Nippon Sanka Fujinka Gakkai Zasshi 1982;34:2253e2256. 140. Sheppard DM, Fisher RA, Lawler SD. Karyotypic analysis and chromosome polymorphisms in four choriocarcinoma cell lines. Cancer Genet Cytogenet 1985;16:251e258. 141. Surti U, Habibian R. Chromosomal rearrangement in choriocarcinoma cell lines. Cancer Genet Cytogenet 1989;38: 229e240. 142. Bettio D, Giardino D, Rizzi N, et al. Chromosomal abnormalities in choriocarcinomas of the female. Cancer Genet Cytogenet 1993;68:149e151. 143. Rodriguez E, Melamed J, Reuter V, et al. Cytogenetic abnormalities detected by direct analysis in a case of choriocarcinoma. Cancer Genet Cytogenet 1995;80:9e12. 144. Ishiwata I, Ishiwata C, Tsuiki A, et al. Establishment and characterization of a human gestational choriocarcinoma cell line (HOCC). Hum Cell 2004;17:33e41. 145. Ahmed MN, Kim K, Haddad B, et al. Comparative genomic hybridization studies in hydatidiform moles and choriocarcinoma: amplification of 7q21-q31 and loss of 8p12-p21 in choriocarcinoma. Cancer Genet Cytogenet 2000;116:10e15. 146. Matsuda T, Sasaki M, Kato H, et al. Human chromosome 7 carries a putative tumor suppressor gene(s) involved in choriocarcinoma. Oncogene 1997;15:2773e2781. 147. Pils D, Horak P, Gleiss A, et al. Five genes from chromosomal band 8p22 are significantly down-regulated in ovarian carcinoma: N33 and EFA6R have a potential impact on overall survival. Cancer 2005;104:2417e2429. 148. 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; 103:642e648. 149. Asanoma K, Kato H, Inoue T, et al. Analysis of a candidate gene associated with growth suppression of choriocarcinoma and differentiation of trophoblasts. J Reprod Med 2004;49:617e626. 150. Alifrangis C, Seckl MJ. Genetics of gestational trophoblastic neoplasia: an update for the clinician. Future Oncol 2010;6: 1915e1923. 151. Fisher RA, Paradinas FJ, Newlands ES, et al. Genetic evidence that placental site trophoblastic tumours can originate from a hydatidiform mole or a normal conceptus. Br J Cancer 1992;65:355e358. 152. Feltmate CM, Genest DR, Goldstein DP, et al. Advances in the understanding of placental site trophoblastic tumor. J Reprod Med 2002;47:337e341. 153. Papadopoulos AJ, Foskett M, Seckl MJ, et al. Twenty-five years’ clinical experience with placental site trophoblastic tumors. J Reprod Med 2002;47:460e464. 154. Palmieri C, Fisher RA, Sebire NJ, et al. Placental site trophoblastic tumour arising from a partial hydatidiform mole. Lancet 2005;366:688. 155. Kohorn EI. Long-term outcome of placental-site trophoblastic tumours. Lancet 2009;374:6e7. 156. Cole LA, Khanlian SA, Muller CY, et al. Gestational trophoblastic diseases: 3. Human chorionic gonadotropin-free betasubunit, a reliable marker of placental site trophoblastic tumors. Gynecol Oncol 2006;102:160e164.

The genetics of gestational trophoblastic disease 157. Wang D, He Y, Hu Y, et al. Placental site trophoblastic tumor with unusual presentation in the uterine cervix. Eur J Obstet Gynecol Reprod Biol 2010;148:100e101. 158. Remadi S, Lifschitz-Mercer B, Ben-Hur H, et al. Metastasizing placental site trophoblastic tumor: immunohistochemical and DNA analysis. 2 case reports and a review of the literature. Arch Gynecol Obstet 1997;259:97e103. 159. Hui P, Wang HL, Chu P, et al. Absence of Y chromosome in human placental site trophoblastic tumor. Mod Pathol 2007;20: 1055e1060. 160. Xue WC, Guan XY, Ngan HY, et al. Malignant placental site trophoblastic tumor: a cytogenetic study using comparative genomic hybridization and chromosome in situ hybridization. Cancer 2002;94:2288e2294. 161. Hui P, Riba A, Pejovic T, et al. Comparative genomic hybridization study of placental site trophoblastic tumour: a report of four cases. Mod Pathol 2004;17:248e251. 162. Shih IM, Kurman RJ. Epithelioid trophoblastic tumor: a neoplasm distinct from choriocarcinoma and placental site trophoblastic tumor simulating carcinoma. Am J Surg Pathol 1998;22:1393e1403. 163. 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;161:1033e1037. 164. Hamazaki S, Nakamoto S, Okino T, et al. Epithelioid trophoblastic tumor: morphological and immunohistochemical study of three lung lesions. Hum Pathol 1999;30:1321e1327. 165. Xu ML, Yang B, Carcangiu ML, et al. Epithelioid trophoblastic tumor: comparative genomic hybridization and diagnostic DNA genotyping. Mod Pathol 2009;22:232e238. 166. Chilosi M, Piazzola E, Lestani M, et al. Differential expression of p57KIP2, a maternally imprinted cdk inhibitor, in normal human placenta and gestational trophoblastic disease. Lab Invest 1998;78:269e276. 167. Castrillon DH, Sun D, Weremowicz S, Fisher RA, et al. Discrimination of complete hydatidiform mole from its mimics

77

168.

169.

170.

171.

172.

173.

174.

175.

176.

177.

by immunohistochemistry of the paternally imprinted gene product p57kip2. Am J Surg Pathol 2001;25:1225e1230. Fukunaga M. Immunohistochemical characterization of p57KIP2 expression in early hydatidiform moles. Hum Pathol 2002;33:1188e1192. Kihara M, Matsui H, Seki K, et al. Genetic origin and imprinting in hydatidiform moles. Comparison between DNA polymorphism analysis and immunoreactivity of p57KIP2. J Reprod Med 2005;50:307e312. Popiolek DA, Yee H, Mittal K, et al. Multiplex short tandem repeat DNA analysis confirms the accuracy of p57(KIP2) immunostaining in the diagnosis of complete hydatidiform mole. Hum Pathol 2006;37:1426e1434. Kato H, Wake N. Differential diagnosis between complete and partial mole using a TSSC3 antibody: correlation with DNA polymorphic marker analysis. J Reprod Med 2006;51: 861e867. Saxena A, Frank D, Panichkul P, et al. The product of the imprinted gene IPL marks human villous cytotrophoblast and is lost in complete hydatidiform mole. Placenta 2003;24: 835e842. Thaker HM, Berlin A, Tycko B, et al. Immunohistochemistry for the imprinted gene product IPL/PHLDA2 for facilitating the differential diagnosis of complete hydatidiform mole. J Reprod Med 2004;49:630e636. Li HW, Tsao SW, Cheung AN. Current understandings of the molecular genetics of gestational trophoblastic diseases. Placenta 2002;23:20e31. Chiang S, Fazlollahi L, Nguyen A, et al. Diagnosis of hydatidiform moles by polymorphic deletion probe fluorescence in situ hybridization. J Mol Diagn 2011;13:406e415. Parry DA, Logan CV, Hayward BE, et al. Mutations causing familial biparental hydatidiform mole implicate C6orf221 as a possible regulator of genomic imprinting in the human oocyte. Am J Hum Genet 2011;89:451e458. Fisher RA, Lavery SA, Carby A, et al. What a difference an egg makes. Lancet 2011;378:1974.