Abnormal Villous Morphology Associated with Triple Trisomy of Paternal Origin

Abstract

The vast majority of trisomies in spontaneous abortions (SAB) are single and of maternal origin, most frequently due to meiosis I errors. Triple trisomies are exceedingly rare (∼0.05% of spontaneous abortions), most often of maternal origin, and associated with increased maternal age. Some trisomic SAB specimens can exhibit abnormal villous morphology simulating a partial hydatidiform mole, a distinct form of hydatidiform mole characterized by diandric triploidy. A SAB specimen from a 27-year-old woman, G1P0 at 8 weeks gestational age, was reviewed in consultation to address the finding of morphological features suggestive of a partial hydatidiform mole but DNA ploidy analysis yielding a diploid result. The villi were irregularly shaped and hydropic but lacked trophoblastic hyperplasia; p57 expression was retained. Since fully developed features of a partial hydatidiform mole were lacking, additional analysis was performed. Molecular genotyping and single nucleotide polymorphism array analysis demonstrated biparental diploidy with trisomy of chromosomes 7, 13, and 20, all of paternal origin. The three trisomies may have originated from paternal meiosis II errors, or from mitotic nondisjunction. We believe this to be the first report of triple trisomy in a SAB confirmed to be of paternal origin.

Aneuploidy is the leading cause of miscarriage.¹ It is estimated that approximately 50% of spontaneous abortions (SAB) are associated with trisomy, the vast majority of which involve a single chromosome.² Trisomy of two chromosomes, double trisomy, occurs in less than 3% of karyotyped SABs. Triple trisomies are exceedingly rare, occurring in ∼0.05% of karyotyped SABs.² Most trisomies originate during oogenesis, often from maternal meiosis I errors, correlating with an increased frequency of trisomies with increasing/advanced maternal age.¹

Some trisomic SAB specimens can exhibit abnormal villous morphology (AVM), including irregular villous shapes and mild trophoblastic hyperplasia, simulating a partial hydatidiform mole (PHM).³

PHM is a distinct form of hydatidiform mole characterized morphologically by irregularly shaped (“scalloped”) and variably sized villi ranging from hydropic to fibrotic, mild to moderate trophoblastic hyperplasia, and occasional trophoblastic inclusions, and genetically by paternally derived (diandric) triploidy.⁴,⁵,⁶,⁷ Distinction of PHMs from SABs with AVM is important for clinical management since the former have a low but known risk of persistent gestational trophoblastic disease (GTD) and are usually followed with serum human chorionic gonadotropin levels.⁸,⁹,¹⁰,¹¹,¹²,¹³,¹⁴ Distinguishing these entities on the basis of morphological features alone can be problematic, given the known poor interobserver reproducibility in the diagnosis of hydatidiform moles.¹⁵,¹⁶,¹⁷ One example of triple trisomy associated with hydatidiform changes of the chorionic villi has been reported, but the parental origin of the trisomy (49, XY, +8, +13, +18) was not determined.¹⁸ Herein we report, to our knowledge, the only molecularly confirmed example of a paternally derived triple trisomy with AVM suggestive of a PHM. Such a case raises the question of how PHMs should be defined at the genetic level.

Materials and Methods

P57 Immunohistochemistry

Formalin-fixed paraffin-embedded (FFPE) tissue sections (5-micron) were stained using a Ventana BenchMark XT automated immunostaining system (Tucson, AZ) with mouse monoclonal antibodies against p57 protein (Neomarkers, Fremont, CA). Validation and criteria for interpretation of immunohistochemical expression of p57 are reported in detail elsewhere.¹⁹

Molecular Genotyping by Short Tandem Repeat Markers

FFPE tissue sections (10-micron, unstained) were prepared. DNA was isolated from well-separated areas of maternal decidua and villous tissue, and purified using QIAamp DNA spin columns (Qiagen, Valencia, CA). The AmpFlSTR Profiler and Identifiler kits (Applied Biosystems, Foster City, CA) were used for molecular genotyping, which included the amelogenin marker and the following short tandem repeats (STRs): TPOX (2p23-2pter), D2S1338 (2q35-37.1), D3S1358 (3p), FGA (4q28), CSF1PO (5q33.3-34), D5S818 (5q21-31), D7S820 (7q11.21-22), D7S820 (7q11.21-22), D8S1179 (8), THO1 (11p15.5), VWA (12p12-pter), D13S317 (13q22-31), D16S539 (16q24-qter), D19S433 (19q12-13.1), and D21(21q11.2-q21). PCR amplification of the STR loci and the amelogenin locus was performed, with thermal cycling conditions and capillary electrophoresis performed according to the manufacturer's instructions. The method and criteria for interpretation of data are reported in detail elsewhere.²⁰ In brief, allele ratios were calculated for each locus by dividing the peak height of the longer allele (allele to the right) by the peak height of the shorter allele (allele to the left). Allele ratios between 0.61 and 1.17 were considered consistent with two copies (diploidy), while ratios between 0.33 and 0.6, and 1.5 and 2.0 were considered to be consistent with three copies. The following additional STRs were also analyzed: D13S159, D13S171, D13S175, and D20S186 (Applied Biosystems). Reactions were individually prepared and thermal cycled according the manufacturer's instructions.

Genome-Wide Copy Number Analysis by Single Nucleotide Polymorphism Array

DNA was isolated from well-separated areas of maternal decidua and villous tissue on unstained tissue sections as described above. Samples were processed with an FFPE-optimized protocol based on the GeneChip Mapping 10 K Xba Assay Kit (Affymetrix, Santa Clara, CA), hybridized on GeneChip Mapping 10 K 2.0 arrays (Affymetrix), and the data analyzed using Affymetrix GeneChip Genotyping Analysis Software (GTYPE) 4.1 as described previously.²¹ Loss of heterozygosity and copy number estimates were obtained using a publicly available analysis package, Copy Number Analyzer for Affymetrix GeneChip arrays (CNAG 3.0).

Case History

The patient was a 27-year-old female, gravida 1 para 0, who underwent a spontaneous abortion at 8 weeks gestational age. A curettage specimen was submitted to the Johns Hopkins Gynecologic Pathology Consultation Service to address the finding of morphological features suggestive of a PHM, despite flow cytometric analysis performed on FFPE tissue by an outside laboratory demonstrating a diploid result. H&E-stained tissue sections revealed villi that were variably sized, irregularly shaped, and variably hydropic, with scalloped villous outlines and focal invagination of trophoblastic epithelium to form trophoblastic inclusions. Circumferential trophoblastic hyperplasia was essentially lacking, with only a rare villus exhibiting focal mild trophoblastic hyperplasia (Figure 1, A–D). An immunohistochemical stain for p57 demonstrated diffuse expression in the villous stromal cells and cytotrophoblast, a finding that excludes a diagnosis of a complete hydatidiform mole but does not distinguish PHMs from non-molar abortuses with AVM (Figure 1E). Since the morphology was abnormal but fully developed features of a PHM were lacking, additional molecular analyses were performed.

Figure 1.

Villi are variably sized, irregularly shaped, and variably hydropic (A). Focal marked villous scalloping creates focal trophoblastic inclusions (B). Circumferential trophoblastic hyperplasia is essentially lacking (A, C), with the exception of a rare villus with focal mild hyperplasia (D, lower right); other normal-appearing small villi with polarized trophoblast are present (D, left). E: p57 expression is diffusely present in villous stromal cells and the cytotrophoblast (expression in intermediate trophoblastic cells serves as internal positive control).

Results

Clinical genetic analysis of the decidua and villous tissue was performed using nine STR markers as described.²⁰ Two markers were not informative because they were homozygous for an allele shared with the decidua. Five of the STR loci yielded two alleles with patterns and allelic ratios consistent with bi-parental diploidy. However, two STR loci (D13S317 and D7S820) yielded two alleles with allelic ratios consistent with three copies (Figure 2). To further understand the genetic makeup of this SAB specimen, GeneChip Mapping 10 K 2.0 array analysis was performed. The array data demonstrated two copies and heterozygosity at all chromosomes except chromosomes 7, 13, and 20, which demonstrated results consistent with three copies (Figure 3, A–C). These results were consistent with biparental diploidy with trisomy of chromosomes 7, 13, and 20. Additional STR analysis confirmed the trisomy of chromosomes 7, 13, and 20, and demonstrated that all three trisomies were of paternal origin (Figure 2). STR markers on each of these three chromosomes demonstrated one paternal allele in double dosage compared with the maternal allele.

Figure 2.

Capillary electrophoresis data of five STR loci for decidua (D) and villous (V). x axis = size in bases; y axis = fluorescence intensity; r = ratio of peak height of longer allele (allele to the right) to peak height of shorter allele (allele to the left). The parental origin of the alleles in the villous tissue [maternal (M) or paternal (P)] is identified. Note that the STR loci on chromosomes 4 (FGA) and 12 (vWA) demonstrate one maternal and one paternal allele in approximately equal ratio, whereas the STR loci on chromosomes 7, 13, and 20 (D7S820, D13S317, and D20S186) demonstrate one maternal and one paternal allele with allelic ratios consistent with a double dosage of the paternal allele (see Materials and Methods for interpretation of the allelic ratios).

Figure 3.

Single nucleotide polymorphism array virtual karyotype for the villous tissue sample. The chromosomes are differentially colored and plotted along the x axis. The top plot (A) is the log2ratio of the test/normal smoothed over 50 single nucleotide polymorphisms. Only single nucleotide polymorphisms on fragments between 300 to 700 bp were used to generate the plot of the log2ratio. The green heterozygous call bars (B) are generated by the software whenever it obtains an AB call in the test sample. Copy number was calculated using the Hidden Markov Model (C) for copy number (dark blue = 0, light blue = 1, yellow = 2, pink = 3, hot pink = 4, and red = 5 or more). The array demonstrated two copies and heterozygosity at all chromosomes except for chromosomes 7, 13, and 20 which demonstrated three copies (noted with arrows).

Discussion

The majority of trisomies in SABs are of maternal origin, often originating from a maternal meiosis I error during oogenesis. Our finding of a SAB with three trisomies of paternal origin appears to be exceedingly rare. In fact, we believe this to be the first report of triple trisomy confirmed to be of paternal origin. Triple trisomies reported in the literature most frequently involve chromosomes 5, 8, 18, 20, 21, 22, and X, and most are associated with advanced maternal age.² In contrast, the current case involves a 27-year-old female with a SAB containing trisomy of chromosomes 7, 13, and 20. These differences may be a reflection of the parental origin of the trisomy of this case compared with those in the literature.

The presence of three trisomic chromosomes with a paternal allele in double dosage suggests that all three could have arisen from a paternal meiosis II error or a mitotic nondisjunction event. To definitively determine the mechanism by which the trisomies originated would require comparison of data from multiple loci from both parents and the villous specimen. Unfortunately the father is unavailable for testing.

The frequency of aneuploidy in sperm from fertile men is estimated to be 1 to 2%.²² Although nondisjunction during spermatogenesis has been reported to affect all chromosomes, the sex chromosomes and chromosomes 21 and 22 have a significantly increased frequency of aneuploidy.^22,23 Several studies now indicate that some infertile men have an increased frequency of sperm containing aneuploidy compared with fertile men.^22,23 Interestingly, in a study of infertile men, anueploidy was demonstrated for each of the 12 autosomes interrogated at an approximately equal frequency, suggesting a possible aberrant meiotic process in these males that affects all chromosomes.²⁴ One report described a healthy male who harbored a level of sperm aneuploidy greater than that typically seen in individuals treated with aneugenic therapies.²⁵ This karyotypic-normal male fathered four consecutive trisomic pregnancies, which were attributed to an inherent genome-wide defect in meiotic disjunction. Our finding of paternal trisomy of chromosomes 7, 13, and 20 in this case might suggest an inherent tendency for nondisjunction in our male individual also. The fertility history of this male is not available.

Spontaneous abortion specimens harboring one or more trisomies can exhibit AVM, resulting in some overlap with the morphological features of PHMs. These shared alterations can lead to misclassification of a trisomic AVM specimen as a PHM in the absence of rigorously applied ancillary molecular testing. Indeed, had this case been analyzed by fluorescence in situ hybridization or STR makers on only one or a combination of chromosomes 7, 13, and 20, the results would have been misinterpreted as being consistent with diandric triploidy, and the specimen would have been misclassified as a PHM. This highlights the limitations of using relatively few markers (fluorescence in situ hybridization or STR) to draw conclusions regarding the genetic composition of SAB specimens. Distinguishing between PHMs and non-molar SABs is important for clinical management. PHMs have a low, but real risk of GTD, estimated at 0.2 to 4%.^9,26,²⁷,²⁸ In contrast, the risk of GTD following a first trimester non-molar SAB is estimated to be ≤0.0002%.¹⁰ For this reason, a first trimester non-molar pregnancy does not inherently warrant follow-up with serum human chorionic gonadotrophin levels after spontaneous abortion or termination whereas patients with PHMs receive follow-up with serum hCG levels in conjunction with contraception until undetectable levels are obtained. These differences in risk of GTD and clinical management support the use of ancillary molecular testing to accurately distinguish non-molar SABs with AVM from PHMs. Avoiding misclassification of SABs, particularly those with AVM, as PHMs is of practical importance to patients with infertility for whom mandated contraception is undesirable. In addition, accurate diagnosis with molecular techniques should be required in current and future investigational studies to assure refined ascertainment of the true risk of persistent GTD associated with these specific entities.

The finding that three diandric trisomies in an otherwise biparental diploid specimen can impart some morphological features suggesting a PHM raises the question of how much excess paternal genetic material is required to create a PHM. Previous reports have associated the single trisomy of chromosome 13 with placental morphology similar to that seen with PHM although the parental origin of the trisomy was not reported.²⁹,³⁰,³¹,³² Since maternally derived (digynic) triploidy is not associated with molar-type morphological alterations of the villi, it is possible that an imbalance of imprinted gene expression is responsible for the villous changes in PHMs. Both chromosomes 7 and 20 contain significant regions of imprinting, which may have contributed to altered gene expression and possibly to the morphological alterations suggestive of a PHM in this case.^33,34 It is possible that both the specific chromosomes involved and the dosage (number) of extra paternal chromosomes may act in concert to influence villous morphology since one study found that the parental origin of trisomy for only chromosome 7 did not affect the placental morphology.³⁵ However, in this study, only four chromosomes (2, 7, 15, and 22) were analyzed, making it possible that additional trisomies or other alterations in their cases affected placental morphology, thus calling into question their conclusions. Genome-wide analysis of a larger number of examples of both PHMs and SABs with AVM would be of interest to explore the genetic spectrum of these entities.