Genome-Wide Cell-Free DNA-Based Prenatal Testing for Rare Autosomal Trisomies and Subchromosomal Abnormalities

Rare Autosomal Trisomy Mosaicism during Pregnancy and at Birth

Chromosomal mosaicism is the presence of two or more distinct cell lines in an individual . In a prenatal setting, chromosomal mosaicism most commonly affects only the placenta (confined placental mosaicism; CPM), but may occasionally extend to the fetus (true fetal mosaicism; TFM). The clinical consequences of chromosomal mosaicism identified during prenatal diagnosis can be difficult to predict, ranging from no apparent phenotypic effect to early fetal lethality. In the absence of fetal anomalies, the outcome of TFM in a prenatal setting remains uncertain .

Autosomal trisomy is a common cause of early miscarriage. Of the 10%–15% of pregnancies that end in clinical miscarriage, about half will do so because of a chromosome abnormality, and of these, the majority will involve an autosomal trisomy . With very rare exceptions, only trisomy for chromosomes 13, 18, and 21 (the so-called live birth trisomies) is compatible with survival to term, recognizing that even these conditions are associated with a high rate of miscarriage and stillbirth . All other autosomal trisomies (the so-called rare autosomal trisomies) are lethal in nonmosaic form, notwithstanding occasional reports of survival into the second, and very rarely the third trimester of pregnancy; stillbirth or neonatal death is expected. A large study reporting on the prevalence and types of rare chromosome abnormalities notified to 16 European congenital anomaly registers recorded 58 nonmosaic rare trisomies from 2.3 million births (0.25 per 10,000), none of whom survived . All were notified following prenatal testing or late fetal death (≥ 20 weeks of pregnancy). In contrast, 141 mosaic rare trisomies were reported (0.6 per 10,000 births), of which 78% were identified prenatally. Of these, 41% were liveborn, 7% stillborn, and 49% resulted in pregnancy terminations associated with fetal anomalies. Mosaicism involving trisomies 8 and 9 were most commonly notified. These findings show that true mosaicism for rare autosomal trisomies contributes to a small but significant part of pre- and perinatal adverse pregnancy outcomes.

Rare Autosomal Trisomies in Amniotic Fluid

Amniocentesis for cytogenetic prenatal diagnosis has been in widespread use since the early 1970s. The cells isolated from amniotic fluid closely reflect the chromosome constitution of the fetus, being derived from sources such as the fetal skin, nasopharyngeal tract, and urogenital tract , with extraembryonic cells being contributed from the amniotic membrane (amnion) . Historically, amniocentesis samples used for conventional chromosome analysis have been divided and grown across several independent culture dishes. Specimens that exhibit the same mosaic chromosome abnormality in at least two culture dishes are considered to exhibit true (Level III) mosaicism, which is present in approximately 0.1%–0.3% of amniocentesis samples analyzed by conventional karyotyping . True mosaicism most commonly involves the autosomal trisomies (48%), followed by sex chromosome aneuploidies (40%) and extra structurally abnormal chromosomes (12%) . Confirmation of mosaicism in fetal or newborn samples occurs in about 60%–70% of cases and in one US collaborative study, approximately 38% of autosomal trisomy mosaics were reported to be associated with noticeable phenotypic abnormalities .

Hsu et al. have reported phenotypic outcome data for 151 rare autosomal trisomy mosaics (excluding trisomy 20) ascertained following amniocentesis . This series was recently updated by Wallerstein et al., who reported summary outcomes for all mosaic autosomal trisomies , including 506 cases of rare trisomy mosaicism ( Table 1 ). Cases with prior abnormal ultrasound findings were excluded to help remove ascertainment bias. With regard to recorded abnormal outcomes, mosaic trisomy for chromosomes 2, 9, 16, 20⁎ [⁎see below], and 22 were classified as very high risk (> 60% with abnormal outcomes); chromosomes 5, 14, and 15 were classified as high risk (40%–59% abnormal); chromosomes 7, 12, and 17 as moderately high risk (20%–39%); chromosomes 6 and 8 as moderate risk (up to 19%); and no rare mosaic trisomies were classified as low risk (0%). Mosaic trisomy for chromosomes 1 and 10 was not observed. Mosaic trisomy for chromosomes 3, 4, 11, and 19 was not assigned a risk in the previous study by Hsu et al. due to insufficient cases ( n < 5). In the Wallerstein et al. series, abnormal outcomes were recorded in 3/4 cases of trisomy 3 mosaicism, 3/5 trisomy 4, 0/4 trisomy 11, and 0/1 trisomy 19. Therefore true mosaicism for trisomies 3 and 4 suggests a high to very high risk, based on these small numbers. Trisomy 20 mosaicism⁎ appears to have been misclassified as very high risk, rather than moderate risk (11% of cases with abnormal outcome), which is consistent with the lower frequency of abnormal outcomes in an earlier study .

The assessment of abnormal outcomes was made mostly by postmortem examination following termination or after birth. The authors note that subtle anomalies may not have been recognized and that neurodevelopmental follow-up after birth was rare. Nonetheless, these summary data are invaluable for helping evaluate possible outcomes following a diagnosis of rare trisomy mosaicism with normal ultrasound and to help guide patient counseling. Genetic counseling in the setting of normal fetal ultrasound remains problematic, but the presence of ultrasound anomalies indicates a very high risk for developmental and physical disabilities following the detection of rare trisomy mosaicism .

One last consideration is the introduction of chromosome microarray (CMA) into prenatal diagnosis. Few data currently exist on the interpretation of chromosomal mosaicism found in uncultured amniotic fluid samples ascertained using CMA. In the State of Victoria, Australia, > 85% of all samples received for cytogenetic prenatal diagnosis are now analyzed using CMA ; the majority of which use DNA extracted directly from uncultured cells. In my own laboratory, using a single-nucleotide polymorphism (SNP) CMA (Illumina Inc.), which has a lower limit of detection of 7%–12% for trisomy mosaicism , we sometimes observe discrepancies between the results of CMA on uncultured cells, and the results of conventional karyotyping on cultured cells. In some, but not all cases of discrepancy, the SNP CMA will exhibit mosaicism, while the cultured cells are karyotypically normal, or perhaps show only a single colony of abnormal cells. Insufficient data currently exist to determine which method more accurately reflects the true fetal karyotype or provides a better prediction of phenotype.

Confined Placental Mosaicism in Chorionic Villi and its Relationship to CELL-FREE DNA-Based NIPT

Chorionic villus sampling (CVS) for cytogenetic prenatal diagnosis emerged in the mid-1980s . The procedure enables prenatal testing from the first trimester of pregnancy, using samples of chorionic villi biopsied from the placenta at around 10 to 12 weeks of gestational age. The basis of CVS is that the karyotype of the placental chorionic villi represents the karyotype of the fetus .

Samples of chorionic villi for conventional chromosome analysis can be prepared using two methods: (i) a direct or short-term (24–48 h) culture method (STC) that analyzes rapidly dividing cells from an outer villi layer of cytotrophoblast, and (ii) a long-term culture method (LTC) that analyzes cells grown from the mesodermal core of the chorionic villi. Over the past two decades, some laboratories have substituted STC for other rapid methods of analysis, using techniques such as fluorescence in situ hybridization (FISH) or quantitative fluorescence polymerase chain reaction (QF-PCR) . Whereas STC provides a low-resolution G-banded karyotype, FISH and QF-PCR typically only target aneuploidy for chromosomes 13, 18, 21, X, and Y. More recently, CMA using DNA extracted from whole chorionic villi has replaced LTC in some laboratories .

Reports of discrepancies between the karyotype of cells from STC and/or LTC, and the chromosome constitution of fetus emerged shortly after CVS was implemented into clinical practice . These discrepancies usually involved chromosomal mosaicism that was present in the chorionic villi but not in the fetus—a phenomenon known as confined placental mosaicism (CPM) , which affects up to 2% of CVS samples . This frequency of mosaicism is at least 10 times higher than the rate of TFM seen after amniocentesis.

CPM can be present in STC only (CPM I), in LTC only (CPM II), or in both (CPM III) . Trisomy for CPM types I and II usually has a mitotic origin, where postzygotic gain of the trisomic chromosome is confined to the cytotrophoblast or the mesenchyme, respectively. CPM type III is more likely to have a meiotic origin and involve a trisomic conception that has lost one of the trisomic chromosomes in the first few cell divisions after fertilization—so-called trisomy rescue . Rarely, a false-negative result may be reported. Here, the chorionic villi have a normal karyotype, but the fetus has a chromosome abnormality, either as a full trisomy or with TFM. False-negative results for the common autosomal trisomies occur almost exclusively during analysis of STC cytotrophoblast cells . From a developmental view point, the cytotrophoblast cells are more distantly related to the embryo proper than are cells from the LTC mesenchyme; the mesenchyme cells being known to more accurately reflect the fetal karyotype . This is because the LTC mesenchyme cells derive from the hypoblast of the inner cell mass (ICM); the ICM also giving rise to the epiblast and embryo proper .

Discrepancies involving mosaicism in chorionic villi are critically important to our understanding and management of cfDNA test results, as the origin of “fetal” cfDNA is apoptotic cytotrophoblast cells. Thus cfDNA-based prenatal testing is analogous to CVS STC and is essentially a liquid biopsy of these placental cells. Several groups have used this knowledge to review large databases of CVS test results to predict the frequency of false-positive cases that will occur during cfDNA analysis and to help guide the choice of follow-up prenatal procedure. In particular, these large reviews are helpful for the interpretation and management of rare autosomal trisomy cases identified during cfDNA-based NIPT . A caveat here is that the cfDNA result is a proxy for the CVS STC result only. No information is provided on the LTC mesenchyme cells, which are available to aid interpretation during the analysis of a diagnostic CVS sample (see Table 2 ). Professional societies governing standards in cytogenetic testing recommend against the analysis of CVS STC alone, because of the increased chance of both false-positive and false-negative results . This recommendation is a salient reminder that cfDNA-based NIPT should always be regarded as a screening test.