Papers in Press. Published November 5, 2013 as doi:10.1373/clinchem.2013.215145
The latest version is at http://hwmaint.clinchem.org/cgi/doi/10.1373/clinchem.2013.215145
Clinical Chemistry 60:1
000 – 000 (2014)
Molecular Diagnostics and Genetics
Maternal Mosaicism Is a Significant Contributor to
Discordant Sex Chromosomal Aneuploidies Associated with
Noninvasive Prenatal Testing
Yanglin Wang,1,2† Yan Chen,2† Feng Tian,3 Jianguang Zhang,3 Zhuo Song,3 Yi Wu,1 Xu Han,1 Wenjing Hu,1
Duan Ma,2 David Cram,3* and Weiwei Cheng2*
BACKGROUND: In the human fetus, sex chromosome
aneuploidies (SCAs) are as prevalent as the common
autosomal trisomies 21, 18, and 13. Currently, most
noninvasive prenatal tests (NIPTs) offer screening only
for chromosomes 21, 18, and 13, because the sensitivity
and specificity are markedly higher than for the sex
chromosomes. Limited studies suggest that the reduced accuracy associated with detecting SCAs is due
to confined placental, placental, or true fetal mosaicism. We hypothesized that an altered maternal karyotype may also be an important contributor to discordant SCA NIPT results.
METHODS:
We developed a rapid karyotyping method
that uses massively parallel sequencing to measure the
degree of chromosome mosaicism. The method was
validated with DNA models mimicking XXX and XO
mosaicism and then applied to maternal white blood
cell (WBC) DNA from patients with discordant SCA
NIPT results.
RESULTS:
Sequencing karyotyping detected chromosome X (ChrX) mosaicism as low as 5%, allowing an
accurate assignment of the maternal X karyotype. In a
prospective NIPT study, we showed that 16 (8.6%) of
181 positive SCAs were due to an abnormal maternal
ChrX karyotype that masked the true contribution of
the fetal ChrX DNA fraction.
CONCLUSIONS: The accuracy of NIPT for ChrX and
ChrY can be improved substantially by integrating the
results of maternal-plasma sequencing with those for
maternal-WBC sequencing. The relatively high frequency of maternal mosaicism warrants mandatory
WBC testing in both shotgun sequencing– and single-
1
Key Laboratory of Molecular Medicine, Ministry of Education, Shanghai Medical
College, Fudan University, Shanghai, China; 2 Prenatal Diagnosis Center, International Peace Maternity and Child Health Hospital, Shanghai Jiaotong University, Shanghai, China; 3 Berry Genomics, Beijing, China.
* Address correspondence to: D.C. at Berry Genomics, Beijing, Bldg. 9, Link Park,
6 Jingshun East St., Chaoyang District, Beijing 100015, China. Fax 86 –1084306824; e-mail [email protected]. W.C. at Key Laboratory of
Molecular Medicine, Ministry of Education, Shanghai Medical College, Fudan
nucleotide polymorphism– based clinical NIPT after
the finding of a potential fetal SCA.
© 2014 American Association for Clinical Chemistry
Chromosomal aneuploidy in the human originates from
either gamete meiotic or mitotic cleavage-stage errors in
the early preimplantation embryo (1 ). The vast majority
of these aneuploidies are lethal, either by causing embryonic growth arrest before implantation or by spontaneous
abortion of the developing fetus during the first trimester
of pregnancy. Autosomal trisomies, polyploidies, and
monosomy X are the main groups of chromosomal abnormalities associated with early pregnancy failure (2– 4 ).
A small proportion of these aneuploidies—such as trisomy 21 (Down syndrome), trisomy 18 (Edwards syndrome), trisomy 13 (Patau syndrome), monosomy X
(Turner syndrome), XXY (Klinefelter syndrome), XXX
(triple X syndrome), and XYY (Jacob syndrome)—are
more tolerated, for reasons still unknown, in the second
and third trimester of fetal development and thus present
as chromosome disease in approximately 0.3% of live
births (5 ). For more than 30 years, the practice of prenatal
diagnosis has enabled early identification of clinically important aneuploidies in the established fetus, leading to a
reduction in the number of children born with chromosomal diseases (6 – 8 ). Prenatal diagnosis of the common
trisomies 21, 18, and 13 has proved highly accurate,
thanks to maternal-serum screening, ultrasound, and
follow-up chorionic villus sampling or amniocentesis, in
combination with fetal karyotyping (6, 9 ). In contrast,
apart from monosomy X, the diagnosis of fetuses with sex
chromosome aneuploidies (SCAs)4 or their mosaic SCA
variants remains more problematic, because most af-
University, Shanghai 200032, China. E-mail [email protected].
Yanglin Wang and Yan Chen contributed equally to the work, and both should
be considered first authors.
Received September 1, 2013; accepted October 21, 2013.
Previously published online at DOI: 10.1373/clinchem.2013.215145
4
Nonstandard abbreviations: SCA, sex chromosome aneuploidy; NIPT, noninvasive prenatal test; CPM, confined placental mosaicism; ChrX, chromosome X;
WBC, white blood cell.
†
1
Copyright (C) 2013 by The American Association for Clinical Chemistry
fected pregnancies do not show any overt clinical signs or
ultrasound abnormalities (10, 11 ).
With the clinical introduction of massively parallel
sequencing of maternal plasma, pregnant women can
now choose to have a noninvasive prenatal test (NIPT)
for the clinically important trisomies 21, 18, and 13, as
well as for SCAs, which represent approximately half
the clinically important chromosome abnormalities
seen in the fetus (12 ). Current data from prospective
NIPT studies (13–15 ) have shown remarkably high
sensitivities and specificities, approaching 100% for the
3 fetal trisomies. False-positive and false-negative
NIPT results do occur, albeit at very low frequency
(⬍0.1%). Limited case study follow-up of discordant
NIPT results via fetal karyotyping and placental cytogenetic analysis has shown that confined placental mosaicism (CPM) (16 –18 ) and placental mosaicism (18 –
21 ), in which differences occur in the distribution and
proportion of euploid and aneuploid cells, are important biological factors that either increase the effective
fetal DNA fraction to yield a false-positive fetal aneuploidy or decrease the effective DNA fraction to yield a
false-negative aneuploidy. Although NIPT results are
also highly accurate, slightly lower sensitivities and
specificities have been consistently reported for SCAs
in blinded studies of known SCAs and euploid samples
(22, 23 ) and in prospective studies (13, 15 ). Even with
improvements in the identification of homologous
chromosome X (ChrX) and ChrY sequences (24 ) that
lead to more accurate calculations of ChrX and ChrY z
scores, it is now widely believed that the vast majority
of discordant SCA NIPT results are also caused by either CPM or placental mosaicism, with true fetal mosaicism being an additional contributing factor (9 ).
An altered or mosaic maternal karyotype represents another possible cause of discordant SCAs, because the effective fraction of ChrX DNA in the maternal plasma would be dramatically different from that of
typical maternal 46,XX somatic cells and thus affect the
calculation of the true fraction of fetal ChrX DNA. We
recently reported a discordant NIPT case of trisomy 18
due to 6% maternal mosaicism for this aneuploidy
(15 ), thus demonstrating in principle that an altered
maternal-DNA plasma can substantially skew the final
fetal z score for the involved chromosome. Furthermore, another unusual case report found (upon substantive clinical follow-up) that the source of a trisomy
13 and monosomy 18 false-positive NIPT result ultimately was due to a metastatic tumor in the mother
(25 ). It is well known that there exists a small percentage of otherwise healthy, fertile women who have maternal SCA mosaicism, or occasionally a full-blown
SCA such as XO or XXX, and who can conceive healthy
euploid offspring (11 ). In addition, some women of
advanced reproductive age have been reported to un2
Clinical Chemistry 60:1 (2014)
dergo gradual and preferential loss of the X chromosome that undergoes X inactivation, thus converting
their blood karyotype from XX to an XO/XX mosaic
(26 ). Indeed, a discordant NIPT result in 1 isolated
report was attributed to low-level maternal 45,XO mosaicism (18 ). We therefore speculated that a small but
noteworthy proportion of women presenting for NIPT
have an altered ChrX karyotype that would skew the
maternal ChrX DNA fraction and thus manifest as discordant fetal SCAs. The present study aimed to develop
a rapid and accurate method for maternal SCA detection and to determine the frequency of discordant SCA
NIPT results due to an altered maternal karyotype.
Methods
PREPARATION OF DNA FROM PATIENT BLOOD SAMPLES AND
KARYOTYPING
For NIPT, we collected 10-mL samples of peripheral
blood from pregnant Chinese women (12–16 weeks’
gestation) in a Cell-Free DNA BCT™ tube (Streck) and
sent them to Berry Genomics, Beijing, for processing
and sequencing. The plasma and white blood cells
(WBCs) from 2.3 mL of blood were separated by 2
rounds of centrifugation to produce an approximately
1.3-mL plasma fraction and an approximately 1.0-mL
fraction of WBCs. DNA was extracted from a 1.0-mL
plasma aliquot with the QIAamp Circulating Nucleic
Acid Kit (Qiagen). Genomic DNA was purified from
0.2-mL aliquots of the maternal-WBC fractions. For
the sensitivity and reproducibility analysis, we processed 2-mL blood samples from a healthy male
(46,XY) and from a patient with triple X syndrome
(47,XXX) in the same manner. We karyotyped amniocytes or maternal WBCs via G-banding analysis
of metaphase chromosome spreads at 450-band
resolution.
NONINVASIVE PRENATAL TESTING
We constructed plasma DNA libraries and performed
massively parallel sequencing on the Illumina HiSeq
2000 platform, as previously described (15, 27 ). We
generated approximately 8 ⫻ 106 single-end reads of 36
bp from each library and aligned them to the unmasked
human genome sequence (28 ). We counted uniquely
mapped reads and then calculated z scores for each
chromosome after GC normalization for Chr13,
Chr18, Chr21, and ChrX (27 ). Chromosome z score
values less than ⫺3.0 or greater than ⫹3.0 were classified as abnormal (15 ).
DETERMINATION OF ALTERED MATERNAL KARYOTYPES BY WBC
GENOMIC DNA SEQUENCING
For each sample, we used 50 ng of WBC genomic
DNA with Nextera DNA Sample Preparation Kits
Maternal Mosaicism and Discordant NIPT
Fig. 1. Sequencing data–analysis pipeline for determination of altered maternal karyotype.
CNV, copy number variation.
(Epicentre/Illumina) to construct the sequencing library. Multiple libraries were indexed and pooled
into a single lane. Library fragments were sequenced
to 43 bp (with 7 bp being the index sequence) on a
HiSeq 2000 instrument. Sequencing reads were analyzed according to the data analysis pipeline summarized in Fig. 1. First, sequencing reads were aligned
to the unmasked human genome sequence (hg19).
Second, we calculated the read density by dividing
each chromosome into contiguous 20-kb bins. For
each bin, i, of a given sample, the reads that were
uniquely and perfectly mapped to that bin (RNBi)
were counted and normalized to the total read number for the sample and the total possible unique
36-bp fragment numbers within bin i (Ui), according to the following equation: NRNBi ⫽ RNBi/(total
read number of the sample) ⫻ (8 ⫻ 106)/Ui, where
NRNBi is the normalized read number uniquely and
perfectly mapped to bin i. For any given bin i of the
reference samples, we calculated the median (␮i) of
NRNBi. We assumed that changes in abnormal copy
numbers among test samples are rare for each bin;
therefore, the median of NRNBi represents the normal copy number. For any bin i of a test sample, we
calculated the ratio (ratioi) between the test sample
and the normal chromosome copies according to the
following equation: ratioi ⫽ NRNBi/␮i. We then
plotted log2(ratioi) values for all of the bins for a
given sample to generate the copy number values
along the length of each chromosome. The gain or
loss of chromosome regions was detected via the
fused lasso algorithm, as described previously (29 ).
Lastly, to determine the level of chromosome mosaicism, we calculated the normalized chromosome
representations (NCRj) for each chromosome of any
given test sample, j, and the reference samples
(NCRf) with the method described previously (27 ).
The percentage of chromosomal mosaicism was then
calculated according to the following equation:
(NCRj ⫺ NCRf)/NCRf.
Clinical Chemistry 60:1 (2014) 3
Results
viations from zero (data not shown), indicating normal
copy numbers.
DEVELOPMENT AND VALIDATION OF A ROBUST SEQUENCING
METHOD TO DETERMINE ALTERED MATERNAL KARYOTYPE
IDENTIFICATION OF ABNORMAL MATERNAL KARYOTYPES IN
To determine the effect of any level of maternal mosaicism on the interpretation of NIPT fetal results for
SCAs, we deemed it imperative to develop a rapid and
accurate method. We speculated that separation of the
WBC fraction from maternal blood samples at the
same time as the plasma DNA would most likely represent the best and most accurate source of cells for
analysis, because WBCs are believed to be the primary
source of cells contributing to the steady-state fraction
of maternal DNA in the plasma (9 ). We hypothesized
that the DNA-sequencing strategy used for NIPT could
be applied equally to the mother’s WBCs for rapid
identification of an altered karyotype and/or to various
levels of maternal mosaicism of the sex chromosomes.
To test this hypothesis, we designed a genomic DNA–
mixing model in which we purified DNA from a
47,XXX patient and a 46,XY male control to mimic
various levels of maternal ChrX mosaicism. In this experimental design, we mixed known 46,XY and
47,XXX genomic DNAs in various ratios, starting with
pure 46,XY (equivalent to 45,XO in this context) DNA
and incrementally increasing the proportion of
47,XXX DNA (to 5%, 10%, 20%, 30%, 40%, 50%,
60%, 70%, 80%, 90%, and 95%) to create samples
mimicking increasing proportions of XO in a background of XX, a balanced XX sample (no ChrX
change), samples with increasing proportions of XXX
in a background of XX, and then finishing with pure
XXX. We sequenced the samples and calculated the
percentage gains in copy number for ChrX.
Examples of ChrX-sequencing plots for XO
(100%), XO (80%)/XXX (20%), XO (50%)/XXX
(50%), XO (20%)/XXX (80%), and XXX (100%) samples are shown in Fig. 2. All levels of “XO mosaicism”
and “XXX mosaicism” in an XX background determined by sequencing were virtually identical to the expected values. To assess the reproducibility of the sequencing method, we performed the experiment with 3
independent technical replicates (prepared by 3 different technicians) at each level of XO and XXX mosaicism. We then plotted the actual XO/XXX ratios (experimental) calculated by sequencing against the
expected XO/XXX ratios (theoretical) for all increased
proportions of ChrX in the 3 independent sample replicates (Fig. 3). The actual and expected XO/XXX ratios
were linearly correlated (r 2 ⫽ 0.99893). Equally important, we were able to detect XO mosaicism levels of
⬍5% accurately, indicating that our “sequencing
karyotyping” method was highly sensitive. As expected, the mean log2 values for all autosomes in these
ChrX mosaic DNA samples showed no significant de-
WBCs FROM DISCORDANT NIPT SCAs
4
Clinical Chemistry 60:1 (2014)
In routine NIPT, we identified 3 samples with unusually high ChrX z scores (greater than ⫹5) and 3 samples
with unusually low ChrX z scores (less than ⫺5) that
could not be accounted for statistically as a fetal SCA
(Table 1). We suspected that the abnormally skewed
ChrX z scores were caused by a deviation in the effective ChrX maternal-plasma DNA fraction produced by
an altered WBC karyotype. To investigate this possibility further, we applied our newly developed sequencing
karyotyping strategy to the WBC fractions derived
from the original NIPT blood samples. Sequencing
ChrX karyotyping profiles for 4 of the 6 maternal WBC
samples are depicted in Fig. 4. According to the percentage increase in ChrX (Table 1), NIPT samples 1, 4,
and 6 with high z scores were associated with a 100%
XXX, a 48% XXX mosaic, and a 94% XXX mosaic maternal karyotype, respectively. In contrast, according to
the percentage loss of ChrX, samples 2, 3, and 5 with
low z scores were associated with a 5%, 20%, and 15%
XO mosaic maternal karyotype, respectively. In all 6
samples, the degree of change in the presumed fetal
ChrX DNA fraction was generally correlated with the
degree of ChrX change in the maternal-WBC DNA.
Given that fetal DNA constitutes only a minor fraction
of the maternal plasma DNA, we concluded that the
altered maternal karyotype was the predominant contributor to the final fetal ChrX z score (Table 1). We
also karyotyped the maternal WBCs by G-banding
analysis as a confirmatory step to validate the sequencing maternal karyotypes (Table 1). Apart from sample
2, the karyotypes determined by G-banding analysis
and those determined by sequencing were highly concordant. Follow-up amniocentesis and G-banding
karyotyping revealed that all 6 fetuses were euploid
(Table 1).
CONTRIBUTION OF MATERNAL MOSAICISM TO POTENTIAL
MISDIAGNOSIS OF SCA BY NIPT
To determine the frequency of maternal mosaicism
contributing to discordant SCAs, we surveyed NIPT
SCAs during a 5-month period of clinical NIPT, from
December 2012 to April 2013. During this time, only
trisomies 21, 18, and 13 were reported. A total of 446
abnormal NIPT samples were identified, of which 259
(58.1%) were trisomy 21, 18, or 13; 187 (41.9%) of
these samples were SCAs. Sequencing analysis for the
maternal karyotype of the 187 patients with a positive
SCA NIPT result (including the 6 patients described
above) identified 16 patients (8.56%) with an altered or
mosaic maternal karyotype: 6 (9.52%) of 63 patients
Maternal Mosaicism and Discordant NIPT
Fig. 2. Sequencing karyotypes for ChrX models of mosaicism.
ChrX sequencing profiles are plotted as log2 values of the normalized sequencing densities (y axis), vs. the number of 20-kb
sequencing bins (x axis). The upper dashed line indicates a 100% ChrX gain [log2(3/2)], and the lower dashed line indicates a
100% ChrX loss [log2(1/2)]. The blue line represents the mean log2 value, red boxes represent repeat regions, and the black
box represents the centromere. Models of 100% XO, 20% XO mosaicism, XX, 80% XXX mosaicism, and 100% XXX are shown
along with the actual mean percentage change in ChrX determined by sequencing (presented in parentheses).
had an abnormal ChrX gain, and 10 of 124 patients
(8.06%) had an abnormal ChrX loss (Table 2). For the
remaining 259 patients with a fetal trisomy result,
maternal-DNA sequencing revealed typical 46,XX
karyotypes in all cases.
Discussion
The limited follow-up studies of discordant NIPT results for trisomy 21 (17, 20, 21 ), trisomy 18 (20 ), and
trisomy 13 (19 ) with placental analyses have revealed
the biological phenomena of CPM and/or placental
mosaicism to be important contributing factors.
Therefore, CPM/placental mosaicism and perhaps true
fetal mosaicism (9 ) have also been assumed to be the
major causes of discordant SCAs. In the present study,
we used a novel maternal-WBC sequencing strategy
with high sensitivity and specificity to analyze 187 cases
with abnormal NIPT results and identified 16 samples
(8.5%) in which the discordance was directly attributable to either an altered or a mosaic maternal karyotype. This relatively high frequency of discordant reClinical Chemistry 60:1 (2014) 5
Fig. 3. Sensitivity and reproducibility of sequencing
karyotyping for measurement of ChrX mosaicism.
Genomic samples of 47,XXX and 46,XY DNA were used to
create different levels of XO and XXX mosaicism in an XX
background. Actual XO/XXX ratios (experimental) are plotted against expected XO/XXX ratios (theoretical), with the
mean ⫾ 3 SDs shown for 3 technical replicates. There is a
linear relationship (r 2 ⫽ 0.99893) for the incremental gain
in ChrX. The ratio of 50% XO/XXX is equivalent to XX and
represents the transition point from XO to XXX mosaicism.
sults due to an altered maternal karyotype suggests that
an abnormal maternal ChrX fraction of the plasma
DNA is an important cause of discordant SCAs. This
finding is the first to explain, in part, the consistently
lower sensitivities and specificities that have been reported for SCAs in the clinical setting (13, 15 ). Largescale patient follow-up studies in which fetal karyotyping of abnormal SCA NIPT results are combined with
targeted placental analyses, which were not possible in
the present study, are now needed to determine the
relative contributions of altered maternal karyotype,
CPM/placental mosaicism, and true fetal mosaicism to
discordant SCA NIPT results.
The study highlights the value of determining the
maternal karyotype in increasing the accuracy of reporting NIPT results for chromosomes X and Y. For
this purpose, we specifically developed and validated a
rapid sequencing karyotyping method that uses the
maternal-WBC fraction from the same blood sample
used to isolate the maternal-plasma DNA. For 5 of the 6
SCA-discordant samples, we showed that the degree of
loss or gain of ChrX as determined by sequencing is an
accurate predictor, as confirmed by conventional
G-banding karyotyping. For the remaining sample (no.
2; Table 1), G-banding analysis revealed a significantly
higher proportion of 45,XO cells than sequencing. Because the ChrX NIPT z score of ⫺7.95 was more consistent with the relatively small loss of ChrX predicted
by sequencing, we concluded that the discordant
G-banding result was possibly caused by other factors
known to be associated with karyotyping, such as
suboptimal culture conditions, pseudomosaicism for
45,XO in extended culture (30 ), and even cellcounting errors. The sequencing karyotyping method
specifically developed in this study also has other useful
advantages over conventional karyotyping. First, it is
rapid, allowing identification of a maternal karyotype
that can be used to interpret the NIPT result. Second,
being a molecular assay, the method is scalable. Third,
the method is highly sensitive, with a capacity to detect
not only an altered maternal ChrX karyotype but also
mosaicism levels of ⬍5%. In rare cases, the method
should also be able to identify changes in the autosomal
karyotype. Given these advantages, we therefore recommend that the maternal karyotype be determined
after identification of any abnormal NIPT result. This
order of analysis would improve the accuracy of reporting NIPT ChrX and ChrY results.
Table 1. Detailed analysis and follow-up of NIPT results with unusually high or low ChrX z scores.
NIPT
sample
no.
a
b
6
Fetal
karyotypea
Maternal WBC
ChrX gain/loss
by sequencing
Maternal
karyotype by
G-bandingb
NIPT ChrX
z score
Calculated NIPT fetal
ChrX gain/loss
1
51.55
⫹79.31%
46,XN
⫹103.93%
2
⫺7.95
⫺13.46%
46,XN
⫺4.82%
45,X[64]/46,XX[36]
3
⫺21.84
⫺31.99%
46,XN
⫺19.97%
45,X[10]/46,XX[40]
47,XXX
4
29.34%
⫹48.95%
46,XN
⫹47.18%
47,XXX[60]/46,XX[40]
5
⫺10.53%
⫺11.82%
46,XN
⫺14.35%
45,X[6]/46,XX[78]
6
62.80
⫹87.03%
46,XN
⫹93.69%
47,XXX
N, ChrX or ChrY.
Square brackets indicate the number of metaphases analyzed.
Clinical Chemistry 60:1 (2014)
Maternal Mosaicism and Discordant NIPT
Fig. 4. Altered and mosaic ChrX karyotypes detected by sequencing of maternal WBCs from NIPT samples.
ChrX sequencing profiles are plotted as log2 values of the normalized sequencing densities (y axis), vs. the number of 20-kb
sequencing bins (x axis). The upper dashed line indicates a 100% ChrX gain [log2(3/2)], and the lower dashed line indicates a
100% ChrX loss [log2(1/2)]. The blue line represents the mean log2 value, red boxes represent repeat regions, and the black
box represents the centromere. On the basis of the mean percentage change in ChrX, the 4 maternal-WBC samples with a
93.7% ChrX gain, a 47.2% ChrX gain, a 20.0% ChrX loss, and no gain or loss (normal 46,XX karyotype) are shown.
The findings of the study also indicate that determination of the maternal karyotype will substantially
decrease the rate of discordant SCAs and in effect increase the overall sensitivity and specificity of NIPT for
SCAs. Nevertheless, difficulties remain for NIPT cases
in which the woman has low-level mosaicism for ChrX.
If women with an altered karyotype can conceive
healthy euploid babies, they can also conceive babies
with SCAs, because this mosaicism may also be present
in the germ line (11, 31 ). Although the true fetal ChrX
measurement will be masked in most cases by the dominant maternal ChrX change, laboratories would still
Table 2. Contribution of an abnormal ChrX maternal karyotype in a prospective study of 187 discordant SCAs.
Clinical
NIPT follow-up
NIPT findings
NIPT ChrX gain
NIPT ChrX loss
Total
Abnormal NIPT for SCA, n
63
124
187
Normal maternal karyotype, n
57
114
171
Altered maternal karyotype, n
6
10
16
Maternal mosaicism rate
9.52%
8.06%
8.56%
Clinical Chemistry 60:1 (2014) 7
be alerted to a potential fetal SCA simply because of an
altered maternal karyotype. It is possible in a very small
number of cases, however, that a fetal SCA will be
missed when the fetal SCA is in balance with the maternal X mosaicism. The following hypothetical scenario illustrates this possibility. If the fetus is XXX and
the mother is 10% XO and 90% XX, and if the fetal
DNA fraction is approximately 10%, the NIPT result
will appear normal in the absence of knowledge of the
maternal karyotype. In cases of knowledge of a presumably normal NIPT result but an altered maternal
karyotype, however, the patient can be identified immediately and referred for amniocentesis and fetal
karyotyping to clarify the NIPT result. If an independent and reliable method for calculating the fetal fraction from genomic sequencing data were available
(other than that associated with ChrX and ChrY), this
method could eventually be used as an alternative to
maternal DNA sequencing in a more holistic approach
for identifying whether a ChrX abnormality is of maternal or fetal origin.
On the basis of our finding that a high percentage
of the discordant NIPT SCA results are due to maternal
mosaicism, we recommend that if NIPT is offered clinically via either shotgun sequencing– based or singlenucleotide polymorphism– based approaches, the maternal karyotype should be tested for all samples that
appear abnormal. This recommendation may raise
ethical concerns, particularly in cases in which the
mother has not previously been karyotyped and is unaware of having an altered X karyotype. Given that possibility, NIPT consent forms would need to be modified to state that a small number of fetal SCAs are
attributable to maternal mosaicism. In addition, any
report of such a case should also recommend genetic
counseling for the mother to discuss any future reproductive issues and any potential effects of the altered
karyotype on her health and well-being. One of the
largest challenges remaining for NIPT is discordant trisomy and SCA results caused by placental mosaicism
that either increase or decrease the effective fetal DNA
fraction for the involved chromosome. Although falsepositive results can be identified via amniocentesis and
karyotyping to avoid termination of pregnancies with
an otherwise healthy fetus, false-negative results remain more problematic, particularly for SCAs that
usually do not show any overt clinical signs during
pregnancy.
Author Contributions: All authors confirmed they have contributed to
the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design,
acquisition of data, or analysis and interpretation of data; (b) drafting
or revising the article for intellectual content; and (c) final approval of
the published article.
Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form.
Disclosures and/or potential conflicts of interest:
Employment or Leadership: F. Tian, Berry Genomics; Z. Song,
Berry Genomics; D. Cram, Berry Genomics.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: Shanghai Committee of Science and Technology,
China (grant no. 134119a4600).
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played a direct role in
the design of study, choice of enrolled patients, review and interpretation of data, and preparation and final approval of the manuscript.
References
1. Kuliev A, Verlinsky Y. Meiotic and mitotic
nondisjunction: lessons from preimplantation genetic diagnosis. Hum Reprod Update 2004;10:
401–7.
2. Eiben B, Bartels I, Bähr-Porsch S, Borgmann S,
Gatz G, Gellert G, et al. Cytogenetic analysis of
750 spontaneous abortions with the directpreparation method of chorionic villi and its implications for studying genetic causes of pregnancy wastage. Am J Hum Genet 1990;47:
656 – 63.
3. Stephenson MD, Awartani KA, Robinson WP. Cytogenetic analysis of miscarriages from couples
with recurrent miscarriage: a case-control study.
Hum Reprod 2002;17:446 –51.
4. Simpson JL. Causes of fetal wastage. Clin Obstet
Gynecol 2007;50:10 –30.
5. Hassold T, Hunt P. To err (meiotically) is human:
the genesis of human aneuploidy. Nat Rev Genet
2001;2:280 –91.
6. Driscoll DA, Gross S. Prenatal screening for aneuploidy. N Engl J Med 2009;360:2556 – 62.
8
Clinical Chemistry 60:1 (2014)
7. Raymond FL, Whittaker J, Jenkins L, Lench N,
Chitty LS. Molecular prenatal diagnosis: the impact of modern technologies. Prenat Diagn 2010;
30:674 – 81.
8. Natoli JL, Ackerman DL, McDermott S, Edwards
JG. Prenatal diagnosis of Down syndrome: a systematic review of termination rates (1995–2011).
Prenat Diagn 2012;32:142–53.
9. Benn P, Cuckle H, Pergament E. Non-invasive
prenatal testing for aneuploidy: current status
and future prospects. Ultrasound Obstet Gynecol
2013;42:15–33.
10. Luthardt FW, Keitges E. Chromosomal syndromes
and genetic disease. eLS; 2001. http://onlinelibrary.
wiley.com/doi/10.1038/npg.els.0001446/abstract
(Accessed October 2013).
11. Demaliaj E, Cerekja A, Piazze J. Sex chromosome
aneuploidies. aneuploidy in health and disease. In:
Storchova Z, ed. Aneuploidy in health and disease.
Shanghai: InTech China; 2012. p 123– 40. http://
www.intechopen.com/books/aneuploidy-in-healthand-disease/sex-chromosome-aneuploidies
(Ac-
cessed October 2013).
12. Grati FR, Barlocco A, Grimi B, Milani S, Frascoli G,
Di Meco AM, et al. Chromosome abnormalities
investigated by non-invasive prenatal testing account for approximately 50% of fetal unbalances
associated with relevant clinical phenotypes.
Am J Med Genet 2010;152A:1434 – 42.
13. Bianchi DW, Platt LD, Goldberg JD, Abuhamad
AZ, Sehnert AJ, Rava RP. Genome-wide fetal
aneuploidy detection by maternal plasma DNA
sequencing. Obstet Gynecol 2012;119:890 –901.
14. Norton ME, Brar H, Weiss J, Karimi A, Laurent LC,
Caughey AB, et al. Non-Invasive Chromosomal
Evaluation (NICE) Study: results of a multicenter
prospective cohort study for detection of fetal
trisomy 21 and trisomy 18. Am J Obstet Gynecol
2012;207:137.e1– 8.
15. Song Y, Liu C, Qi H, Zhang Y, Bian X, Liu J.
Noninvasive prenatal testing of fetal aneuploidies
by massively parallel sequencing in a prospective
Chinese population. Prenat Diagn 2013;33:
700 – 6.
Maternal Mosaicism and Discordant NIPT
16. Choi H, Lau TK, Jiang FM, Chan MK, Zhang HY,
Lo PS, et al. Fetal aneuploidy screening by maternal plasma DNA sequencing: ‘false positive’
due to confined placental mosaicism. Prenat Diagn 2013;33:198 –200.
17. Pan M, Li FT, Li Y, Jiang FM, Li DZ, Lau TK, Liao
C. Discordant results between fetal karyotyping
and non-invasive prenatal testing by maternal
plasma sequencing in a case of uniparental disomy 21 due to trisomic rescue. Prenat Diagn
2013;33:598 – 601.
18. Lau TK, Jiang FM, Stevenson RJ, Lo TK, Chan LW,
Chan MK, et al. Secondary findings from noninvasive prenatal testing for common fetal aneuploidies by whole genome sequencing as a clinical service. Prenat Diagn 2013;33:602– 8.
19. Hall AL, Drendel HM, Verbrugge JL, Reese AM,
Schumacher KL, Griffith CB, et al. Positive cellfree fetal DNA testing for trisomy 13 reveals
confined placental mosaicism. Genet Med 2013;
15:729 –32.
20. Canick JA, Palomaki GE, Kloza EM, LambertMesserlian GM, Haddow JE. The impact of maternal plasma DNA fetal fraction on next generation sequencing tests for common fetal
aneuploidies. Prenat Diagn 2013;33:667–74.
21. Wang Y, Zhu J, Chen Y, Zhao X, Wu Y, Han X,
et al. Two cases of placental T21 mosaicism:
challenging the detection limits of non-invasive
prenatal testing. Prenat Diagn [Epub ahead of
print 2013 Aug 8].
22. Samango-Sprouse C, Banjevic M, Ryan A, Sigurjonsson S, Zimmermann B, Hill M, et al. SNPbased non-invasive prenatal testing detects sex
chromosome aneuploidies with high accuracy.
Prenat Diagn 2013;33:643–9.
23. Mazloom AR, Džakula Ž, Oeth P, Wang H, Jensen
T, Tynan J, et al. Noninvasive prenatal detection
of sex chromosomal aneuploidies by sequencing
circulating cell-free DNA from maternal plasma.
Prenat Diagn 2013;33:591–7.
24. Skaletsky H, Kuroda-Kawaguchi T, Minx PJ, Cordum HS, Hillier L, Brown LG, et al. The malespecific region of the human Y chromosome is a
mosaic of discrete sequence classes. Nature
2003;423:825–37.
25. Osborne CM, Hardisty E, Devers P, KaiserRogers K, Hayden MA, Goodnight W, Vora NL.
Discordant noninvasive prenatal testing results
in a patient subsequently diagnosed with met-
astatic disease. Prenat Diagn 2013;33:609 –11.
26. Russell LM, Strike P, Browne CE, Jacobs PA. X
chromosome loss and ageing. Cytogenet Genome
Res 2007;116:181–5.
27. Liang D, Lv W, Wang H, Xu L, Liu J, Li H, et al.
Non-invasive prenatal testing of fetal whole chromosome aneuploidy by massively parallel sequencing. Prenat Diagn 2013;33:409 –15.
28. Li H, Durbin R. Fast and accurate short read
alignment with Burrows-Wheeler transform.
Bioinformatics 2009;25:1754 – 60.
29. Tibshirani R, Wang P. Spatial smoothing and hot
spot detection for CGH data using the fused
lasso. Biostatistics 2008;9:18 –29.
30. Moertel CA, Stupca PJ, Dewald GW. Pseudomosaicism, true mosaicism, and maternal cell contamination in amniotic fluid processed with in
situ culture and robotic harvesting. Prenat Diagn
1992;671– 83.
31. Onalan G, Yilmaz Z, Durak T, Sahin FI, Zeyneloglu HB. Successful pregnancy with preimplantation genetic diagnosis in a woman with mosaic Turner syndrome. Fertil Steril 2011;95:
1788.e1–3.
Clinical Chemistry 60:1 (2014) 9