Non-invasive prenatal diagnostic testing for β-thalassaemia using cell-free fetal DNA and next generation sequencing

This article is protected by copyright. All rights reserved.

Non-invasive prenatal diagnostic testing for -thalassaemia using cell-free

fetal DNA and next generation sequencing

Li Xiong1,2*, Angela N Barrett1*, Rui Hua1,2, Tuan Zea Tan3, Sherry Sze Yee

Ho4,Jerry KY Chan5,6, Mei Zhong2#, Mahesh Choolani1#

1Department of Obstetrics and Gynaecology, Yong Loo Lin School of Medicine,

National University of Singapore, Singapore, 119228.

2Department of Gynecology & Obstetrics, Nanfang Hospital, Southern Medical

University, Guangzhou, China.

3Cancer Science Institute of Singapore, National University of Singapore,

Singapore.

4Department of Laboratory Medicine, Molecular Diagnosis Centre, National

University Hospital, Singapore, S119074

5Experimental Fetal Medicine Group, Department of Obstetrics and

Gynaecology, Yong Loo Lin School of Medicine, National University of

Singapore, Singapore, 119228.

6Department of Reproductive Medicine, KK Women's and Children's Hospital,

Singapore, S229899

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/pd.4536


* Authors contributed equally

# Corresponding Authors:

Mahesh Choolani, Department of Obstetrics & Gynaecology, National University

of Singapore, NUHS Tower Block, Level 12, 1E Kent Ridge Road, Singapore

119228.

Tel: +65 67722672; Fax: +65 67794753; e-mail: [email protected].

Mei Zhong, Department of Obstetrics & Gynaecology, Nan Fang Hospital,

Southern Medical University, Guangzhou, Guangdong, 510515,PR China.

Tel: +86 20-61641901; Fax: +86 20-62787562; e-mail: [email protected]

Running title: NIPD for β-thalassaemia using next generation sequencing

Word count: 3,414

Figure count: 2

Tables: 3

Funding: We acknowledge the National Medical Research Council /Clinician

Scientist Award (NMRC/CSA/007/2009) for supporting our work. The funders

played no part in the study design, data collection, data analysis or manuscript

preparation.

Conflict of Interest: The authors have no conflicts of interest to declare.

What is already known about this topic?

Cell-free fetal DNA in maternal plasma can be used to detect paternally

inherited mutations absent from the mother’s DNA.

If the fetus has not inherited the paternal mutation, no further invasive

testing is required.


What does this study add?

A simple next generation sequencing test with a 100% detection rate can

detect inheritance of paternal -thalassaemia alleles.

This sequencing method can be applied to any paternally inherited point

mutation or small insertion/deletion.

ABSTRACT

Objective To develop an accurate non-invasive prenatal test using next

generation sequencing (NGS) for HbE and the four most common β-

thalassaemia mutations found in South East Asia (namely -28A>G, CD17A>T,

CD41/42(-TTCT), and IVS-II-654C>T).

Methods: Cell-free DNA was extracted from maternal plasma from 83 families

where both parents were carriers of the HbE mutation or one of four common β-

thalassaemia mutations. Overlapping PCR amplicons covering each mutation

were generated, pooled, and sequenced using the Illumina Miseq. Fastq files

were analysed to detect inheritance of the paternal mutation.

Results: In two cases where the fathers were compound heterozygotes for HbE

and -28A>G, the fetus was correctly diagnosed as having inherited one of the

paternal mutations. In 35/85 cases, the paternal mutation was not detected, and

in 50/85 cases it was classified as inherited. Overall sensitivity for detection of

paternal mutations was 100% (95% CI: 92.4%-100%), and specificity was

92.1% (95% CI: 79.2%-97.3%).


Conclusion: We demonstrated that detection of paternal mutations using NGS

can be readily achieved with high sensitivity and specificity, removing the need

for an invasive test in 50% of pregnancies at risk of a thalassaemia in cases

where the father and mother carry a different mutation.

INTRODUCTION

Beta-thalassaemia is a common autosomal recessive disorder caused by

reduced (β+) or absent (β0) synthesis of the -globin chains of the haemoglobin

tetramer1. Mutations occur on the HBB gene located on chromosome 11. When

both parents are carriers of a thalassaemia mutation, there is a 25% risk for

thalassaemia major in each pregnancy. Thalassaemia can result in profound

anaemia from early life and, if not treated with regular blood transfusions, can

lead to death in the first year2,3. These blood transfusions bring their own

complications, mainly caused by iron-overload, resulting in endocrinopathies,

cardiac complications and death3. Over 200 different HBB mutations have been

identified to date, the majority being single nucleotide substitutions, deletions, or

insertions of nucleotides leading to a frame-shift4; rarely, -thalassaemia results

from gross gene deletion. In South East Asia between up to 11% of the

population are estimated to be carriers for -thalassaemia, with approximately

65% of all β-thalassaemia mutations caused by four mutations in the HBB gene:

-28A>G (under HGVS nomenclature, HBB:c.-78A>G), CD17A>T


(HBB:c.52A>T), CD41/42(-TTCT) (HBB:c.126_129delCTTT), and IVS-II-

654C>T (HBB:c.316-197C>T)1,5. Haemoglobin E is another common structural

haemoglobin variant in South Asia, caused by a CD26GAG>AAG

(HBB:c.79G>A) mutation in the HBB gene, often occurring as compound β-

thalassaemia leading to serious anaemia and transfusion dependency6.

Prenatal diagnosis for thalassaemia syndromes is an essential part of

preventive genetics, and is currently dependent on the use of invasive

diagnostic tests, which carry a small but significant risk of miscarriage of up to

1%7. Although more recent data suggests that the risk is generally much lower

than 1%8, it may be the case that in centres with less experienced staff, the risk

is far greater. Because of this potential risk, a major goal in prenatal diagnosis

has been to develop methods to test non-invasively using cell-free DNA

(cfDNA) from a maternal blood sample, thus eliminating the risk of miscarriage.

Cell-free fetal DNA (cffDNA) constitutes approximately 10-15% of the total

cfDNA9, and has been shown to represent the entire fetal genome10.

Detection of inheritance of paternal β-thalassaemia mutations using NIPD has

been previously reported11-16. Many different methods have been used in the

development of these tests, including allele-specific real-time PCR (AS-

PCR)11,12, single-allele base extension reaction (SABER)13, mass arrayed

primer extension (APEX)14, digital PCR15, and MALDI-TOF mass spectrometry16

methods. The sensitivity of any test for NIPD is hampered by the low proportion

of cffDNA present in maternal plasma. In order to overcome this problem,

strategies for enrichment of circulating cffDNA and depletion of maternal

background cfDNA have been used. These strategies are dependent on the fact


that cffDNA molecules have been shown to be shorter than maternally derived

cfDNA10, with a length of less than 146bp on average, but all have their

disadvantages. Agarose gel electrophoresis followed by DNA purification can

be used to enrich smaller DNA fragments17, but is relatively labour-intensive

and prone to contamination, so is not ideal for implementation into clinical

practice. Digital nucleic acid size selection (NASS) gives a 32.6-38.0%

enrichment of fetal DNA15; however this methodology requires expensive digital

PCR equipment and has not been widely adopted. Enrichment of the fetal

fraction by suppression of maternal cfDNA can be achieved by the use of a

peptide-nucleic-acid (PNA) clamp17,18 and COLD-PCR19, requiring extensive

optimization for each amplicon. Addition of formaldehyde to a blood sample was

shown by Dhallan et al. to enrich the fetal fraction20; however these results are

controversial, since other groups have not found them to be reproducible21, 22.

The advent of next generation sequencing (NGS) has allowed us to count

millions of DNA molecules from a single sample, greatly increasing the

sensitivity of molecular tests using cfDNA. Several groups are working to

develop NIPD for β-thalassaemia using NGS. Lo et al.10 performed massively

parallel whole genome sequencing on a sample from a family where both

parents were carriers of different β-thalassaemia mutations, and successful

diagnosis of the fetus as a heterozygous carrier revealed that it is feasible to

use NIPD for single gene disorders. However, the expense associated with this

high coverage sequencing precludes its use for routine prenatal diagnosis.

Targeted sequencing is a far more cost-effective strategy to capture cfDNA

molecules from specific genomic regions. Using in-solution capture, sequence


coverage of targeted regions can be increased by over 200 times23. This

approach was utilized to capture thousands of SNPs on the HBB gene cluster

and targeted RHDO analysis was used to diagnose β-thalassaemia in two

families24. In another study, using NGS for four informative SNPs in the HBB

locus, Papasavva et al. achieved correct classification of paternal inheritance in

eight samples out of ten using haplotype analysis25.

In this study, we directly targeted the five most common South East Asian

mutation loci on the HBB gene using overlapping PCR amplicons, and used the

Illumina MiSeq next generation sequencing platform to detect the presence of

paternally inherited mutant alleles in maternal plasma. Our aim was to develop

a simple, rapid, and cost-effective method for detection of paternal mutations

using NIPD, thereby removing the need for an invasive test in 50% of cases at

risk of β-thalassaemia where parents are heterozygous for different mutations.

MATERIALS AND METHODS

Patient recruitment and sample collection

Blood was collected at the Department of Gynaecology and Obstetrics at

Nanfang Hospital, Southern China from 140 pregnant couples at risk of β-

thalassaemia, attending a clinic for prenatal diagnosis. This study was

performed with the approval of the Institutional Review Board for Nanfang

Hospital (study number NFEC-2014-048). 10mL of maternal peripheral blood

was collected into two K3-EDTA tubes and 2mL of paternal peripheral blood

was collected into a single K3-EDTA tube. Specimens for genotyping the fetus

obtained between 11-14 weeks of gestation were collected by chorionic villus


sampling (CVS), and from 14+1 - 24 weeks of gestation were obtained by

amniocentesis. Cord blood was collected during the third trimester.

Sample processing and DNA extraction

Plasma was separated from 10 mL of maternal blood within 4 hours of blood

draw using previously published protocols26. Plasma DNA was extracted from 2

mL plasma using the QIAamp Circulating Nucleic Acid Kit (Qiagen, California,

USA) according to manufacturer's instructions. Plasma DNA was eluted into a

final volume of 60 µL AVE elution buffer, provided as part of the kit. Genomic

DNA (gDNA) was extracted from buffy coat and fetal material (CVS samples,

amniotic fluid samples) using a FuJi-film gDNA extraction kit (FuJi, Japan)

according to manufacturer's instructions.

Molecular genotyping of β-thalassaemia mutations

Parental buffy coat DNA and fetal DNA were genotyped using a PCR-based

reverse dot blot method27. Only samples from couples carrying different

mutations were selected for further analysis.

Primer design

Three pairs of primers specific for each of the four most common South East

Asian β-thalassaemia mutations, namely -28A>G, CD17A>T, CD41/42(-TTCT),

and IVS-II-654C>T, were designed using Primer 3 software, using the reference

sequence for the haemoglobin gene locus (from the NCBI database, accession

number NG_000007.3; http://www.ncbi.nlm.nih.gov/nuccore/NG_000007.3).

Since a fifth mutation, HbE (CD26GAG>AAG), and the CD17A>T mutations are

located very close together on the HBB gene, the same primer pairs were used

to amplify both. Forward primers contained P5 adaptor sequences

http://www.ncbi.nlm.nih.gov/nuccore/NG_000007


complementary to those found on the Illumina MiSeq flow cell, a unique patient-

specific index sequence to allow sample multiplexing, and amplicon-specific

sequences; reverse primers contained a P7 primer sequence (all sequences

detailed in Supplementary Table 1). Amplicons ranged in size from 178-212bp

(including adaptor and index sequences). All primer sequences were examined

for the presence of SNPs using SNPCheck3

(https://secure.ngrl.org.uk/SNPCheck/snpcheck.htm).

Library preparation and NGS

Targeted PCR was performed using three overlapping amplicons per mutation.

Each reaction was carried out using 10 µL of 2 x SensiFAST SYBR Lo-Rox kit

(Bioline, UK), 10 M each primer and 5 µL of cfDNA in a total final volume of 20

µL. Thermal cycling conditions were as follows: 2 minutes at 95°C, followed by

32 cycles of 95°C for 20 s, 60°C for 30 s, and 72°C for 15 s. Amplicons were

purified using Isolate II PCR Kit clean-up columns (Bioline) and eluted into 20

µL of elution buffer, then quantified using a Qubit Fluorometer (Life

Technologies, USA). Amplicon quality was assessed using a Bioanalyzer

(Agilent, USA) and a DNA1000 chip to check for presence of primer dimers.

Purified PCR products were diluted to 2 nM in Elution Buffer (Qiagen, USA) and

equal amounts were pooled for up to 20 patients to yield a single 2 nM library.

The sample was denatured using sodium hydroxide according to MiSeq v3

sample preparation kit instructions (Illumina, California, USA), and then the

library was diluted to a final concentration of 8 pM. This library was mixed with

an 8 pM PhiX control to give a 5% PhiX spike, providing sequence diversity.


Cluster generation was carried out on board the MiSeq followed by initiation of a

single-end 100 cycle sequencing protocol.

Data Analysis

De-multiplexing and base-calling were performed using Illumina MiSeq Reporter

Software version 2.4 to generate FASTQ sequence files for each sample. A

Visual Basic script was used to search each FASTQ file for wild-type and

mutant sequences, with five to six base pairs up- and down-stream of each

mutation included in the search terms. Frequency of wild-type and mutant

alleles were generated by counting the occurrence of the wild-type and mutant

sequences in each FASTQ file, with only exact matches being detected. We

calculated the proportion of the paternal allele present in each sample using the

formula: f = (2p/(p+q)) x 100%, where f is the percentage of fetal DNA, p is the

read count for the fetus-specific allele (paternal origin), and q is the read count

for the allele shared by the maternal and fetal genomes15. A threshold of 1%

was set below which the paternal mutation was said not to be inherited. Since

each mutation locus was targeted by three overlapping amplicons the

proportion of paternal mutation allele was counted, and two out of three

amplicons were required to be negative for us to consider that the paternal

mutation had not been inherited by the fetus; if two out of three were positive,

we considered that the fetus had inherited the paternal mutation.


RESULTS

Genotyping of the parental mutations

Eleven different mutations were found in the HBB gene of the 140 couples

recruited to the study, with 92.9% (262/282) of these being the five common

South East Asian mutations (Table 1). Two fathers were compound

heterozygotes, hence the 282 mutations for 140 couples. Of those carrying the

common mutations, eighty-three couples were found to be carriers of different

β-thalassaemia mutations (Table 2), the remaining forty-one carrying the same

mutations. Among those cases where parents carried different mutations,

twenty-seven fathers were carriers of the CD41/42(-TTCT) mutation, 21 carried

the IVS-II-654 mutation, 14 carried CD17A>T, and 18 carried the -28A>G

mutation. In one case the paternal mutation was HbE. For the two remaining

families, the fathers were compound heterozygotes for HbE and -28A>G

mutations. Fetal genomic DNA was also genotyped for each family where the

parents carried different mutations, and we found that 24/83 were affected with

β-thalassaemia (and therefore all carrying a paternal mutation), 47/83 were

carriers (of which 23 were carrying the paternal mutation) and the remaining

12/83 were unaffected. cfDNA was extracted from maternal plasma from all

cases where parent’s carried different mutations, and was analysed to detect

inheritance of the paternal mutation.

Detection of paternal mutations in cfDNA

All paternal mutation loci were targeted by three overlapping amplicons. Single-

end sequencing produced an average of 25 million reads per run, with a mean

read count of 82,282 per individual amplicon, giving an average read count per


mutation for each patient of 246,846. Raw NGS data is presented in

Supplementary Table 2.

Thirty-three of the eighty-one cases in which the fathers carried a single

mutation were correctly found to be negative for the paternal mutation, with less

than 1% paternal mutation detected per amplicon (Figure 1). Of these, 24 of 33

cases were shown to be negative by all three overlapping amplicons and the

remaining nine by two negative amplicons. Three cases that had not inherited

the paternal mutation were found to be positive using our test (false positives).

Amongst these false positives, there was one case of CD41/42(-TTCT), one of

CD17A>T and one of -28A>G. We correctly detected paternal inheritance

(greater than 1% mutant sequence) in all 45 cases where the father carried a

single mutation (Figure 2), as well as detecting the three false positives

mentioned previously, giving a total count of 48 mutation-positive samples.

Thirty-six of these forty-eight cases were positive according to results for all

three overlapping amplicons and the remaining 12 cases were shown to be

positive in two amplicons each (all data presented in Supplementary table 2).

In the two families where the fathers were compound heterozygotes for HbE

and -28A>G, both fetuses were correctly diagnosed as having inherited only

one paternal mutation, in one case only inheriting the HbE mutation and in the

other case only inheriting the -28A>G mutation (Figures 1 and 2). Counting a

total of 85 mutations (81 cases of a single mutation, and two cases of a

compound mutation), inheritance of the paternal mutation was not detected in

35 of 85 cases (41.2%), with inheritance being diagnosed in 50 of 85 cases

(58.8%) (Table 1). There were three cases (3.5%) that gave a false positive


result using the overlapping amplicon sequences. Diagnostic accuracy was

96.5% (82/85), sensitivity was 100% (95%CI: 92.4%-100%), and specificity was

92.1% (95%CI: 79.2%-97.3%). The positive predictive value is therefore 94.0%

(95% CI: 83.8%-97.9%) and negative predictive value is 100% (90.1%-100%).

DISCUSSION

In this study we have shown that NIPD using cfDNA can detect paternally

inherited β-thalassaemia mutations with 100% sensitivity. We used overlapping

amplicons to target the four most common β-thalassaemia mutations in South

East Asia, as well as HbE (together affecting 92.9% of our study cohort), and

performed targeted sequencing on 83 cases where the parents were carriers of

different β-thalassaemia mutations. As well as 100% sensitivity, we achieved a

specificity of 92.1%, with a positive predictive value of 94% and a negative

predictive value of 100%. The latter indicates that the β-thalassaemia major can

be reliably excluded in the absence of a paternal mutation, removing the

requirement for an invasive test in 50% of cases where the father is a carrier for

one of five common South East Asian mutations. Assuming that our data

reflects the incidence of the mutations in the South East Asian population as a

whole, we would expect that 59.3% of couples would carry a different mutation.

In 50% of cases, the baby will not have inherited the father’s mutant allele, and

therefore the test will remove the need for an invasive procedure in 29.6% of all

pregnancies at risk of -thalassaemia.

The use of overlapping amplicons is essential for reducing potential for false

negative results due to the presence of SNPs under the primers (and indeed, in

our study, no false negatives were seen). The HBB gene contains a high


number of SNPs, and so it is challenging to design primer pairs avoiding all

known SNPs. Although we attempted to minimize the number of SNPs covered

by any given primer pair using SNPCheck, to increase our confidence in a

negative result, we decided to include three pairs of primers per mutation that

covered the same mutation, but avoided using the same priming regions. If the

paternal allele happens to be located in a region near to a linked SNP, this

strategy should help to prevent misclassification due to allele dropout. Cell-free

DNA is highly fragmented, and although evidence suggests that the fragments

are approximately 146bp in length, and are associated with nucleosomes10

there is no accurate method to predict exactly where the fragmentation will be

located. It is therefore possible that any given pair of primers may not bind to all

fragments containing a particular mutation. Ideally, amplicons will be designed

to be as small as possible, increasing their chance of annealing, but another

benefit of using three overlapping amplicons is that even if one pair of primers

cannot bind to a particular fragment, another should be able to, thus reducing

the likelihood of false negatives.

Sequencing errors caused by the MiSeq itself, library preparation, and starting

quantity of template can affect the results for a particular sample28-31, but it is

unlikely that the same error would have been introduced to all three amplicons.

We found that this was particularly important for the CD41/42(-TTCT) site,

where accuracy was enhanced from 81.5% to 96% by use of overlapping

amplicons (see Table 3). Another study using amplicons targeting four linked

SNPs found in the HBB gene to detect the paternally inherited mutant allele25

returned four false negative and three false positive results from a total of 34


cases. Our study yielded a significant improvement of more than 17% in

diagnostic accuracy on a larger cohort, possibly attributable to the use of three

overlapping amplicons for each mutation.

We obtained three false positives, one for a sample where the father carried

the four base pair CD42/42(-TTCT) deletion, with all three overlapping

amplicons indicating that the fetus had inherited the paternal mutation. The

other two false-positive cases were one -28A>G and one CD17A>T, both cases

having two out of three overlapping amplicons providing false positive results.

Our laboratory is developing a panel of insertion/deletion polymorphisms that

can be added to the sequencing reaction; this may help to prevent false positive

results due to contamination going undetected, since it should alert us to the

presence of an unexpected proportion of alleles being found in the sample,

indicating the possibility of a sample mix-up. There is also the possibility of PCR

bias being introduced during the library preparation32, although it seems unlikely

that this would affect only a single sample for each of three mutations,

especially considering that the overlapping amplicons do not share a priming

site. Our one-step process for library preparation involves significantly less

sample-handling and subsequent PCR than a standard multi-step library

preparation protocol, and so less bias should be introduced33. Even so, high

numbers of PCR cycles can still lead to higher error rate, and reduction of the

PCR cycles may improve the data further. The GC content of our targeted

amplicons ranged from 30.0% to 60.3%, and so we would not expect there to

be any contribution of GC content to PCR-bias. With the exception of the

CD41/42(-TTCT) mutation, the wild-type and mutant amplicons are the same


length, and just have one base-pair substitution, so it would be expected that

the amplification efficiency differences would be minimal. We introduced a 1%

threshold below which a sample was considered to be negative for a particular

mutation. Low levels of background counts can occasionally be seen for some

samples, and are most likely attributable to inherent raw sequencing errors that

can be found in data from all NGS platforms34. Using a conservative threshold

of 1%, we ensure that we reduce false positives caused by these stochastic

sequencing chemistry errors.

Women will be encouraged to take an invasive test to confirm whether the

fetus has also inherited the maternal mutation in the case of a positive result for

paternal inheritance. Using targeted sequencing specific for the mutations of

interest means that this method is applicable to any disorder where there is a

point mutation or small indel, in contrast to approaches that rely on haplotyping

for the paternal allele, where there may be no informative SNPs present near to

the mutation of interest25. We expect to be able to expand the panel of

mutations to cover all -thalassaemia mutations. The main limitation will be that

in areas of lower genetic diversity, the chances may be high that parents will

both carry the same mutation, and so paternal detection using our test cannot

be carried out. In the cohort that we tested, the parents carried a different

mutation in two thirds of cases.

The assay only requires a simple PCR to generate the libraries, and with the

addition of sample-specific index sequences, samples from a number of

different patients can be examined in a single MiSeq run. Therefore our


procedure is cost-effective, and has a very short turnaround time, since the

MiSeq run itself takes approximately six hours.

CONCLUSIONS

We have demonstrated that overlapping targeted amplicon sequencing for

the four most common South East Asian β-thalassaemia mutations plus HbE

can be used for reliable detection of paternally inherited alleles. A test with

100% sensitivity will allow us to detect all cases of paternally inherited β-

thalassaemia mutations, and so in 29.6% of all the cases referred for prenatal

testing for -thalassaemia, the requirement for an invasive test will be removed.

ACKNOWLEDGEMENTS

We would like to acknowledge Chaoqun Xiao and Siping Liu for assistance in

collection of the samples.

REFERENCES

1. Weatherall DJ, Clegg JB. The thalassemia syndromes. Oxford. 4th edition:

Blackwell Scientific Publications, 2001.

2. Cao A, Galanello R. Beta-thalassemia. Genet Med 2010; 12:61-76.

3. Galanello R, Eleftheriou A, Traeger-Synodinos J, et al. Prevention of

Thalassaemias and Other Haemoglobin Disorders: Thalassaemia

International Federation, 2003.

4. Giardine B, van Baal S, Kaimakis P, et al. HbVar database of human

hemoglobin variants and thalassemia mutations: 2007 update. Hum Mutat

2007; 28:206.


5. Colah R, Gorakshakar A, Nadkarni A. Global burden, distribution and

prevention of beta-thalassemias and hemoglobin E disorders. Expert Rev

Hematol 2010; 3:103-17.

6. Vichinsky E. Hemoglobin e syndromes. Hematology Am Soc Hematol Educ

Program 2007; 79-83.

7. Tabor A, Alfirevic Z. Update on procedure-related risks for prenatal

diagnosis techniques. Fetal Diagn Ther 2010; 27:1-7.

8. Akolekar R, Beta J, Picciarelli G, et al. Procedure-related risk of miscarriage

following amniocentesis and chorionic villus sampling: a systematic review

and meta-analysis. Ultrasound Obstet Gynecol 2014.

9. Lun FM, Chiu RW, Chan KC, et al. Microfluics digital PCR reveals a higher

than expected fraction of fetal DNA in maternal plasma. Clin Chem 2008;

54:1664-72.

10. Lo YM, Chan KC, Sun H, et al. Maternal plasma DNA sequencing reveals

the genome-wide genetic and mutational profile of the fetus. Sci Transl Med

2010; 2:61ra91.

11. Chiu RW, Lau TK, Leung TN, et al. Prenatal exclusion of beta thalassaemia

major by examination of maternal plasma. Lancet 2002; 360:998-1000.

12. Papasavva T, Kalakoutis G, Kalikas I, et al. Noninvasive prenatal diagnostic

assay for the detection of beta-thalassemia. Ann N Y Acad Sci 2006;

1075:148-53.

13. Ding C, Chiu RW, Lau TK, et al. MS analysis of single-nucleotide

differences in circulating nucleic acids: Application to noninvasive prenatal

diagnosis. Proc Natl Acad Sci U S A 2004; 101:10762-7.


14. Papasavva T, Kalikas I, Kyrri A, Kleanthous M. Arrayed primer extension for

the noninvasive prenatal diagnosis of beta-thalassemia based on detection

of single nucleotide polymorphisms. Ann N Y Acad Sci 2008; 1137:302-8.

15. Lun FM, Tsui NB, Chan KC, et al. Noninvasive prenatal diagnosis of

monogenic diseases by digital size selection and relative mutation dosage

on DNA in maternal plasma. Proc Natl Acad Sci USA 2008; 105:19920-5.

16. Li Y, Di Naro E, Vitucci A, et al. Size fractionation of cell-free DNA in

maternal plasma improves the detection of a paternally inherited beta-

thalassemia point mutation by MALDI-TOF mass spectrometry. Fetal Diagn

Ther 2009; 25:246-9.

17. Li Y, Di Naro E, Vitucci A, Zimmermann B, Holzgreve W, Hahn S. Detection

of paternally inherited fetal point mutations for beta-thalassemia using size-

fractionated cell-free DNA in maternal plasma. JAMA 2005; 293:843-9.

18. Galbiati S, Foglieni B, Travi M, et al. Peptide-nucleic acid-mediated

enriched polymerase chain reaction as a key point for non-invasive prenatal

diagnosis of beta-thalassemia. Haematologica 2008; 93:610-4.

19. Galbiati S, Brisci A, Lalatta F, et al. Full COLD-PCR protocol for noninvasive

prenatal diagnosis of genetic diseases. Clin Chem 2011; 57:136-8.

20. Dhallan R, Au WC, Mattagajasingh S, et al. Methods to increase the

percentage of free fetal DNA recovered from the maternal circulation. JAMA

2004; 291:1114-9.

21. Chung GT, Chiu RW, Chan KC, et al. Lack of dramatic enrichment of fetal

DNA in maternal plasma by formaldehyde treatment. Clin Chem 2005;

51:655-8.


22. Chinnapapagari SK, Holzgreve W, Lapaire O, et al. Treatment of maternal

blood samples with formaldehyde does not alter the proportion of circulatory

fetal nucleic acids (DNA and mRNA) in maternal plasma. Clin Chem 2005;

51:652-5.

23. Liao GJ, Lun FM, Zheng YW, et al. Targeted massively parallel sequencing

of maternal plasma DNA permits efficient and unbiased detection of fetal

alleles. Clin Chem 2011; 57:92-101.

24. Lam KW, Jiang P, Liao GJ, et al. Noninvasive prenatal diagnosis of

monogenic diseases by targeted massively parallel sequencing of maternal

plasma: application to β-thalassemia. Clin Chem 2012; 58:1467-75.

25. Papasavva T, van Ijcken WF, Kockx CE, et al. Next generation sequencing

of SNPs for non-invasive prenatal diagnosis: challenges and feasibility as

illustrated by an application to beta-thalassaemia. Eur J Hum Genet 2013;

12:1403-10.

26. Chan KCA, Yeung SW, Lui WB, et al. Effects of preanalytical factors on the

molecular size of cell-free DNA in blood. Clin Chem 2005; 54:781-4.

27. Li D, Liao C, Li J, et al. Prenatal diagnosis of beta-thalassemia by reverse

dot-blot hybridization in southern China. Hemoglobin 2006; 30:365-70.

28. Loman NJ, Misra RV, Dallman TJ, et al. Performance comparison of

benchtop high-throughput sequencing platforms. Nat Biotechnol 2012;

30:434-9.

29. Lou DI, Hussmann JA, McBee RM, et al. High-throughput DNA sequencing

errors are reduced by orders of magnitude using circle sequencing. Proc

Natl Acad Sci USA 2013; 110:19872-7.


30. Kinde I, Wu J, Papadopoulos N, et al. Detection and quantification of rare

mutations with massively parallel sequencing. Proc Natl Acad Sci USA

2011; 108:9530-5.

31. Casbon JA, Osborne RJ, Brenner S, Lichtenstein CP: A method for

counting PCR template molecules with application to next-generation

sequencing. Nucleic Acids Res 2011; 39:e81.

32. Aird D, Ross MG, Chen WS, et al. Analyzing and minimizing PCR

amplification bias in Illumina sequencing libraries. Genome Biol 2011; 12:

R18.

33. Kozarewa I, Turner DJ: Amplification-free library preparation for paired-end

Illumina sequencing. Methods Mol Biol 2011; 733:257-66.

34. Quail MA, Smith M, Coupland P, et al. A tale of three next generation

sequencing platforms: comparison of Ion Torrent, Pacific Biosciences and

Illumina MiSeq sequencers. BMC Genomics. 2012;13:341-54.

http://www.ncbi.nlm.nih.gov/pubmed/22827831




Table 1: The allele frequency of β-thalassaemia and HbE mutations in 140

couples. HGVS nomenclature is also given.

Mutations HGVS Nomenclature Number (n) Frequency (%)

Five Common Mutations 262 92.9

CD 41/42(-TTCT) HBB:c.126_129delCTTT 102 36.2

IVS-II-654C>T HBB:c.316-197C>T 79 28.0

CD17A>T HBB:c.52A>T 38 13.5

-28A>G HBB:c.-78A>G 35 12.4

CD26GAG>AAG

(HbE)

HBB:c.79G>A 8 2.8

Rare Mutations 20 7.1

CD27-28(+C) HBB:c.84_85insC 5 1.8

CD14-15(+G) HBB:c.45_46insG 5 1.8

CD71-72(+A) HBB:c.216_217insA 4 1.4

-29A>G HBB:c.-79A>G 3 1.1

IVS-1-1G>T HBB:c.92+1G>T 2 0.7

CD43G>T HBB:c.130G>T 1 0.4

Total 282


Table 2: Genotyping results for parental mutations in 124 couples

Father Mother Cases(n)

Parents carrying different mutations 83

Hetero CD41/42(-TTCT) Hetero IVS-II-654C>T 15

Hetero CD41/42(-TTCT) Hetero CD17A>T 4

Hetero CD41/42(-TTCT) Hetero-28 A>G 6

Hetero CD41/42(-TTCT) Hetero-29A>G 1

Hetero CD41/42(-TTCT) Hetero HbE 1

Hetero IVS-II-654C>T Hetero CD41/42(-TTCT) 10

Hetero IVS-II-654C>T Hetero CD17A>T 7

Hetero IVS-II-654C>T Hetero-28A>G 2

Hetero IVS-II-654C>T Hetero-29A>G 1

Hetero IVS-II-654C>T Hetero HbE 1

Hetero CD17A>T Hetero CD41/42(-TTCT) 7

Hetero CD17A>T Hetero IVS-II-654C>T 3

Hetero CD17A>T Hetero-28A>G 2

Hetero CD17A>T Hetero HbE 2

Hetero-28A>G Hetero CD41/42(-TTCT) 10

Hetero-28A>G Hetero IVS-II-654C>T 6

Hetero-28A>G Hetero CD17A>T 1

Hetero-28A>G Hetero HbE 1

Hetero HbE Hetero IVS-II-654C>T 1

Hetero-28A>G/HbE Hetero IVS-II-654C>T 1

Hetero -28A>G/HbE Hetero CD41/42(-TTCT) 1


Parents carrying the same mutations

41

Hetero CD41/42(-TTCT) Hetero CD41/42(-TTCT) 22

Hetero IVS-II-654C>T Hetero IVS-II-654C>T 14

Hetero CD17A>T Hetero CD17A>T 5

Total 124


Table 3: Diagnostic accuracy for the three overlapping amplicons for each

mutation.

Mutation Amplicon 1 Amplicon 2 Amplicon 3 Totala

CD41/42(-TTCT) 81.5%

(22/27)

92.6%

(25/27)

81.5%

(22/27)

96.3%

(26/27)

IVS-II-654C>T 90.5%

(19/21)

90.5%

(19/21)

85.7%

(18/21)

95.2%

(20/21)

CD17A>T 92.9%

(13/14)

85.7%

(12/14)

100%

(14/14)

100%

(14/14)

HbE 100% (3/3) 100% (3/3) 100% (3/3) 100% (3/3)

-28A>G 90%

(18/20)

95%

(19/20)

95%

(19/20)

95%

(19/20)

aTotal accuracy refers to the number of samples classified correctly by at least

two out of three amplicons


Figure 1: The relative proportion of mutant paternal allele identified in the

samples where the paternal mutation has not been inherited. The result is an

average obtained from the three overlapping amplicons. The two cases labelled

with * were analysed for both the HbE and -28A>G mutations since the fathers

are compound heterozygotes. Three false positives are indicated by arrows.


Figure 2: The relative proportion of mutant paternal allele identified in the

samples where the paternal mutation has been inherited. The result is an

average obtained from the three overlapping amplicons. The two cases labelled

with * were analysed for both the HbE and -28A>G mutations since the fathers

are compound heterozygotes.

Documents

Non-invasive prenatal diagnostic testing for β-thalassaemia using cell-free fetal DNA and next generation sequencing