Abstract
Objective
Noninvasive Prenatal Diagnosis has recently been introduced for a limited number of monogenetic disorders. However, the majority of DNA diagnostics still require fetal material obtained using an invasive test. Recently, a novel technique, TRIC (Trophoblast Retrieval and Isolation from the Cervix), has been described, which collects fetal trophoblast cells by endocervical sampling. Since this technique has not been successfully replicated by other groups, we aimed to achieve this in the current study.
Method
Pregnant women referred for transvaginal chorionic villous sampling (CVS) were asked for an endocervical sample prior to CVS. The TRIC samples were processed to isolate trophoblast DNA. TRIC DNA was used in ForenSeq to determine the amount of maternal DNA contamination, and for Sanger sequencing in case of a monogenic disorder.
Results
23%–44% of samples had a sufficiently high fetal DNA fraction to allow genetic testing, as calculated by Sanger sequencing and ForenSeq, respectively.
Conclusion
We have been able to successfully replicate the TRIC protocol, although with a much lower success rate as described by the original study performing TRIC. As we obtained the samples in the actual clinical setting envisioned, the method in its current setup is not advisable for use in prenatal diagnostics.
Key points
What's already known about this topic?
Trophoblasts can be retrieved from the cervix.
A previous publication has shown that fetal DNA can be derived from these cells.
What does this study add?
This study tries to replicate the original TRIC protocol to assess whether it is applicable for routine diagnostic testing.
Although we were able to replicate the method, the success rate was very low. We therefore conclude that the current TRIC protocol cannot be successfully used for routine genetic testing in early pregnancy.
1. INTRODUCTION
The introduction of noninvasive prenatal testing (NIPT) using massively parallel sequencing of fetal DNA in maternal blood has had a huge impact on worldwide clinical practice for prenatal screening, 1 allowing detection of trisomies with high sensitivity (99%) and specificity (99.9%). 2 Consequently, both pregnant women and their health care providers have greatly welcomed this technique. Although the possibilities of NIPT are expanding to noninvasive prenatal diagnosis (NIPD) for monogenic disorders, 3 there are still a large number of limitations since not every prenatal test situation is suited for this approach; the techniques are based on the fact that cell‐free fetal DNA fragments derived from trophoblast cells are circulating in maternal blood and use massively parallel sequencing, followed by complex bioinformatics for interpretation, to distinguish fetal from maternal haplotypes. 1 The fetal fraction must reach over 4% for accurate testing, which can be an issue in maternal obesity, for example. 4 Secondly, as false positives in chromosomal aberrations do occur due to confined placental mosaicism, confirmation of a positive NIPT test result by an invasive method, preferably amniocentesis, is advised. 5 The low fetal fraction, but also issues, such as low incidence, makes NIPT/NIPD highly challenging in detecting a birth defect of genetic origin other than trisomies. Moreover, in pregnancies with an overrepresentation of the fetal genotype in the maternal fraction of DNA, such as a male fetus at risk for an X‐linked disorder or a fetus at risk for an autosomal dominant disease that is affecting the mother, 6 there is currently no noninvasive diagnostic alternative available.
At the moment, prenatal diagnostic testing is performed mainly on invasively obtained fetal material by either amniocentesis (performed from 15 weeks onward) or chorionic villous sampling (CVS, performed from 11 weeks onward). 7 Both techniques have a (low) risk of miscarriage (around 0.1%) in women with a similar risk profile for chromosomal abnormalities. 8
It has been known for many years that, through a mechanism largely unknown, fetal cells with a trophoblast‐like phenotype are naturally shed from the conceptus into the reproductive tract. Many have tried to use these cells for prenatal diagnostic purposes but were always hampered by maternal cell co‐purification. Recently, an approach has been developed that appears to be simple, reliable, and safe and that can interrogate the intact fetal genome as early as 5 weeks of gestation using a Papanicolaou (Pap) smear. 9 The trophoblast retrieval and isolation from the cervix (TRIC) method has been developed to isolate these fetal cells in numbers sufficient for prenatal genetic testing. It has been shown that the cell numbers are unaffected by gestational age and maternal obesity. 10 Using this method, the original study group obtained DNA from an average of 700 cells with a fetal DNA fraction/purity of 92.2 ± 6.5% providing 100% correct fetal haplotyping at single nucleotide resolution. 11 In theory, this means that prenatal diagnostic testing on fetal DNA obtained using the TRIC technique can be performed using genotyping methods that are already established in clinical genetic labs and currently used for genotyping cells obtained using CVS, thus making this minimal‐invasive method an excellent addition to NIPD in fetuses at risk for a genetic birth defect. To date however, this technique including obtaining a high fetal DNA fraction has not been replicated in other research centers.
The objective of the current TRIC‐PT study was to replicate the original published study 11 using samples obtained prior to scheduled CVS. Additionally, as the CVS sample was obtained for prenatal diagnostic purposes, we also had the opportunity to compare the genetic test outcome of the TRIC sample to the diagnostic test outcome of the CVS sample.
2. METHODS
2.1. Sample collection
The Medical ethical committee of the Amsterdam Medical Center approved the TRIC‐PT study (METC #2017_291), and each participating patient provided written informed consent. For this study, pregnant women (n = 160), referred for transcervical CVS to the Fetal Medicine Unit at the Amsterdam UMC location AMC, were asked for an endocervical sample during their CVS procedure between 11 and 15 weeks of gestation. The endocervical sampling during CVS was done directly after inserting the speculum, before collection of the CVS sample. The endocervical sampling was done using a nylon cytobrush introduced approximately 2 cm into the endocervical canal and rotated 360° as it was withdrawn. The collected cells were immediately fixed using 20 ml CytoLyt solution (Hologic) and transported to the laboratory for further workup.
2.2. Isolation of endocervical fetal trophoblast cells
Trophoblast cells were isolated from each specimen using the published TRIC procedure. 11 Briefly, cells were centrifuged and resuspended in 10 ml of phosphate‐buffered saline (PBS) and washed three times. After a final resuspension in 1.5 ml of PBS, anti–HLA‐G–coated magnetic nanoparticles were added and incubated overnight at 4°C with mixing. Preparation of the HLA‐G‐coated nanoparticles was done beforehand by combining 100 μL sterile PBS, 10 μL mouse monoclonal anti‐HLA‐G antibody (Becton Dickinson #557577, 0.5 mg/ml), and 10 μL goat anti‐mouse IgG magnetic nanoparticles (Clemente). The nanoparticles were incubated overnight with mixing on a rocker at 4°C. The next day, the particles were placed on a magnet and washed 3 times, followed by resuspension in 100 μL PBS and subsequently added to the cells. The cell nanoparticle mixture was incubated overnight with mixing on a rocker at 4°C. The following day, the unbound (maternal) cells were collected after magnetic immobilization of HLA‐G–positive (fetal) cells. After three washings in PBS, the bound fetal cells were resuspended in 100 μL PBS and trophoblast numbers were estimated using a counting chamber.
2.3. Immunohistochemistry
Small aliquots of cells present in the endocervical sample were stained using standard immunohistochemistry procedures. Blocking was done in 5% bovine serum albumin (BSA), followed by overnight incubation with the mouse monoclonal anti‐HLA‐G antibody (Becton Dickinson #557577). Finally, Powervision Poly‐HRP secondary antibodies were used, followed by 3,3'‐diaminobe staining and counterstaining using hematoxylin.
2.4. Immunofluorescence
To investigate the successful isolation of fetal trophoblast‐like cells, an aliquot of randomly picked samples were spotted on microscope slides. These cells were blocked in 3% BSA, and the β‐hCG primary antibody (Thermo Scientific #MA1‐35020) was added overnight at 4°C. Finally, alexa‐fluor‐488 labeled secondary antibody was used following 4',6‐diamidino‐2‐phenylindole counterstain.
2.5. DNA extraction
DNA from maternal cells (the unbound fraction) was extracted using the Qiamp DNA micro kit (Qiagen) according to protocol. Fetal cells in PBS as obtained by TRIC were deposited on microscope slides and dried on a slide warmer at 40°C. Before DNA extraction, the slides were immersed in pepsin solution (0.01 g pepsin (Roche #10108057001) in 100 ml of 0.1 N HCl) for 11 min, followed by a PBS wash to remove cell membranes and contaminating maternal DNA fragments. The remaining nuclei on the slides were lysed by overnight incubation at 65°C with 0.5 μL per cell of Arcturus PicoPure DNA lysis buffer (Applied Biosystems) and then inactivated at 95°C for 30 min. Subsequently, the Qiamp DNA micro kit is used for genomic DNA cleanup. For samples with a high number of trophoblasts, DNA cleanup was done in multiple batches. DNA concentration of the fetal DNA was increased by glycogen‐mediated ethanol precipitation. In short, to each DNA sample, 1/10 volume (10 μL) of 3 M sodium acetate pH 5.2 was added, followed by 2.5 μL glycogen solution (Roche #10901393001, 20 mg/ml) and 2.5 volumes of ethanol. This was incubated overnight at −20°C. The next day, the samples were centrifuged for 10 min at 12,000 rpm, the supernatant was discarded, and the pellet was washed with cold 70% ethanol. After air drying, the pellet was dissolved in 10 μL nuclease‐free water.
2.6. ForenSeq
ForenSeq DNA library preparation and sequence runs were performed at the Genomics Core Facility of the Universitat Pompeu Fabra in Barcelona, Spain. Both maternal and 10 μL concentrated fetal TRIC DNA samples were prepared with the ForenSeq DNA Signature Prep Kit (Verogen) using primer mix A and sequenced on a MiSeq FGx system, including positive and negative controls (Verogen).
ForenSeq with primer mix A analyzes 94 single nucleotide variant (SNV) loci and 59 short tandem repeat (STR) loci, representing all human chromosomes. The ForenSeq Universal Analysis software provides a run quality report and detailed genotype. From this report, data were extracted, and allele percentages were determined using the read depth of each STR and SNV. For STR data, the two dominant alleles with the highest read depths were analyzed. For all fetal–maternal sample couples, each STR marker contained at least one identical STR, confirming the parent–offspring relation.
To determine the purity of the trophoblast DNA (named fetal fraction), the extent of contamination with maternal DNA was calculated from the SNV data. For each fetal–maternal sample couple, the homozygosity or heterozygosity of each SNV was determined, and the percentage deviation of the fetal DNA fraction was calculated. Maternal and fetal SNV profiles were then compared, and the non‐informative SNVs (identical maternal and fetal alleles) were filtered out. For the informative SNVs, the deviations of the fetal sample were multiplied by two (to account for two maternal alleles contributing to the fetal sample), yielding the percentage of maternal contamination.
2.7. PCR and Sanger sequencing
Polymerase chain reaction (PCR) primers covering the indicated mutations potentially carried by the TRIC samples were designed using Primer3 (www.primer3.org). Primer sequences can be found in Supplementary Table 1. Polymerase chain reaction reactions on 2 μL (or more when no PCR product was obtained) concentrated fetal TRIC DNA were performed using HOT FIREPol DNA polymerase (Solis BioDyne). Sanger sequencing of the purified PCR products was performed at the Genomics Core Facility of the Amsterdam UMC, location AMC. The results obtained using Sanger sequencing were subsequently compared to the diagnostic test outcome of the CVS sample.
3. RESULTS
3.1. Isolation of trophoblasts, and its DNA, from the cervix
We collected a total of 160 endocervical cell samples from women also undergoing CVS and performed the extraction and isolation of the trophoblasts as described by the original study group. 11 Staining of the endocervical samples with HLA‐G confirmed the presence of trophoblasts in this material, while staining with anti‐β‐hCG after the isolation procedure indeed confirmed that we isolated trophoblasts (Figure 1). We do not have complete certainty about the subtype of trophoblasts that we isolated; however, the fact that they express HLA‐G makes it likely they are of extravillous origin. In accordance with this, multiple studies have shown that β‐hCG is expressed by extravillous trophoblasts as well. 12 , 13 We successfully collected 154 cell samples, for 6 samples the isolation of cells failed. The indications for CVS varied: 36 samples were obtained as the NIPT screening was positive for a trisomy 13, 18 or 21; 68 samples were collected from pregnancies with ultrasound anomalies; 14 samples had indications for large chromosomal anomalies such as translocations, large deletions, or Turner syndrome; and 36 samples had an indication for a monogenic disorder (pathogenic SNV in the family). An overview of the samples collected is provided in Figure 2.The amount of isolated cells ranged from 100 to 5000 with an average of around 700 cells, similar as in the original study.
FIGURE 1.
Trophoblasts in and isolated from endocervical sampling stain positive for HLA‐G and β‐hCG. (A) Endocervical cells were stained with the HLA‐G antibody also used in the TRIC procedure. Only trophoblasts stain brown (black arrow) in a pool of maternal epithelial cells with a similar morphology (white arrows). (B) Isolated cells from the cervix using the TRIC method stain positive for β‐hCG (green), a trophoblast‐specific marker. Stainings on four different samples and a no primary antibody negative control are shown. 4',6‐diamidino‐2‐phenylindole (blue) staining was used to indicate the nuclei.
FIGURE 2.
Overview of the samples collected and the results obtained. Overview summarizing inclusions, chorionic villous sampling (CVS) indications, and results of ForenSeq and Sanger sequencing.
To isolate DNA from the trophoblasts, we initially followed the original protocol, that is, spotting the cells on microscope slides, followed by pepsin treatment and DNA isolation. Upon ForenSeq analysis, as described below, we could detect fetal DNA but also had multiple samples with too little DNA input to be analyzed by the ForenSeq platform, while other samples provided the maternal genotype only. We therefore did test a variety of adaptations to the DNA isolation part of the protocol to decrease maternal DNA contamination and increase fetal DNA yield; these included in‐tube DNase treatments, isolation of the nuclei using Nuclear Core Complex antibodies, and RepliG single cell whole genome amplification as suggested by Huang et al. 14 All of these adaptations were tested individually and in combination on at least 3 different TRIC samples, but all of them either yielded no DNA or maternal DNA only. We therefore discarded the potential adaptations and performed the study using DNA isolated using the original protocol only. As we needed DNA input in ForenSeq as well as monogenic PCR sequencing to be as high as possible, for most samples, we could only perform one of the two analyses.
3.2. ForenSeq reveals a TRIC success rate of 44 percent
In total, four different ForenSeq runs were performed, in which 53 samples were obtained by using the original protocol, and the analyzed ForenSeq data of these 53 samples are combined in Table 1. In ForenSeq, both maternal and fetal trophoblast DNA samples were analyzed to determine the amount of maternal contamination. Examples of a couple of SNV and STR results of three different samples with varying amounts of maternal contamination are provided as Supplementary data. In 10 samples, the amount of input DNA was too low to obtain sufficient reads for analysis. In the remaining 43 samples, the success rate, that is, a maternal contamination of ≤20% (as described to be allowed in the original study 11 ), of the TRIC method was 44% (19/43 samples). A summary of the ForenSeq results can also be found in Figure 2.
TABLE 1.
TRIC samples analyzed by ForenSeq
| TRIC‐PT | # Informative autosomal STRs a | Total # SNVs detected b | # Informative SNVs c | % Maternal contamination d | NB | |
|---|---|---|---|---|---|---|
| TRIC homozygous | Maternal homozygous | |||||
| 001 | 0 | 3 | 0 | 0 | 100% maternal | |
| 002 | 17 | 87 | 27 | 18 | 6 | |
| 003 | 1 | 6 | 0 | 1 | 8 | |
| 004 | 1 | 0 | 0 | 0 | Input too low | |
| 005 | 0 | 1 | 0 | 0 | Input too low | |
| 006 | 2 | 3 | 0 | 0 | Input too low | |
| 007 | 2 | 6 | 0 | 2 | 16 | |
| 008 | 0 | 0 | 0 | 0 | Input too low | |
| 009 | 0 | 0 | 0 | 0 | Input too low | |
| 010 | 0 | 0 | 0 | 0 | Input too low | |
| 022 | 2 | 70 | 7 | 1 | 11 | |
| 023 | 0 | 94 | 0 | 0 | 100% maternal | |
| 024 | 0 | 94 | 0 | 0 | 100% maternal | |
| 025 | 5 | 59 | 7 | 0 | 13 | |
| 026 | 8 | 51 | 8 | 1 | 12 | |
| 027 | 0 | 94 | 0 | 0 | 100% maternal | |
| 028 | 0 | 94 | 0 | 0 | 100% maternal | |
| 029 | 4 | 38 | 4 | 0 | 6 | |
| 030 | 0 | 94 | 0 | 0 | 100% maternal | |
| 031 | 5 | 59 | 6 | 1 | 10 | |
| 032 | 8 | 49 | 4 | 0 | 9 | |
| 033 | 2 | 78 | 2 | 2 | 22 | |
| 034 | 5 | 33 | 7 | 3 | 10 | |
| 035 | 2 | 78 | 4 | 0 | 9 | |
| 045 | 4 | 45 | 14 | 0 | 10 | |
| 049 | 0 | 5 | 1 | 0 | Input too low | |
| 054 | 5 | 27 | 12 | 0 | 7 | |
| 057 | 5 | 48 | 18 | 5 | 8 | |
| 059 | 0 | 94 | 0 | 0 | 100% maternal | |
| 065 | 0 | 92 | 7 | 1 | 30 | |
| 068 | 0 | 90 | 5 | 3 | 55 | |
| 070 | 5 | 88 | 11 | 7 | 36 | |
| 071 | 0 | 92 | 0 | 0 | 100% maternal | |
| 074 | 2 | 80 | 15 | 0 | 22 | |
| 080 | 0 | 89 | 3 | 0 | 41 | |
| 091 | 0 | 93 | 0 | 0 | 100% maternal | |
| 093 | 2 | 10 | 2 | 2 | 50 | |
| 102 | 0 | 94 | 0 | 0 | 100% maternal | |
| 116 | 0 | 86 | 0 | 0 | 100% maternal | |
| 119 | 2 | 92 | 4 | 0 | 44 | |
| 124 | 1 | 90 | 5 | 5 | 61 | |
| 135 | 0 | 88 | 0 | 0 | 100% maternal | |
| 137 | 0 | 90 | 0 | 0 | 100% maternal | |
| 139 | 1 | 15 | 1 | 0 | Input too low | |
| 141 | 4 | 70 | 6 | 3 | 31 | |
| 143 | 0 | 2 | 0 | 0 | Input too low | |
| 146 | 1 | 10 | 3 | 0 | 16 | |
| 149 | 9 | 42 | 13 | 0 | 4 | |
| 151 | 0 | 1 | 0 | 0 | Input too low | |
| 153 | 5 | 28 | 7 | 0 | 6 | |
| 156 | 0 | 80 | 1 | 0 | 18 | |
| 158 | 4 | 65 | 8 | 0 | 29 | |
| 159 | 2 | 27 | 11 | 0 | 13 | |
Note: In bold, the samples with a maternal contamination ≤20%.
Abbreviation: STR, short tandem repeat.
Informative STR: TRIC DNA has 2 alleles of which one differs from the allele carried by the maternal sample.
SNVs detected in TRIC DNA with a minimal sequencing depth of 50.
Informative SNV: TRIC DNA is heterozygous (minimal sequencing depth is 50) where maternal DNA is homozygous or TRIC DNA is homozygous (minimal sequencing depth is 100) where maternal DNA is heterozygous.
Percentage maternal contamination: Determined using informative SNVs only.
3.3. Monogenic single nucleotide variant indications can be confirmed in 23 percent of cases
In total, there were 36 samples with a monogenic SNV indication. We designed PCR primers to Sanger sequence all the chromosomal locations containing the different SNVs. Although some sequences failed, either due to the low amount of DNA input or due to multiple PCR products, we could successfully sequence the SNV locations in 30 samples. The results of the different successful PCRs (Table 2) show that in 7 samples (23%), the SNV carried by the fetus was correctly identified, while in 40% (12 samples), we most likely amplified the maternal genome. To complicate the results, in some sequence samples (e.g., TRIC‐PT088) it also appeared that allele dropout had occurred, that is, only one of the (maternal) alleles amplified, probably due to the low DNA input. Allele dropout might have also disturbed the results of the 7 samples with correctly identified homozygous or hemizygous fetal SNVs, and although some did contain other heterozygous variants, this was not the case for all. In contrast to ForenSeq, using Sanger sequencing, it is not possible to discriminate between the maternal and fetal genotype in case they carry the same SNVs, and thus, we are unsure about the origin of the genotype in 11 samples (37%). But it can be expected, based on the ForenSeq results, that some of these samples will reflect the maternal genotype and the sequence was not derived from fetal DNA. A summary of the monogenic SNV results can also be found in Figure 2.
TABLE 2.
TRIC samples analyzed using polymerase chain reaction and Sanger sequencing
| TRIC‐PT | SNV | Disorder | Inheritance pattern | Result TRIC | Result CVS | NB |
|---|---|---|---|---|---|---|
| 001 | WWOX c.550C > T | Epileptic encephalopathy | AR | Absent | Present HZ | |
| 002 | ALPL c.46_49del | Hypophosphatasia | AR | Present | Absent | |
| ALPL c.668G > A | Absent | Present | ||||
| 005 | CTSA c.1406C > A | Galactosialidosis | AR | Present | Present | |
| CTSA c.413T > C | Absent | Absent | ||||
| 013 | PEX1 c.2097insT | Zellweger spectrum disorder | AR | Present HZ | Present | |
| 015 | ETFDH c.1414G > A | Multiple acyl‐CoA dehydrogenase deficiency | AR | Absent | Present | |
| 023 | CFTR F508del | Cystic fibrosis | AR | Present | Present | Maternal genotype? |
| 027 | MED13L c.2606C > T | Asadollahi‐Rauch‐syndrome | AD | Absent | Absent | Maternal genotype? |
| 030 | NIPBL c.6893G > C | Cornelia de Lange syndrome | AD | Absent | Absent | Maternal genotype? |
| 035 | TCOF1 c.4218dupG | Treacher‐Collins syndrome | AD | Present | Absent | |
| 036 | MUSK c.1724T > C | Pena‐Shokeir syndrome type 1 | AR | Absent | Absent | |
| 037 | PEX7 c.875T > A | Rhizomelic chondrodysplasia punctata type 1 | AR | Present | Absent | |
| 039 | HEY2 c.318‐319delAG | Familial TAAD | AR | Present HZ | Present HZ | |
| 041 | UPB1 c.105‐2A > G | Beta‐ureidopropionase deficiency | AR | Absent | Absent | |
| 056 | HBB c.20A > T | Beta‐thalassemia | AR | Absent | Absent | |
| 060 | CHD7 c.5405‐17G > A | CHARGE syndrome | AD | Absent | Absent | Maternal genotype? |
| 066 | DYNC1H1 c.874C > T | Charcot‐Marie‐Tooth disease type 20 | AD | Absent | Absent | Maternal genotype? |
| 069 | CFTR c.579 + 1G > T | Cystic fibrosis | AR | Absent | Present | |
| 072 | LAMB3 c.1903C > T | Epidermolysis bullosa | AR | Present | Not HZ | Maternal genotype? |
| 077 | DCT c.125C > T | Oculocutaneous albinism type 8 | AR | Present | Present | Maternal genotype? |
| 079 | CDKL5 c.2578C > T | CDKL5 deficiency disorder | X | Absent | Absent | Maternal genotype? |
| 084 | LDLR c.2140 + 103G > T | Familial hypercholesterolemia | AR/AD | Absent | Present | |
| 085 | ASXL3 c.4403C > G | Bainbridge‐Ropers syndrome | AD | Absent | Absent | Maternal genotype? |
| 088 | FBN1 c.7477C > T | Marfan syndrome | AD | Present HZ | Absent | |
| 095 | NEXN c.1174C > T | Familial dilated cardiomyopathy | AD | Present | Unknown | Miscarriage |
| 114 | PCDH19 c.1902_1903insG | Dravet syndrome | AD | Absent | Absent | |
| 120 | LAMA2 c.498G > A | Congential muscular dystrophy type 1A | AR | Present | Present | |
| 122 | FKTN c.919C > T | Fukuyama type muscular dystrophy | AR | Present | Absent | |
| 126 | LDLR c.2140 + 103G > T | Familial hypercholesterolemia | AR/AD | Absent | Absent | Maternal genotype? |
| 144 | MTM1 c.49G > T | Centronuclear myopathy | X | Present HZ | Absent | |
| 147 | KCNB1 c.1041C > G | KCNB1 encephalopathy | AD | Absent | Absent | Maternal genotype? |
Note: In bold, the samples in which the genotype sequenced is similar to CVS and different from mother.
Abbreviations: AD, autosomal dominant; AR, autosomal recessive; HZ, homozygous or hemizygous; X, X‐linked.
4. DISCUSSION
In this study, we attempted to exactly replicate the TRIC method 11 using samples obtained from women undergoing CVS. This study population is identical to the population that would mostly benefit from this method as for these women, no minimal invasive alternative yet exists. As we can conclude from the results obtained from the ForenSeq analyses, as well as the monogenic SNV PCR sequencing, in our hands, the method was less successful as described in the original publication. In 44% of cases, we were able to achieve a high fetal fraction, that is, low level (max 20%) of maternal contamination, while in 23%, we were certain in correctly identifying the SNV carried by the unborn child.
It is not completely certain how the discrepancy between our study and the original study can be explained as exactly the same materials were used as in the original study. The most obvious difference however is that both the sampling and the extraction and isolation of the cells and DNA were performed in a different hospital and laboratory, respectively. Especially, the sampling before CVS might have had a substantial impact; the original study was performed in women undergoing a termination of pregnancy, while in our study, the majority of women were not considering a termination when sampling took place. This might have had an impact on the thoroughness of obtaining the endocervical samples. In our study, the clinicians performing the sampling were always careful not to disturb the endocervix too much as this could cause bleeding or even potentially increase the miscarriage risk. A potential increase in miscarriage risk was retrospectively examined and was found not to have occurred. Another recent study has also shown that taking a sample from the exocervix even yields 10 times less trophoblast‐like cells. 15 Although we retrieved similar amounts of trophoblast cells, we do assume that the most likely cause of our reduced success compared to the original study is due to more careful sampling. We suggest that collecting from the endocervix but closer to the exocervix, as done is this study, yields samples with more excreted maternal cell debris, which contains high amounts of free‐floating DNA. But since our samples were taken in the actual clinical setting as also envisaged for TRIC, we have to conclude that this method in its current setup does not yield consistent results to be added to the sampling options for prenatal diagnosis. In the past years, there have been other studies that tried to isolated cells from the cervix and/or replicated parts of the current method used, 16 , 17 but apart from initial (positive) preliminary results, no studies have been published in which the technique was sufficiently advanced to be implemented in prenatal diagnostics.
5. CONCLUSION
We have been unable to successfully replicate the results of the original study performing TRIC; although we did obtain fetal trophoblast DNA in 23–44% of the samples, the lack of control over the final fetal DNA yield in samples obtained in the actual clinical setting envisioned makes the method in its current setup not advisable for use in prenatal diagnostics.
CONFLICT OF INTEREST
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Supporting information
Supporting Information S1
Table S1
ACKNOWLEDGMENTS
The authors would like to thank all women that participated in the TRIC‐PT study and all personnel working at the prenatal diagnosis clinic at Amsterdam UMC. Funding for this research was provided by a grant from the Amsterdam Reproduction and Development Research Institute.
van Dijk M, Boussata S, Janssen D, et al. Tricky TRIC: a replication study using trophoblast retrieval and isolation from the cervix to study genetic birth defects. Prenat Diagn. 2022;42(13):1612‐1621. 10.1002/pd.6260
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
REFERENCES
- 1. Chitty LS, Bianchi DW. Noninvasive prenatal testing: the paradigm is shifting rapidly. Prenat Diagn. 2013;33(6):511‐513. 10.1002/pd.4136 [DOI] [PubMed] [Google Scholar]
- 2. Taylor‐Phillips S, Freeman K, Geppert J, et al. Accuracy of non‐invasive prenatal testing using cell‐free DNA for detection of Down, Edwards and Patau syndromes: a systematic review and meta‐analysis. BMJ Open. 2016;6(1):e010002. 10.1136/bmjopen-2015-010002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Scotchman E, Shaw J, Paternoster B, Chandler N, Chitty LS. Non‐invasive prenatal diagnosis and screening for monogenic disorders. Eur J Obstet Gynecol Reprod Biol. 2020;253:320‐327. 10.1016/j.ejogrb.2020.08.001 [DOI] [PubMed] [Google Scholar]
- 4. Wang E, Batey A, Struble C, Musci T, Song K, Oliphant A. Gestational age and maternal weight effects on fetal cell‐free DNA in maternal plasma. Prenat Diagn. 2013;33(7):662‐666. 10.1002/pd.4119 [DOI] [PubMed] [Google Scholar]
- 5. Committee Opinion No. 640: cell‐free DNA screening for fetal aneuploidy. Obstet Gynecol. 2015;126:e31‐e37. [DOI] [PubMed] [Google Scholar]
- 6. Wong FC, Lo YM. Prenatal diagnosis innovation: genome sequencing of maternal plasma. Annu Rev Med. 2016;67(1):419‐432. 10.1146/annurev-med-091014-115715 [DOI] [PubMed] [Google Scholar]
- 7. Wapner RJ. Invasive prenatal diagnostic techniques. Semin Perinatol. 2005;29(6):401‐404. 10.1053/j.semperi.2006.01.003 [DOI] [PubMed] [Google Scholar]
- 8. Salomon LJ, Sotiriadis A, Wulff CB, Odibo A, Akolekar R. Risk of miscarriage following amniocentesis or chorionic villus sampling: systematic review of literature and updated meta‐analysis. Ultrasound Obstet Gynecol. 2019;54(4):442‐451. 10.1002/uog.20353 [DOI] [PubMed] [Google Scholar]
- 9. Bolnick JM, Kilburn BA, Bajpayee S, et al. Trophoblast retrieval and isolation from the cervix (TRIC) for noninvasive prenatal screening at 5 to 20 weeks of gestation. Fertil Steril. 2014;102(1):135‐142.e6. 10.1016/j.fertnstert.2014.04.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Fritz R, Kohan‐Ghadr HR, Sacher A, et al. Trophoblast retrieval and isolation from the cervix (TRIC) is unaffected by early gestational age or maternal obesity. Prenat Diagn. 2015;35(12):1218‐1222. 10.1002/pd.4681 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Jain C, Kadam L, Van Dijk M, et al. Fetal genome profiling at five weeks of gestation after noninvasive isolation of trophoblast cells from the endocervical canal. Sci Transl Med. 2016;8(363):363re4. 10.1126/scitranslmed.aah4661 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Handschuh K, Guibourdenche J, Tsatsaris V, et al. Human chorionic gonadotropin produced by the invasive trophoblast but not the villous trophoblast promotes cell invasion and is down‐regulated by peroxisome proliferator‐activated receptor‐gamma. Endocrinology. 2007;148(10):5011‐5019. 10.1210/en.2007-0286 [DOI] [PubMed] [Google Scholar]
- 13. Eikmans M, van der Keur C, Anholts JDH, et al. Primary trophoblast cultures: characterization of HLA profiles and immune cell interactions. Front Immunol. 2022;13:814019. 10.3389/fimmu.2022.814019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Huang Y, Situ B, Huang L, et al. Nondestructive identification of rare trophoblastic cells by endoplasmic reticulum staining for noninvasive prenatal testing of monogenic diseases. Adv Sci. 2020;7:1903354. 10.1002/advs.201903354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Bourlard L, Manigart Y, Donner C, et al. Rarity of fetal cells in exocervical samples for noninvasive prenatal diagnosis. J Perinat Med. 2021;50(4):476‐485. 10.1515/jpm-2021-0291 [DOI] [PubMed] [Google Scholar]
- 16. Lee MJ, Kim SH, Shim SH, Jang HY, Park HJ, Cha DH. Optimization protocol of fixation method for trophoblast retrieval from the cervix (TRIC): a preliminary study. Diagnostics. 2020;10(5):300. 10.3390/diagnostics10050300 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Pfeifer I, Benachi A, Saker A, et al. Cervical trophoblasts for non‐invasive single‐cell genotyping and prenatal diagnosis. Placenta. 2016;37:56‐60. 10.1016/j.placenta.2015.11.002 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information S1
Table S1
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.