Jean-Marc Costa, Centre de Diagnostic Prénatal, American Hospital of Paris, 63 bd Victor Hugo, 92200 Neuilly sur-Seine, France. E-mail: firstname.lastname@example.org
Summary. Fetal RHD genotype determination is useful in the management of sensitized RhD-negative pregnant women. It can be ascertained early during pregnancy by chorionic villus sampling (CVS) or amniocentesis. However, these procedures are invasive, resulting both in an increased risk of fetal loss and in an increased severity of immunization due to fetomaternal haemorrhage. A reliable determination of RHD genotype by fetal DNA analysis in maternal serum during the first trimester of pregnancy is reported in this study. One hundred and six sera from RhD-negative pregnant women were obtained during the first trimester of pregnancy. These sera were tested for the presence of RHD gene using a new real-time polymerase chain reaction assay and the results compared with those obtained later in pregnancy on amniotic fluid cells and by RHD serology of the new-born. All sera from women carrying a RhD-positive fetus (n = 62) gave positive results for RHD gene detection and sera from women carrying a RhD-negative fetus (n = 40) were negative. The high level of accuracy of fetal RHD genotyping obtained in this study could enable this technique to be offered on a routine basis for the management of RhD-negative patients during the first trimester of pregnancy.
Many studies have recently demonstrated that RHD status could be accurately determined by molecular study at the DNA level (Aubin et al, 1997; Maaskant-vanWijk et al, 1998). For clinical purposes, fetal RHD genotyping using polymerase chain reaction (PCR) is a significant advance. Indeed, fetal RHD genotype determination early during pregnancy is useful in the management of sensitized RhD-negative pregnant women. Fetal RHD status can be ascertained after chorionic villus sampling (CVS) or amniocentesis (Fisk et al, 1994; Crombach et al, 1997, 1999; Chan et al, 2001). However, these sampling procedures are invasive, resulting both in an increased risk of fetal loss and in an increased severity of immunization due to fetomaternal haemorrhage (Urbaniak, 1998). Fetal RHD genotyping is also useful in RhD-negative pregnant women at risk for RHD immunization (American College of Obstetrics & Gynecology (ACOG) practice bulletin, 1999), in order to adapt prophylactic anti-D immunoglobulin infusion to avoid unnecessary administration in case of a RhD-negative fetus. Unfortunately, this procedure cannot be applied to those RhD-negative pregnant women when fetal tissue is not sampled.
A reliable non-invasive approach to determine fetal RHD genotype would be valuable because it can be offered to the general RhD-negative obstetric population, whether or not fetal tissue sampling is performed. Furthermore, this approach would be especially advantageous during the first trimester of pregnancy to plan for further investigation or treatment in obstetrical situations such as miscarriage, antenatal haemorrhage or pregnancy termination.
Recovery of fetal cells from maternal blood or cervical mucus have been reported, but questions regarding the sensitivity and cost-effectiveness are still unresolved for routine use (Adinolfi & Sherlock, 1997; Pertl & Bianchi, 1999; Daryani et al, 2000; Holzgreve & Hahn, 2000). Alternatively, an RNA-based assay on fetal erythroblasts isolated from maternal blood was also evaluated. It was suggested that the presence of multiple copies of RhD-specific mRNA per cell could enhance the sensitivity of fetal RhD genotyping (Al-Mufti et al, 1998; Cunningham et al, 1999), but results were not conclusive. On the other hand, the amount of fetal nucleic acids in maternal plasma or serum is sufficient to enable analysis without any complex procedure, in contrast to the use of fetal cells, which require isolation and enrichment (Lo et al, 1998a). Fetal RHD genotyping has thus been successfully achieved by PCR on maternal plasma (Lo et al, 1998b; Zhang et al, 2000; Zhong et al, 2000). However, the described protocols never reached 100% sensitivity during the first trimester of pregnancy. In order to improve management of pregnant women carriers for an X-linked disorder, a real-time PCR assay was recently proposed, sensitive enough to achieve successful fetal sex determination early in pregnancy using maternal serum analysis (Costa et al, 2001). We have used a similar approach to determine the fetal RHD genotype in this first large study, focusing on the first trimester of pregnancy.
Patients and methods
Patients and samples. One hundred and six RhD-negative pregnant women were recruited from the Centre de Diagnostic Prénatal of the American Hospital of Paris and from the Maternité, Hôpital Notre-Dame de Bon Secours. These patients underwent genetic counselling either before prenatal diagnosis of fetal genetic disorders and chromosomal abnormality, or before first trimester screening of Down's syndrome, using fetal nuchal translucency study and maternal serum biochemical markers.
Maternal blood, after signed consent, was obtained during the first trimester of pregnancy, prior to amniocentesis or CVS if indicated. The mean gestational age was 12·7 weeks (range: 8–14).
Blood (5 ml) was collected into Vacutainer SST® tubes (Becton Dickinson, Meylan, France). Immediately after clotting, serum was obtained by centrifugation at 4°C for 10 min at 3000 g, aliquoted and stored at −80°C. If not treated immediately, the blood sample was kept at room temperature for a maximum of 24 h before processing. An additional 5 ml of blood were collected on EDTA anticoagulant, aliquoted and stored at −30°C.
If amniocentesis was performed later during pregnancy (n = 91), 2 ml of amniotic fluid was centrifuged for 10 min at 10 000 g and the supernatant was discarded. The pellet was then stored at −30°C until further processing.
Duplex real-time PCR for the RHD gene in maternal blood and amniotic fluid. Total DNA was extracted from 200 μl of total EDTA blood and from the resuspended (in 200 µl of NaCl 154 mmol/l) amniotic fluid cell pellet by the PCR template preparation kit (Roche Biochemicals, Meylan, France). In order to control the amplifiability of DNA extracted from maternal blood and amniotic fluid, both the RHD gene and a second autosomal reference gene, the superoxide dismutase (SOD) gene, were co-amplified by duplex real-time PCR in a LightCycler® instrument (Roche Biochemicals). Each of the PCR products was simultaneously and specifically detected using hybridization probes, labelled either with LCRed640 (for the RHD gene) or LCRed705 (for the SOD gene). Primers (MWG-Biotech, Courtaboeuf, France) and probes (TibMolBiol, Berlin, Germany) are described in Table I. PCR reactions were set up in a final volume of 20 µl using the Fast DNA master hybridization probes kit (Roche Biochemicals), with 0·5 mmol/l of each primer, 0·25 μmol/l of each probe, 1·25 units of uracil DNA glycosylase (UDG) (Biolabs, Saint-Quentin en Yvelines, France), 4·5 mmol/l of magnesium chloride and 5 µl of extracted DNA. After an initial 1 min incubation at 50°C to allow UDG to cleave putative contaminant PCR products from previous reactions, a first denaturation step of 8 min at 95°C was followed by amplification, performed for 40 cycles of denaturation (95°C, 10 s, ramping rate 20°C/s), annealing (56°C, 10 s, ramping rate 20°C/s) and extension (72°C, 15 s, ramping rate 2°C/s). During each run, 5 ng of RhD-positive DNA was used as a positive control and elution buffer for DNA extraction as a negative control.
Table I. Characteristics and sequences of primers and probes used in the PCR assays.
Real-time PCR for the RHD gene in maternal serum. The procedure was similar to that previously described for fetal sex determination (Costa et al, 2001) except that primers (MWG-Biotech) and probes (TibMolBiol) were targeted at the RHD gene (Table I). Briefly, as a tracer for the DNA extraction and amplification steps, a low amount (250 pg) of mouse DNA (Sigma, Grenoble, France) was added to each patient's sample (400 μl of serum) immediately prior to DNA extraction. Total DNA was then extracted by the PCR Template Preparation Kit (Roche Biochemicals) according to the manufacturer's recommendations, except that adsorbed DNA was eluted from the column with 50 µl instead of 200 µl of elution buffer. Amplification was carried out in a LightCycler® instrument as previously described (Costa et al, 2001). Each sample was analysed in duplicate. During each run, a low amount of DNA (10 pg) isolated from a RhD-positive patient was used as a positive control and the elution buffer for DNA extraction as a negative control. In case of a negative result for RHD gene amplification, the presence and correct amplification of the tracer mouse DNA in the extracted maternal serum DNA was checked. The same extracted serum DNA used for RHD gene amplification was assayed in a second PCR targeted at the mouse galactose-1-phosphate uridyltransferase (GALT) gene. This step enabled the determination of a true-negative from a false-negative result (lack of internal control) because of a deficient DNA extraction process or the presence of PCR inhibitors in the eluted DNA. If no amplification signal was obtained for the tracer DNA, the result was considered inconclusive and the assay repeated.
All sera were blind tested and the results compared with those obtained later in pregnancy on amniotic fluid cells and by RHD serology of the new-born.
Specificity of RhD genotyping using the using the RHD/SOD duplex real-time PCR was preliminary demonstrated on 101 blood samples of individuals who underwent blood group typing. All Rh-D positive (n = 91) and RhD-negative (n = 10) were correctly genotyped. A sample was considered positive when a fluorescent signal was detected for both the RHD and SOD genes. Efficiencies of both RHD and SOD PCR being equivalent, the time of appearance of fluorescence, defined as the crossing point (Cp), was similar for the two genes in the case of RHD positive samples. A sample was declared RHD negative when a signal was detected for the SOD gene only. Using this assay, no RHD gene was detected in leucocyte DNA of the patients included in this study. This result confirmed that the RhD-negative phenotype of all patients was related to a homozygous deletion of RHD gene and that the results obtained for fetal DNA analysis on maternal serum were not influenced by maternal RHD genotype.
The concordance of newborn RHD status determined by serology and fetal RHD genotyping on maternal serum was analysed in 102 of the 106 patients recruited to this study. In four patients, newborn serology could not be assayed because pregnancy termination in relation to fetal chromosomal abnormality or genetic disease was performed. All sera from women carrying a RhD-positive fetus (n = 62) gave positive results for RHD gene detection whereas all sera from women carrying a RhD-negative fetus (n = 40) gave negative results. Figure 1 illustrates the results obtained on a positive and a negative maternal serum for RHD gene amplification. The RHD gene assay was sensitive enough to detect about 10 copies of fetal DNA per ml of maternal serum but the sensitivity was lowered if a duplex PCR for both RHD gene and mouse GALT gene was performed. Therefore, all sera samples were analysed by a RHD monoplex assay. As expected, no amplification was observed with mouse DNA using the RHD gene set of primers, attesting to the specificity of the assay for the human RHD gene.
Among the 106 pregnant women tested for RHD gene presence in their serum, 77 underwent amniocentesis later in pregnancy and the amniotic fluid was analysed for the presence of the RHD gene; results for RHD gene detection in these samples were in complete agreement with those obtained with the corresponding sera, including the four patients where pregnancy was terminated. All samples were analysed and neither false-negative nor false-positive results were observed. Results obtained on a positive and a negative amniotic fluid for RHD gene amplification using duplex real-time PCR are presented in Fig 2.
Fetal RHD genotype was determined by maternal serum analysis during the first trimester of pregnancy. The assay was based on a similar real-time PCR approach to the one previously used to determine fetal sex in order to improve management of pregnant women carriers for an X-linked disorder (Costa et al, 2001). This PCR assay was applied in a blinded study on the sera from 106 RhD-negative pregnant women recruited from two different prenatal diagnosis centres. Results obtained on maternal serum were in complete concordance with those obtained on fetal cells isolated from amniotic fluid and, above all, with the RHD serotype of the new-born. Thus, this study demonstrated that a reliable fetal RHD genotype determination could be achieved with 100% sensitivity and specificity in maternal serum during the first trimester of pregnancy.
Determination of fetal RHD genotype using a non-invasive approach is an important challenge as it has crucial implications in the clinical management of sensitized RHD-negative pregnant women. Fetal DNA analysis in maternal serum offers opportunity to achieve this challenge. Unfortunately, published reports to date show a lack of sensitivity during the first trimester of pregnancy (Lo et al, 1998b). This lack of sensitivity may be a result of poor amplification efficiency. As previously demonstrated (Costa et al, 2001), the different real-time PCR assay used here (FRET® instead of Taqman® technology), achieved a high level of sensitivity associated with a high level of safety as it is a closed-tube system (Heid et al, 1996; Wittwer et al, 1997); no false-positive results related to PCR product carry-over were observed. Poor efficiency of DNA recovery or an inhibitory effect of DNA extract may also explain a low sensitivity. As a control for amplifiability, the use of an endogenous gene (such as β-globin gene) seems inappropriate (Lo et al, 1998b) as it evaluates mainly maternal DNA which is present in large amounts in the plasma. The addition of a low amount of a heterologous DNA as internal control is more sensitive to monitor both extraction and amplification steps. The use of this internal control is highly important because technical error could not be excluded. Finally, the use of maternal serum instead of plasma used by Lo et al (1998b) may also explain these discrepancies.
Typical errors of genotyping may be caused by rare variants of the RHD gene as 3′D gene fragments could be present in the genome of some individuals serotyped as RhD negative (Daniels et al, 1997; Okuda et al, 1997; Wagner et al 2001). Therefore, it has been suggested that different regions of the RHD gene should be examined to increase the accuracy of RHD genotyping (Maaskant-vanWijk et al, 1998). Although the assay described in this study was only targeted at the 3′untranslated region specific of the RHD gene (exon 10), no false-positive or negative results were observed when compared with newborn RHD serotype. This complete concordance may be the result of newly designed primers and probes. As suggested by Aubin et al (1997), the shortest PCR fragments in the 3′RHD gene can lead to more accurate results. Furthermore, other groups have already successfully determined fetal RHD genotype using a single RHD gene region assay (Lo et al, 1998b; Zhang et al, 2000).
Although a false-positive RHD genotype relative to a RHD gene variant cannot be excluded, such a situation is exceptional, particularly in the Caucasian population (Wagner et al, 2001), and without any medical consequence except an acceptable, if unnecessary, immunoglobulin injection.
The high level of accuracy of fetal RHD genotyping obtained in this study could enable this technique to be offered on a routine basis for the management of RhD-negative patients during the first trimester of pregnancy. This strategy offers many advantages.
The potential risk for maternal infection from contaminated anti-D immunoglobulin has to be taken into consideration and women should be informed about the risk involved (Saha, 1998). Even if an anti-D product was safe regarding viral transmission, the risk could not be formally excluded (Yap, 1996; Budka, 2000).
Moreover, as unnecessary infusion of anti-D immunoglobulin would be avoided, this strategy could contribute to saving this limited product until clonal anti-D becomes available. Furthermore, no further biological or clinical investigations would be performed in case of a RhD-negative fetus, thus contributing to lower maternal anxiety.
Finally, the fact that the assay is proposed during the first trimester is of interest because several clinical events may cause RHD alloimmunization very early in pregnancy such as therapeutic or spontaneous abortions and antenatal bleeding. Knowledge of fetal RHD genotype will thus be instrumental in the adequate management of these obstetrical situations.
The first trimester of pregnancy is clearly a critical period for prenatal diagnosis using non-invasive procedures (Nicolaides et al 2000). The assay here described offers the opportunity to combine systematic analysis of fetal RHD genotyping in RhD-negative women in conjunction with fetal nuchal translucency and first trimester screening of Down's syndrome by serum biochemical markers.
We are indebted to Dr Lavergne and Dr Benachi for reviewing this manuscript.