* Professor K. H. Nicolaides, Harris Birthright Research Centre for Fetal Medicine, King's College London School of Medicine, King's College Hospital, Denmark Hill, London SE5 9RS, UK.
Objective To investigate whether chorionic villus sampling (CVS) is associated with an increase in fetomaternal cell trafficking.
Design Prospective study.
Setting King's College London School of Medicine, King's College Hospital.
Sample Eighteen singleton pregnancies undergoing CVS for fetal karyotyping at 11–14 weeks of gestation and subsequently found to have chromosomal defects.
Method Maternal blood samples were obtained immediately before and at 3–14 (median 5) days after CVS. Fetal erythroblasts were isolated using triple density gradient separation and anti-CD71 magnetic cell sorting techniques. The enriched erythroblasts were stained with Kleihauer–Giemsa and with fluorescent antibodies for the epsilon (ɛ) and gamma (γ) globin chains. The percentage of fetal cells positive for each stain was calculated. Fluorescence in situ hybridisation (FISH) for X- and Y-chromosomes was also performed. Comparison was made in the proportion of enriched fetal cells between the pre-CVS and post-CVS samples.
Main outcome measures The proportion of fetal erythroblasts in maternal blood.
Results The percentage of erythroblasts enriched from maternal blood that stained positive for ɛ and γ globin chains and with Kleihauer–Giemsa was significantly higher in the post-CVS samples compared with the pre-CVS samples. FISH analysis for the Y-chromosome confirmed the increase in fetal cell proportion in the post-CVS samples. The percentage difference in fetal cells decreased significantly with time interval from CVS.
Conclusion CVS results in an increase in fetomaternal cell trafficking, which continues to be present for several days after the procedure.
Invasive intrauterine procedures, such as chorionic villus sampling (CVS), can result in fetomaternal haemorrhage and routine clinical practice is to give anti-D injections to Rhesus D negative pregnant women to prevent sensitisation. Previous studies have shown that CVS is associated with a rapid rise in maternal serum α-fetoprotein, human chorionic gonadotrophin (hCG), pregnancy-associated plasma protein A (PAPP-A), decidual pregnancy protein 12 (PP12) and Achwangerschafts protein 1 (SP1)1–3.
Fetal cells in the maternal circulation are present at a low concentration of one fetal cell in 103–108 maternal cells4,5. The proportion of these cells is higher in pregnancies with fetal chromosomal abnormalities6,7 and in certain conditions such as pre-eclampsia, fetal growth restriction and fetal anaemia8–11, which are thought to be due to the increase in the passage of these cells across the placenta. There is also some controversial evidence that invasive procedures may cause an increase in the number of fetal cells enriched from maternal blood. Gänshirt-Ahlert et al.12,13 used triple density gradient and MACS to determine the total number of enriched erythroblasts before and after invasive procedures and reported no significant differences. In contrast, Jansen et al.14, studied 19 pregnant women carrying male fetuses at 11–14 weeks of gestation, and reported that CVS was associated with an increase in the number of fetal cells, using Y-signals on fluorescence in situ hybridisation (FISH) in enriched erythroblasts from maternal blood, in 10 of the 19 male pregnancies. This cellular increase was accompanied by an elevation in maternal serum α-fetoprotein levels.
The aims of this study were, firstly, to investigate whether there is an increase in fetomaternal cell trafficking after iatrogenic invasion of the placenta with CVS, and secondly, if such an increase in fetal cells exists, to determine its magnitude and duration.
Triple density gradient separation and anti-CD71 magnetic cell sorting techniques were used to isolate fetal erythroblasts from maternal blood in 18 singleton chromosomally abnormal pregnancies in maternal blood samples obtained immediately before and at 3–14 (median 5) days after CVS.
We have previously reported on a study of blood samples obtained immediately before CVS for fetal karyotyping in women identified by nuchal translucency screening to be at high risk for chromosomal abnormalities6. In 18 of these pregnancies, which were subsequently found to have chromosomal defects and the parents elected to have termination of pregnancy, blood samples were also obtained just before the operation. In each case on both occasions, 20 mL of venous maternal blood was collected into lithium heparinised vacutainers (Beckton Dickinson, Franklin Lakes, New Jersey, USA), sorted at 4°C, and processed within 24 hours of collection.
Triple density gradient centrifugation was carried out as previously described and the middle layer containing the erythroblasts were separated and isolated6,8–10,15. Cells were incubated with magnetically labeled CD71 antibody to the transferrin receptor antigen (Miltenyi Biotech, Bergisch Gladbach, Germany) for 15 minutes at 4°C. Magnetic cell sorting was then performed to enrich these erythroblasts as previously described6,8–10,15. In each case, three aliquots were obtained from the positively selected cells of each pre-CVS and post-CVS sample. Cells were cytocentrifuged at 14.3 ×g for 10 minutes (Shandon, Frankfurt, Germany) and were cytospinned onto three slides. Six slides were analysed for each case (three pre- and three post-CVS). The fetal cells were detected and quantitated using two methods: firstly, morphologically after staining by the Kleihauer–Betke method (GTI, North Patrick Boulevard, Brookfield, Wisconsin, USA) and counterstaining with methylene blue (Gurr–Giemsa, BDH Merck, Poole, England), and secondly, by immunocytochemistry using monoclonal fluorescein isothiocyanate (FITC) conjugate fluorescent antibody against epsilon (ɛ) and gamma (γ) haemoglobin chains. Cells were fixed and permeabilised, as previously described15,16, using commercial ‘Fix and Perm’ reagents (Caltac, Burlingame, California). Slides were then washed in phosphate buffered saline (PBS) solution and incubated with monoclonal FITC conjugate fluorescent antibody for the ɛ and γ chains, respectively15,16 (Fig. 1). After antibody incubation, the slides were washed in PBS solution, mounted with 4,6-diamidino-2-phenylindole (DAPI) and visualised under a fluorescence microscope (Zeiss Axioskop microscope, Carl Zeiss, Gottingen, Germany). Nucleated cells that showed specific staining above the DAPI background stain were counted as positive. At least 100 nucleated cells were counted.
The remaining cells in the positive fractions were centrifuged, treated with KCl and fixed with methanol/glacial acetic acid. Cells were transferred to glass slides and FISH was carried out using a dual chromosome-specific DNA probes kit (Vysis, Downers Grove, Illinois, USA) to detect chromosomes X and Y (Fig. 1) as previously described8–10,15. The number of nucleated cells and Y-signal positive cells was calculated. At least 100 nucleated cells were examined on each slide and the percentage of cells with one signal for the Y-chromosome probe, and one, two and three signals for the X-chromosome probe was calculated. Only intact cells that were not overlapping were chosen for the analysis. The slides were examined under fluorescence microscope (Zeiss Axioskop microscope, Carl Zeiss), using DAPI/FITC/TRITC triple band pass filter set. Image capture and processing was by a Microsoft computerised system (Vysis). Enrichment of fetal cells and analysis were carried out without knowledge of the clinical details of the patients.
For morphology and immunocytochemistry, the number and percentage of positive cells was calculated out of the total nucleated cells. Comparison was made in the proportion of positive cells between the pre-CVS and post-CVS samples, using paired t test. Similarly, for FISH analysis, the total number of cells with positive signals for the Y- and X-chromosomes was calculated out of the total cell number present on the slides and the proportion of these positive signals was compared between the pre-CVS and post-CVS samples using the same statistical test above. The association between the different staining methods was determined by the Spearman correlation coefficient. The percentage of difference (percentage change) in fetal cell proportion between the pre-CVS and post-CVS samples was determined for all methods of staining. This was calculated from the equation: Percentage difference = (post − pre) ÷ pre × 100; 100% means the difference is 1, and 300% means the difference is three times. The relation between the percentage difference and the time interval between the pre-CVS and post-CVS samples was also calculated. The type of regression line was determined by the residual sum of squares.
In the 18 women studied, the median maternal age was 34 (range 19–43). The median gestation at CVS was 13 (range 11–14) weeks and at termination was 14 (range 12–15) weeks.
The proportion of enriched fetal erythroblasts was significantly higher in the post-CVS samples compared with the pre-CVS samples in all methods of fetal cells detection (Table 1, Fig. 2). CVS was associated with a threefold increase in fetomaternal cell trafficking with an overall median percentage difference between the post-CVS and pre-CVS samples of about 300 (range 0–421) (Table 1). The median percentages of positive erythroblasts for the Kleihauer–Giemsa stain, γ-chain and ɛ-chain in the pre-CVS samples were 5, 3 and 3, respectively, compared with 16, 12 and 9 in the post-CVS samples (Fig. 2). In 12 of the 18 cases, the fetuses were males and in these 12 pregnancies, the median percentage of cells positive for Y-signals was 5 in the pre-CVS samples, compared with 14.5 in the post-CVS samples (Fig. 2). In the six female pregnancies, no Y-signals were detected.
Table 1. Percentage of fetal erythroblasts enriched from maternal blood in the pre-CVS and post-CVS samples, identified by morphology, immunocytochemistry and FISH methods. Values are given in median (range).
Percentage of fetal erythroblasts enriched from maternal blood
Percentage difference between the pre- and post-CVS samples
Paired t test (P*)
*P values obtained from comparison between pre- and post-samples.
Kleihauer–Giemsa stain (n= positive cases)
5 (1–9) (n= 18)
16 (1–37) (n= 18)
γ-Chains (n= positive cases)
3 (1–8) (n= 18)
12 (1–31) (n= 18)
ɛ-Chains (n= positive cases)
3 (0.8–8) (n= 18)
9 (1–25) (n= 18)
Y-signals (12 cases) (n= positive cases)
5 (3–9.5) (n= 10)
14.5 (0.45–21) (n= 12)
There was a significant association between the Kleihauer–Giemsa stain and the two globin chains (pre-CVS, Kleihauer–Giemsa vsγ-chain, r= 0.85, P < 0.001; Kleihauer–Giemsa vsɛ-chain, r= 0.77, P < 0.001; post-CVS, Kleihauer–Giemsa vsγ-chain, r= 0.96, P < 0.001; Kleihauer–Giemsa vsɛ-chain, r= 0.62, P < 0.007). There was also a significant association between the Kleihauer–Giemsa and FISH for Y-signals in these pre-CVS and post-CVS samples (pre-CVS, r= 0.88, P < 0.001; post-CVS, r= 0.90, P < 0.001).
There was a linear relation between the time interval and the percentage difference between the pre- and post-CVS samples (Fig. 3). The percentage difference decreased significantly with time interval from CVS, with a large fall observed after six days.
The data of this study demonstrate that CVS is associated with a threefold increase in the proportion of fetal erythroblasts enriched from maternal blood. We found an increase in cells positive for the embryonic and fetal haemoglobin chains and an increase in the percentage of cells positive for Y-signals from pregnancies with male fetuses.
We previously reported the use of embryonic and fetal haemoglobins as fetal cell markers to detect the presence of fetal erythroblasts enriched from maternal blood in the first trimester of pregnancy and confirmed these findings by the presence of Y-signals on FISH in the male cases15. We have also shown that embryonic haemoglobin chains were absent in the non-pregnant female and male adults, providing strong evidence of the specificity of the staining and utility of this technique in identifying fetal cells by immunocytochemistry15. The findings of the current study therefore exclude the possibility that enriched erythroblasts that positively stained for embryonic haemoglobins are maternal in origin. The finding was also confirmed by the increase in Y-signals on FISH in the post-CVS samples of male cases.
Our findings are at variance to those of Gänshirt-Ahlert et al.12,13, who also used triple density gradient centrifugation and anti-CD71 magnetic cell sorting to enrich fetal cells in maternal blood and reported no significant difference in the number of erythroblasts between pre-CVS and post-CVS samples. In contrast, Jansen et al.14, who isolated fetal cells by Ficoll single density centrifugation and maternal depletion with CD45 and CD14 and CD71 magnetic cell sorting enrichment and then applied FISH for the Y-chromosome, from blood samples obtained immediately before and 5–20 minutes after CVS, found an increase in the number of fetal cells in 10 out of 19 male pregnancies examined. Possible reasons for the difference in findings in the studies of Gänshirt-Ahlert et al.12,13 are that firstly, the pre-CVS and post-CVS blood samples were obtained from two different groups of women; secondly, the range of gestational age of the pregnancies examined was not indicated, particularly in relation to the similarity of gestation of the two groups, and previous studies have shown a biological variation in the number of fetal cells present in maternal blood at different gestations17–19; and thirdly, the distribution of normal and abnormal karyotype in the pregnancies examined before and after CVS is not given and it is known that in trisomic pregnancies the number of fetal cells in maternal blood is increased6,7.
Unlike the study by Jansen et al.14, we obtained the post-CVS blood samples several days after the procedure, and we were able to show the extent of this placental leak and cellular transfusion and the time interval for which it continues. Interestingly, the increased difference in fetal cell proportion declined after about six days. Because the life span of erythroblasts is 60–70 days20, the most likely explanation for this sudden decline in the percentage difference is the closure of the temporary passage that was created by the procedure.
We have previously demonstrated an increase in fetomaternal cell trafficking in pregnancies complicated by pre-eclampsia, IUGR and fetal anaemia8–10. This study established that CVS is associated with an increase in fetomaternal cell trafficking, which continues to be present for several days after the procedure. This may constitute a good in vivo system to optimise fetal cell isolation procedures and to study fetal cell dynamics and characteristics.
The study was funded by the Fetal Medicine Foundation. Registered Charity No. 1037116.