Fetal granulocytes in maternal venous blood detected by in situ hybridization

Biochip for separating fetal cells from maternal circulation

Isolation of fetal cells from maternal circulation is the subject of intense research to eliminate the need for currently used invasive prenatal diagnosis procedures. Fetal cells can be isolated using magnetic-activated cell sorting or fluorescence-activated cell sorting, however no technique to specifically isolate and use fetal cells for genetic diagnosis has reached routine clinical practice. This paper demonstrates the use of a micromachined device to separate fetal cells from maternal circulation based on differences in size and deformation characteristics. Nucleated fetal red blood cells range in diameter from 9 to 12 microm can deform and pass through a channel as small as 2.5 microm wide and 5 microm deep. Although the white blood cells range in diameter from 10 to 20 microm, they cannot deform and are retained by the 2.5 microm wide and 5 microm deep channels under our experimental conditions. Fetal cells were isolated from cord blood and DNA analysis confirmed their fetal origin with ruled out maternal contamination.

The utility of an erythroblast scoring system and gender-independent short tandem repeat (STR) analysis for the detection of aneuploid fetal cells in maternal blood

OBJECTIVE

The aim of this study was to determine whether fetal nucleated red blood cells (NRBCs) could be distinguished from maternal cells in peripheral blood using an erythroblast scoring system based on the unique morphological and hemoglobin staining characteristics of this cell type. Presumptive fetal NRBCs were further analyzed for the presence of paternally inherited DNA polymorphisms to prove fetal origin.

METHODS

NRBCs were isolated by density gradient separation, CD15/45 depletion, and gamma hemoglobin positive selection from peripheral blood of nine women following termination of pregnancy for trisomy 21 (n=4), 18 (n=1), 13 (n=2), and other genetic abnormalities (n=2). Candidate fetal NRBCs, based on four discrete morphological and hemoglobin staining criteria, were then subjected to fluorescent PCR (polymerase chain reaction) amplification of chromosome 21 (D21S1411, D21S11) and chromosome 18 (D18S535) short tandem repeat (STR) DNA polymorphisms.

RESULTS

In all cases, candidate fetal NRBCs were accurately identified on the basis of morphologic and hemoglobin staining characteristics and confirmed to be fetal in origin based on the presence of shared and nonshared polymorphic DNA alleles when compared to DNA isolated from maternal cells.

CONCLUSIONS

Using the erythroblast scoring system and subsequent analysis of inherited DNA polymorphisms, we were able to distinguish fetal NRBCs from maternal cells and prove fetal origin independent of gender. These results suggest that this novel combined approach to fetal cell isolation and genetic analysis is a promising method for noninvasive prenatal diagnostic applications.

Non-invasive prenatal diagnosis: on the horizon?

The launch of the genomics and postgenomics era has greatly expanded our understanding of the genetic basis of many diseases. In conjunction with the sociocultural trend to delay childbirth and to maintain smaller family units, extra demand may be placed on the existing prenatal diagnostic services. The inherent risk of fetal loss associated with current prenatal diagnostic procedures, such as amniocentesis and chorionic villus sampling, has spurred research into non-invasive prenatal diagnosis. Much research has been conducted on the exploitation of fetal genetic material present in the maternal circulation. The initial focus was on the isolation of intact fetal cells and subsequently, the existence of extracellular fetal DNA in maternal plasma was realized. Exciting developments have been achieved in recent years. A large-scale trial to evaluate the clinical utility of fetal cell isolation from maternal blood for fetal aneuploidy diagnosis was launched and data were recently published. Much has taken place in the research of fetal DNA analysis in maternal plasma and in one example, namely prenatal RhD determination, this type of analysis has been used in the clinical setting. This paper reviews the technological developments in non-invasive prenatal diagnosis.

A simple and sensitive erythroblast scoring system to identify fetal cells in maternal blood

OBJECTIVES

Nucleated red blood cells (NRBCs) of fetal origin appear to have distinguishable characteristics from that of maternal NRBCs in both nuclear morphology and properties of hemoglobin staining. However, these differences have yet to be quantified. Our aim was to develop an erythroblast scoring system using four distinct phenotypic parameters (nuclear roundness, nuclear morphology, gamma hemoglobin staining intensity, and peripheral brightness of the stained cytoplasm) to address this issue.

METHODS

NRBCs were isolated from four maternal blood samples by density gradient separation, CD15/45 depletion, and gamma hemoglobin positive selection after elective termination of a trisomy 21 male fetus (47,XY,+21). All cells were deposited onto microscope slides and every NRBC was analyzed according to the scoring system. Each of the four individual parameters was given a value from 0 to 3 points and a combined score was obtained for each cell (range 0-12). Fluorescence in situ hybridization (FISH) using X- and Y-specific probes was performed to determine, on the basis of interphase karyotype, whether the cell was maternal or fetal.

RESULTS

The majority of maternal NRBCs were found to have a combined score of 6 or less (103/117) and the majority of fetal NRBCs were found to have a score of 7 or greater (43/53). The proportion of cells that were identified correctly as fetal increased with each ascending category of combined score. For example, 5.7% of NRBCs with a combined score of 5 points or less were found to be fetal, whereas 19.2% of NRBCs with a combined score of 6 points were fetal. At combined scores of 11 and 12 points, 100% of NRBCs were found to be fetal.

CONCLUSION

Fetal NRBCs have characteristic morphology and a gamma hemoglobin staining appearance that makes them distinguishable from maternal NRBCs. The scoring system presented here is a simple and sensitive method to distinguish fetal NRBCs from adult cells in maternal blood. This system may have clinical utility for noninvasive prenatal diagnosis as well as applications for basic research into the developmental biology of NRBCs. In addition, these defined parameters may serve as computational classifiers for the automated detection of fetal cells in maternal blood.

Kinetics of fetal cellular and cell-free DNA in the maternal circulation during and after pregnancy: implications for noninvasive prenatal diagnosis

BACKGROUND

Fetal genetic material is detectable in the maternal circulation and has been used for noninvasive prenatal diagnosis. However, few data are available concerning its quantity and natural history during gestation.

STUDY DESIGN AND METHODS

This study prospectively characterized the kinetics of cellular and cell-free fetal DNA in the circulation of 25 healthy women during and after uncomplicated pregnancy. Real-time kinetic PCR was used to quantitate human Y-chromosome sequences, and liquid oligomer hybridization with (32)P-labeled probes was used to verify the identity of amplified products.

RESULTS

In all male pregnancies, but no female pregnancies, low-level fetal Y-chromosome DNA was detected in both cellular and cell-free compartments beginning at 7 to 16 weeks but increasing steadily after 24 weeks and reaching a peak at parturition. The fetal DNA decreased rapidly after birth.

CONCLUSION

Fetal genetic material can be detected throughout pregnancy, and its quantity is a function of gestational age and of whether the plasma or cellular compartment is examined. Both the absolute quantity of fetal DNA and its ratio to total DNA (maternal + fetal) are greater in the plasma than in the cellular compartment. Fetal DNA is cleared rapidly from both compartments after parturition, which suggests that turnover is dynamic. Because they provide prospective and quantitative data concerning fetal DNA levels, these observations and kinetic PCR methods may have implications for noninvasive prenatal diagnosis. Further studies will be needed to determine the immunologic implications of fetal-maternal DNA exchange and cellular microchimerism.

[Fetal cells in the maternal blood and prenatal diagnosis]

Fetal loss after amniocentesis or chorionic villus sampling is a limit to prenatal diagnosis practice and to its generalization. The existence of fetal cells in the blood of pregnant women is now well established. Recognizing these cells with specific antibodies and isolating them with fluorescent or magnetic systems are the subject of numerous studies. However, to date, neither the sensitivity nor the specificity of these methods are sufficient to allow a non invasive prenatal diagnosis.

Trophoblastic cells expressing human chorionic gonadotropin genes in peripheral blood of patients with trophoblastic disease

We attempted to identify the cells expressing alpha and beta subunits of human chorionic gonadotropin (hCG) in the peripheral blood of patients with trophoblastic disease and normal pregnant women by using reverse transcriptase polymerase chain reaction (RT-PCR) and Southern blot. By this method, the mRNAs of hCG alpha and hCG beta were detected in the peripheral blood mononulear cells (PBMNC) from 3 of 7 hydatidiform mole (mole) and 1 of 4 choriocarcinoma patients as well as from normal pregnant women during the first trimester. None of the mRNAs of hCG subunits was detected in the PBMNC from healthy male and nonpregnant healthy women examined. The expression of hCG alpha and hCG beta in patients with trophoblastic disease and normal pregnant women almost correlated with their plasma levels of intact hCG. The present study indicates that the cells expressing hCG alpha and hCG beta, which virtually represent trophoblasts, are circulating in the peripheral blood of patients with trophoblastic disease as well as of normal pregnant women.

First trimester prenatal diagnosis: fetal cells in the maternal circulation

The recovery of fetal cells from the maternal circulation represents a promising approach to noninvasive prenatal diagnosis. Advances in techniques of sensitive molecular genetic analysis have enabled the conclusive demonstration of the presence of fetal cells in maternal blood. In most pregnancies, there are few fetal cells detectable. In some abnormal pregnancies, there appears to be increased fetomaternal transfusion, which facilitates recognition of aneuploid fetal cells. This review article describes general strategies of fetal cell isolation, current technical challenges, and clinical applications that are envisioned for the future. The increased appreciation of fetal cell microchimerism, and its association with complications of pregnancy and the postpartum development of autoimmune disease, is also discussed.

Transcervical recovery of fetal cells from the lower uterine pole: reliability of recovery and histological/immunocytochemical analysis of recovered cell populations

The aim of this work was to isolate, enumerate and attempt the identification of fetal cells recovered from the lower uterine pole. Immediately before elective termination of pregnancy at 7-17 weeks gestation, samples were recovered by transcervical flushing of the lower uterine pole (n = 108) or transcervical aspiration of mucus from just above the internal os (n = 187), and their contents examined using histological, immunohistochemical and molecular techniques. Syncytiotrophoblasts were identified morphologically in 28 out of 89 (31%) and 50 out of 180 (28%) flushings and aspirates respectively (mean 29%). Immunocytochemistry with monoclonal antibodies (mAbs) recognizing trophoblast or epithelial cell antigens on a smaller number of samples (n = 69) identified putative placental cells in 13 out of 19 (68%) and 25 out of 50 (50%) flushings and aspirates respectively (mean 55%). These included groups of distinctive cells with a small, round, hyperchromatic nucleus, strongly reactive with mAbs PLAP, NDOG1 and FT1.41.1. Smaller groups of larger, amorphous cells, usually containing multiple large, pale staining nuclei, reactive with mAb 340 and to a lesser degree with mAb NDOG5 were also observed. Taking cellular morphology and immunophenotype into consideration, the smaller uninucleate cells were likely to be villous mesenchymal cells, while the larger cells were possibly degrading villous syncytiotrophoblast. There was no significant difference in the frequency of fetal cells obtained by the two recovery methods. Squamous or columnar epithelial cells, labelled strongly with antibodies to cytokeratins or human milk fat globule protein, were observed in 97% (29 out of 30) of aspirates. The use of cervagem in a small number of patients prior to termination of pregnancy did not appear to influence the subsequent recovery of placental cells. Y-specific DNA was detected by polymerase chain reaction (PCR) in 13 out of 26 (50%) flushings and (99 out of 154) 64% aspirates analysed (mean 62%). In-situ hybridization (ISH) revealed Y-specific targets in 40 out of 69 (60%) of aspirates analysed. A comparison of PCR data obtained from transcervical recovered samples and placental tissues showed a concordance of 80% (76 out of 95), with 10 false positives. Comparing the PCR data from tissues with data derived by ISH from 41 aspirates gave a concordance of 90% with two false positives. Although syncytiotrophoblasts were much more likely to be present in samples containing immunoreactive placental cells, the detection rates of fetal-derived DNA were similar regardless of the morphological and/or immunological presence of placental cells. We conclude that the transcervical recovery of fetal cells, while promising, requires considerable additional effort being expended in further research and development, particular in the sampling procedure.

The application of individualised fetal growth curves

Individually adjusted or 'customised' growth charts aim to optimise the assessment of fetal growth by taking individual variation into account, and by projecting an optimal curve which delineates the potential weight gain in each pregnancy. This results in an increased detection rate of true growth restriction and a reduction in false positive diagnoses for IUGR. An adjustable standard can apply across geographical boundaries, as individual variation exceeds that between different maternity populations.

Fetal cells in maternal blood: the use of primed in situ (PRINS) labelling technique for fetal cell detection and sex assessment

Prenatal diagnosis is presently performed following invasive procedures with variable risks of fetal loss; non-invasive procedures using fetal cells in maternal blood would be welcome for the early detection of fetal sex or aneuploidy. We describe a simple and rapid protocol to detect fetal cells and thus to assess fetal sex. In a first step, nucleated blood cells were separated into mononuclear and polynuclear cells using a double density gradient centrifugation. In a second step primed in situ (PRINS) labelling technique was performed to label Y-chromosomes. 15 samples were studied and correct gender assignment was made in 13/15. The number of labelled nuclei was higher in polynuclear cell phases than in mononuclear cell phases. Moreover, the polylobular aspect of labelled nuclei from polynuclear cell phases strongly suggested that they could belong to fetal polynuclear cells. The PRINS technique combines some advantages of FISH, such as visual assessment of in situ chromosome labelling and the powerful specificity and sensitivity of PCR. In association with a simple enrichment procedure it constitutes a rapid protocol for fetal cell detection, non-invasive early prenatal sex assessment, and could further be applied to detect the main viable aneuploidies.

PCR quantitation of fetal cells in maternal blood in normal and aneuploid pregnancies

Fetal cells in maternal blood are a noninvasive source of fetal genetic material for prenatal diagnosis. We determined the number of fetal-cell DNA equivalents present in maternal whole-blood samples to deduce whether this number is affected by fetal karyotype. Peripheral blood samples were obtained from 199 women carrying chromosomally normal fetuses and from 31 women with male aneuploid fetuses. Male fetal-cell DNA-equivalent quantitation was determined by PCR amplification of a Y chromosome-specific sequence and was compared with PCR product amplified from known concentrations of male DNA run simultaneously. The mean number of male fetal-cell DNA equivalents detected in 16-ml blood samples from 90 women bearing a 46,XY fetus was 19 (range 0-91). The mean number of male fetal-cell DNA equivalents detected in 109 women bearing a 46,XX fetus was 2 (range 0-24). The mean number of male fetal-cell DNA equivalents detected when the fetus was male compared with when the fetus was female was highly significant (P = .0001). More fetal cells were detected in maternal blood when the fetus was aneuploid. The mean number of male fetal-cell DNA equivalents detected when the fetal karyotype was 47,XY,+21 was 110 (range 0.1-650), which was significantly higher than the number of male fetal-cell DNA equivalents detected in 46,XY fetuses (P = .0001). Feto-maternal transfusion of nucleated cells appears to be influenced by fetal karyotype. The sixfold elevation of fetal cells observed in maternal blood when the fetus had trisomy 21 indicates that noninvasive cytogenetic diagnosis of trisomy 21 should be feasible.

Prenatal diagnosis using fetal cells from the maternal circulation

Current prenatal diagnosis relies on invasive methods such as amniocentesis and chorionic villus sampling. Because these methods carry a low, but finite risk of pregnancy loss, noninvasive genetic screening techniques are the focus of intense research. Isolating fetal cells from maternal blood for genetic analysis is the least invasive method currently being investigated. We discuss the various methods that have been used to isolate these cells. Nucleated red blood cells have emerged as the ideal fetal cell type. This is because they have the DNA material necessary for genetic analysis, they are consistently present in maternal blood, they can be easily identified based on their morphology, and they have a definite gestational life span.

Maternal origin of transferrin receptor positive cells in venous blood of pregnant women

We studied the origin of transferrin receptor (CD71) positive cells in blood from seven women pregnant with a male fetus in order to explore if fetal cells could be detected among them. We used a technique that allows direct chromosomal analysis by in situ hybridization on immunologically and morphologically classified cells. Enrichment was performed by magnetic activated cell sorting (miniMACS) using an anti-CD71 monoclonal antibody. The cells were immunophenotyped by alkaline phosphatase anti-alkaline phosphatase immunostaining with the same antibody. The origin of the immunophenotyped cells was studied by in situ hybridization using an X cosmid Y repeat chromosome specific probe cocktail. CD71 positive cells were found in six of the seven women at the range of 4 to 43 in respective samples. Over 90% of the CD71 positive cells were nucleated erythrocytes. None of the detected positive cells were shown to be fetal. Thus, the use of transferrin receptor antigen alone in combination with the miniMACS may not be sufficient for enrichment of fetal cells.

Prenatal diagnosis by analysis of fetal cells in maternal blood

The data accumulated thus far indicate that fetal NRBCs are the target cell type of choice in maternal blood for most investigators, although some groups continue to work with the trophoblast. Reports of persistent circulation of hematopoietic stem cells, lymphoid/myeloid progenitors, and lymphocytes mandate that removal of these cell types must occur before clinical diagnosis of the current pregnancy can be made. In selected cases, accurate detection of fetal aneuploidy has been made from fetal cells in maternal blood; the clinical evaluation sponsored by the National Institute of Child Health and Human Development will determine the sensitivity and specificity of cytogenetic diagnosis in a larger group of pregnant women, but this information will not be available for several years. At present, detection of uniquely fetal, paternally inherited gene polymorphisms or mutations such as the Rh(D) antigen is possible only because the mother lacks these genes; hence, maternal cell contamination does not hinder diagnosis. Currently the presence of large numbers of maternal cells in enriched samples precludes single-gene diagnosis for conditions in which the mother carries a mutant gene, because her cells are preferentially amplified and difficult to distinguish from those of the fetus. It is likely, however, that as techniques of individual fetal cell isolation are perfected, maternal cell contamination will no longer be an issue, and the entire fetal genome will become available for diagnosis and therapy. Pediatricians need to be aware of the progress of research in this field, because fetal cell isolation from maternal blood not only could change prenatal diagnosis but would change the amount of genetic information that arrives with a newborn infant at birth. The ultimate goal of this work is to diagnose noninvasively, in the first trimester, the common fetal aneuploidies and single-gene disorders, to permit in utero treatment, or to allow low-risk pregnant women carrying an abnormal fetus an opportunity for reproductive choice.

Immunohistochemical characterization of cells retrieved by transcervical sampling in early pregnancy

Trophoblastic cells can be retrieved from the endocervix and the lower uterine segment in early pregnancy by aspiration or lavage (Rodeck et al., 1995). The feasibility of using this technique for prenatal diagnosis depends on how frequently fetal cells can be retrieved and whether such cells can be purified from the predominant maternal cell population. In this study, specimens retrieved from the lower uterine segment prior to elective first-trimester termination of pregnancy were examined histologically and characterized using a panel of monoclonal antibodies in an avidin-biotin-peroxidase technique. Lavage samples generally contained fewer cervical epithelial cells than aspirates. Syncytial fragments or cytotrophoblast were identified in 9 of 12 lavage samples but in only 4 of 10 aspirates. Trophoblast cells were reactive with various anti-trophoblast monoclonal antibodies but the trophoblast cells present displayed considerable antigenic heterogeneity. For positive selection of trophoblast cells from these samples, it is likely that the best yield will be achieved by using a panel of carefully characterized monoclonal antibodies directed against various villous and extravillous trophoblast populations.

Maternal origin of nucleated erythrocytes in peripheral venous blood of pregnant women

We studied the origin of nucleated red blood cells (NRBC) in peripheral venous blood samples from 40 pregnant women carrying a male fetus, using a technique that allows direct chromosomal analysis by in situ hybridisation on immunologically and morphologically classified cells. Samples from ten nulligravid women were studied as controls. NRBC were enriched by negative magnetic activated cell sorting (miniMACS) using anti-CD45 monoclonal antibody. NRBC were detected by alkaline phosphatase anti-alkaline phosphatase immunostaining using a monoclonal anti-glycophorin A antibody. The origin of the NRBC was determined by fluorescence in situ hybridisation using X and Y specific probes. NRBC were found in 37 of the 40 pregnant women at a range of 1 to 230 per 20 ml of venous blood and in 6 of the 10 controls at a range of 1 to 3 per 20 ml of venous blood. All NRBC detected in the pregnant women were evidently of maternal origin, and in the pregnant women the number of NRBC was significantly higher (P < 0.05) than in the controls. Pregnancy per se seems to induce the appearance of maternal NRBC in the circulation, and it cannot therefore be assumed that NRBC isolated from the maternal blood are of fetal origin on the basis of morphology alone. Discrimination of fetal NRBC must occur for prenatal diagnosis of fetal genetic disorders.

Fetal erythroblasts from maternal blood identified with 2,3-bisphosphoglycerate (BPG) and in situ hybridization (ISH) using Y-specific probes

Different types of fetal nucleated cells can be found in maternal blood, providing the possibility of non-invasive prenatal diagnosis. For this purpose, we have studied fetal erythroblasts. We discovered that haemoglobin-containing cells treated with 2,3-bisphosphoglycerate (BPG) can be visualized by a peroxidase reaction, which at the same time visualizes an in situ hybridization (ISH) signal, specific for the X, Y or 21 chromosome. In order to prove that the BPG-positive cells were erythroid, an anti-glycophorin A (GPA) antiserum combined with a staphylococcal rosette technique was used. To enrich for erythroblasts, leukocytes were depleted from maternal blood by treatment with anti-CD45 monoclonal antibody and passage over an anti-mouse IgG-coated glass bead column. To evaluate the potential of the method for clinical use, we studied maternal blood samples from 18 women referred to us for prenatal diagnosis between 6 and 19 weeks of gestation. Erythroblasts were found in 13 out of 14 normal pregnancies. Erythroblasts with a Y-signal were found as early as 9 weeks of gestation, but at 6 weeks the Y-signal was seen in BPG-negative cells only. These cells showed an epithelioid morphology indicating that they were cytotrophoblasts. The BPG-ISH method provides a simple technique for identifying erythroblasts and simultaneously visualizing a desired probe.

Mid-trimester fetal sex determination from maternal peripheral blood by fluorescence in situ hybridization without enrichment of fetal cells

To determine the fetal sex on 30 women who were 16-20 weeks pregnant, about 100,000 maternal blood nucleated cells were analysed by means of fluorescence in situ hybridization (FISH) with a Y-chromosome-specific DNA probe. Cells with the hybridization signal were detected in 12 of the 30 women. All the 12 mothers gave birth to a male child. Of the other 18 women who had no Y-positive cells in the peripheral blood, 14 gave birth to a female child and four gave birth to a male child. These false-negative results probably occurred because the number of cells examined was inadequate. The data obtained in this study suggest that fetal sex determination using maternal peripheral blood with FISH is possible and that this diagnostic method will be clinically useful when more cells are analysed.

Isolating fetal cells in maternal circulation for prenatal diagnosis

Fetal cells unequivocally exist in and can be isolated from maternal blood. Erythroblasts, trophoblasts, granulocytes and lymphocytes have all been isolated by various density gradient and flow sorting techniques. Chromosomal abnormalities detected on isolated fetal cells include trisomy 21, trisomy 18, Klinefelter syndrome (47,XXY) and 47,XYY. Polymerase chain reaction (PCR) technology has enabled the detection of fetal sex, Mendelian disorders (e.g. beta-globin mutations), HLA polymorphisms, and fetal Rhesus (D) blood type. The fetal cell type that has generated the most success is the nucleated erythrocyte; however, trophoblasts, lymphocytes and granulocytes are also considered to be present in maternal blood. Fetal cells circulate in maternal blood during the first and second trimesters, and their detection is probably not affected by Rh or ABO maternal-fetal incompatibilities. Emphasis is now directed toward determining the most practical and efficacious manner for this technique to be applied to prenatal genetic diagnosis. Only upon completion of clinical evaluations could it be considered appropriate to offer this technology as an alternative to conventional invasive and non-invasive methods of prenatal cytogenetic diagnosis.

Prenatal sex determination by in situ hybridization on fetal nucleated cells in maternal whole venous blood

Our aim was to evaluate whether the sex of a fetus could be determined in maternal whole venous blood by in situ hybridization without enrichment of fetal cells. This procedure is virtually without risks to the fetus or the mother. Blood samples were obtained from 59 women at different stages of pregnancy. Twenty preparations were discarded because they were technically unfit for in situ hybridization. Of the remaining 39 pregnant women, 18 had a male fetus, one had male twins, and 20 had a female fetus. Y-positive cells were detected in 12 of the 19 pregnancies with male fetuses and in two of the 20 pregnancies with a female fetus. The frequencies of cells with Y-signals ranged from 1 in 100,000 to 1 in 639. Our results show that fetal cells in maternal blood cannot be reliably used for prenatal diagnosis without prior enrichment of fetal cells.

ISH and PCR study with Y-specific probe/primers

Our main aim was to evaluate whether maternal whole venous blood could be used for determination of fetal sex, when no enrichment of fetal cells was attempted and when "standard" interphase cytogenetics and PCR analysis were adopted. Altogether 39 pregnant women were studied by using ISH and 59 by using PCR. Out of the 59 pregnant women, 26 carried a male fetus and 33 a female fetus. By ISH, Y-positive cells were detected in 12 of 19 pregnancies with a male fetus and in two of the 20 pregnancies with a female fetus. The frequency of the fetal cells ranged from 1 in 639 to 1 in 100,000. By nested PCR with primers flanking a Y-specific repeat sequence, the positive band indicating a male fetus was found in one of the 26 pregnancies with a male fetus and in one of the 33 pregnancies with a female fetus. According to our results, fetal cells in maternal blood cannot be reliably used for prenatal diagnosis without enrichment of fetal cells.

Clinical trials and experience: Boston

Our cumulative experience continues to validate the fetal nucleated erythrocyte as the target fetal cell type of choice, primarily because it reflects the cytogenetic status of the current pregnancy. Additional cell types, such as the granulocyte, await further study. Quantitative PCR is a sensitive and useful new method that can facilitate rapid comparisons between cell separation methods or different monoclonal antibodies. It can also be used on patient material to determine final purity of the enriched maternal samples. If the purity is too low, FISH studies will be complicated by the presence of thousands of maternal cells. Our planned studies include an analysis of why aneuploid pregnancies appear to have a higher number of fetal cells in the maternal circulation. We are also studying the timing of the fetomaternal transfer of cells with qPCR analysis of sorted maternal samples drawn weekly from well-dated women. We are continuously improving our methods (both in separations and antibodies) to reach a fetal cell purity of at least 20% for cytogenetic diagnosis by FISH studies. With the knowledge obtained thus far by us and by others, such a goal appears to be achievable within the near future.