- Email: [email protected]
MORPHOLOGY OF RAPIDLY ADHERING AMNIOTIC-FLUID CELLS AS AN AID TO THE DIAGNOSIS OF NEURAL-TUBE DEFECTS
Saturday MORPHOLOGY OF RAPIDLY ADHERING
AMNIOTIC-FLUID CELLS AS AN AID TO THE DIAGNOSIS OF NEURAL-TUBE DEFECTS CHRISTINE M. GOSDEN M.R.C. Clinical and Population Cytogenetics Unit, Western General Hospital, Edinburgh EH4 2XU
J. H. BROCK Department of Human Genetics, University of Edinburgh, Western General Hospital, Edinburgh EH4 2XU D.
In 20 amniotic-fluid samples taken in the second trimester from pregnancies in which the fetus had a neural-tube defect, the proportion of the total viable cells which adhered to glass surfaces after 20 hours in culture ranged from 9 to 100%. In 92 normal amniotic fluids this proportion was less than 6%. Furthermore, the morphology of the rapidly adhering cells was characteristic in spina bifida (8 cases) and anencephaly (12 cases) and distinct from the epithelioidlike cells seen in normal amniotic fluids, including many which were grossly blood-stained. A sample of amniotic fluid from fetal exomphalos and one from a pregnancy in which the placenta had been repeatedly traversed during amniocentesis had proportions of adherent cells in the pathological range, but the morphologies were very different from those seen in neural-tube defects. It is suggested that the techniques described here will be useful adjuncts to amniotic-fluid alpha-fetoprotein determination in the early diagnosis of fetal abnormality, particularly in blood-stained samples.
Summary
Introduction THE early prenatal diagnosis of neural-tube defects through measurement of amniotic-fluid alpha-fetoprotein (A.F.P.) concentrations is well established and widely
used.’ It has three main limitations. The first is in the
diagnosis of some cases of spina bifida, where potentially disabling lesions may be too small to cause more than a marginal elevation in the amniotic-fluid A.F.P. concentration.2 Physical measurements, such as ultrasonar scan or amniography, are usually not helpful in confirming or excluding these abnormalities.3 The second is in elevated A.F.P. values associated with a fetal neural-tube defect from those caused by less serious disorders, such as exomphalospilonidal sinus,5 or duodenal atresia.6 The third problem arises when the amniotic fluid is contaminated by fetal blood, since fetal serum-A.F.p. concentrations are about 150 times those in the amniotic fluid at a corresponding gestation.7 Since A.F.P. determination is now applied routinely to most amniotic-fluid samples, whatever the reason for
distinguishing
30
April 1977
amniocentesis, the possibility of an increasing number of misdiagnoses has become serious." An additional amniotic-fluid parameter which can be used as an adjunct to A.F.P. is urgently needed. Some years ago Sutherland et al. observed that amniotic fluids from pregnancies with fetal neural-tube defects had a high proportion of
rapidly adhering cells which were fetal in origin and had macrophage-like properties. The original "macrophage"-counting technique had serious limitations1o and has not been widely used. We have now modified this procedure and extended it to include a quantitative estimate of the relative proportions of seven types of rapidly adhering cells. The method has promise in confirming A.F.P.-dependent diagnoses of neural-tube defects and in distinguishing these defects from other fetal abnormalities and from fluids contaminated with fetal blood.
Materials and Methods Amniotic Fluid
Amniotic fluids
were
obtained
at
15 to 22 weeks’
by transabdominal amniocentesis, because high values indicated
a
risk of the fetus
having a
gestation
serum-A.F.p.
neural-tube defect.
A.F.P. Determination
concentrations were determined by rocket immunothe method previously described." The upper limit of normal for each week of gestation has already been defined. 12 A.F.P.
electrophoresis by
Determination
of Cell Numbers
Amniotic-fluid cells
were
stained with
trypan-blue diluted
1/5 with amniotic fluids." This mixture was incubated at 37°C for 5 minutes and then counted in Neubauer countingchambers. The total cell number, viable-cell count, and number of red blood cells were recorded, and the cell numbers for total, viable, and blood cells per millilitre of amniotic fluid were calculated. Smears were made from each blood-stained liquor, and these were stained by the Shepherd and Kleihauer methods to determine the proportion of maternal and fetal blood cells.
Plating for 20-hour Cellular Adherence 0.5ml of amniotic fluid was added to 4-5ml of Ham’s F10 medium buffered with Hepes and supplemented with 25% fetal-calf serum, 50 units/ml penicillin, and 50 g/ml streptomycin. Enough cells were taken so that at least 2000 viable cells were available for coverslip counting. 2.5ml of this mixture was added to each of two 50 mm sterile Falcon plastic petri dishes each containing 3 glass coverslips, 19 mm x 35 mm. The dishes were incubated at 37°C for 20 hours without disturbance. The covers lips were removed, rinsed with phosphate-buffered saline, fixed, and stained with one of the following : May-Grunwald-Giemsa, peroxidase, haematoxylm and eosin, cresyl-fast-violet, or gallocyanin. Each coverslip was mounted, and the total number of cells adherent to each coverslip was counted. The area of each coverslip was mea8018
c
920 sured as a proportion of the total area of each dish, and thus the number of viable cells plated per unit area could be calculated. The morphology and staining characteristics of each adherent cell were examined, and a differential count was obtained. The adherence in 20 hours (A2oh) was given by the average number of cells adherent per coverslip expressed as a percentage of the viable cells plated per coverslip.
Classification of Cell Types 1. Long bipolar cells.-These are very extended spindleshaped cells, some of which widen into bulbous ends with pseudopodia or lamelliform membranes. The cytoplasm, especially in the central wide area near the nucleus, often shows prominent vacuolation (see figure, a). 2. Cells with multiple filamentous pseudopodia (processes).-These cells vary in size and may have prominent distinctive feature is the presence of 3-9 filamentous pseudopodia, some of which may be very long (see vacuoles. Their
most
figure, c).
3. Large vacuolated cells.-These resemble amoebocytes derived from blood monocytes or tissue macrophages. They are large and irregular in shape and may adopt a variety of forms. They are usually isolated, although some may be clustered together, and they have prominent vacuoles and may have darkly staining cytoplasmic inclusions (see figure, d). 4. Giant multinucleate cells.-These cells range in size from those containing two nuclei to those with as many as 10-15 nuclei. These cells sometimes have vacuolated cytoplasm
(figure, b). 5. Long triangularfibroblastic cells.-These cells seem to be
polarised and are very long and triangular with a wide ruflling membrane at the base of the triangle. They may have many vacuoles. 6. Cells with small dense nuclei and streak cytoplasm.These resemble the form shown by some lymphocytes. The pale-staining "streak cytoplasm" is caused by the gliding movement of the cell on the glass surface. 7. Cells with an eccentric nucleus.-These cells have dense cytoplasm and cover a smaller area after 20 hours on glass than the other rapidly adherent cell-types. The nucleus is so eccentrically positioned that it appears as a cap at one end. Some of these cells have filamentous pseudopodia, but they do not have prominent vacuoles (figure, e).
Results The proportion of rapidly adhering cells in normal amniotic fluids was less than 6% of the viable cells plated in the 92 samples tested. The calculated proportion of adherent cells was higher when the viable-cell count was low, owing to the errors involved in plating small numbers of cells. The proportion of rapidly adhering cells found in amniotic fluids from pregnancies with neural-tube defects was always greater than that found in normal controls (table i). With one exception (A70/6), the proportion of these cells in anencephaly was greater than 20% and often nearly 100% of the viable cells plated. Though the proportion of adhering cells in spina bifida was lower, it was always at least 50% greater than the upper limit of normal. Many of the amniotic fluids were contaminated with blood of maternal, fetal, or mixed maternal and fetal origin, but, though blood contamination appears to influence the adherence of other cells in certain cases of anencephaly, it does not have this effect in normal amniotic fluids. In amniotic fluids from normal fetuses the few cells which adhere to the coverslips are epithelioid. However, in case A8/6, where the sample was the product of third penetration of a large anterior placenta, the proportion of adherent cells was in the anencephalic range. The cells are classified by morphology into seven types (see table n and figure), and the proportion of each celltype among the adherent cells has been estimated for each abnormal amniotic fluid. Cytoplasmic inclusions and large prominent vacuoles were seen in some of the long bipolar, long triangular fibroblastic, multiple filamentous pseudopodial, and multinucleate cells, as well as in the large vacuolated cells. The proportion of adherent cells with vacuoles is shown in table n. The most characteristic adherent cell-types in amniotic fluids from pregnancies with neural-tube defects are the long bipolar cell (see figure, a), the filamentous pseudopodial cell (figure, c), the large vacuolated cell (figure, d), and the multinucleate cell (figure, b). The long bipo-
Some cell types found in amniotic fluid. a, long bipolar cell with prominent vacuoles; b, giant multinucleate cell with prominent vacuolation; c, cell with multiple filamentous pseudopodia; d, large vacuolated cell; e, small dense cell with eccentrically placed nucleus forming a "cap" at one end (arrowed).
921
lar cell and
multiple-filamentous pseudopodial
cell
were
essentially absent in the one case of exomphalos (A104/6, table n). The proportion of long triangular fibroblastic cells with a wide ruffling membrane was usually less than 10%, except in case A70/6 which, as
already indicated, was an unusual anencephalic with a low proportion of adherent cells and a comparatively low A.F.P. (table i). Cells with an eccentric nucleus in non-vacuolated cytoplasm are comparatively rare but were seen at a low frequency in the case of exomphalos
TABLE I-PROPORTION OF RAPIDLY ADHERING CELLS AND A.F.P. CONCENTRATIONS IN AMNIOTIC FLUIDS FROM CASES OF FETAL ABNORMALITY
’sB=spina bifida, A=anencephaly, ISB=iniencephaly with spina bifida, ASB=anencephaly with spma bifida, o=exomphalos, NAP=3rd transplacental amniocentesis from normal fetus with large anterior placenta, AP=anterior placenta, XX, XY indicate
respectively female and male. fUpper limit of normal in brackets. #=fetal, M=maternal. Normal range of cellular adherence in 20his 0-1-6%. TABLE II-PROPORTION OF CELLS WITH DIFFERENT MORPHOLOGIES IN AMNIOTIC FLUIDS FROM CASES OF FETAL ABNORMALITY
’Abbreviations
as
in table
t.
fTheproportion of vacuolated cells is indicated
as
follows:
+
0-20%,
++
21-40%,
+++
41-60%,
++++
61-80%,+++++ 81-100‘% .
922 and as the predominant cell type in the placental aspiration (case A8/6). In case A2/7, in which there was a large anterior placenta, the fluid contained a mixture of cells with eccentric nuclei and the four cell-types characteristic of a neural-tube defect. The fetus had a large spina bifida. Fluids containing cells with small dense nuclei and streak cytoplasm (which may be derived from
types of white blood-cells) were usually heavily contaminated with blood and were found only in cases of neural-tube defect. In only one case was an A.F.P. concentration outside the normal range observed in a sample where the fetus did not have a neural-tube defect or exomphalos. This sample was heavily contaminated with blood cells, 5% of which were of fetal origin. The proportion of adherent cells was distinctly abnormal (table i). However, the absence of the four cell-types characteristic of neuraltube defects (long bipolar, multiple-filamentous-pseudopodial, vacuolated, and multinucleate) and the presence of cells with eccentric nuclei excluded the possibility of neural-tube defect. some
Discussion The presence of rapidlyadhering cells in amniotic fluids where the fetus has an open neural-tube defect was first noted by Sutherland et awl. A quantitative technique was developed in which cells which became fixed to a glass or plastic surface after 18 hours in culture were expressed as a proportion of the total amnioticfluid cells. The limitation of the method was in the finding of about 5% false-positive counts-amniotic fluids with high proportions of rapidly adhering cells, in which
the outcome of the pregnancy was a normal infant.10, The amended procedure described here, in which adherent cells are expressed as a percentage of viable amniotic-fluid cells, appears to have eliminated this problem. Among 92 normal amniotic fluids, including 19 which were contaminated with varying amounts of maternal and fetal blood, none had more than 6% rapidly adhering cells. None of the 20 cases of neuraltube defect had less than 9% of such cells, and some cases with larger open lesions had more than 90% (table
i). An amniocentesis sample, obtained after a large anterior placenta had been traversed three times and which was heavily contaminated with maternal and fetal blood, showed 47% adherent cells. This suggested to us that examination of the morphology of the adherent cells in addition to an estimation of their proportion might allow more refined discrimination of categories of abnormality. In this case (A8/6) the A.F.P. concentration was 44 flg/ml at 20 weeks’ gestation, the upper limit of normal for this gestation being 17 g/ml. The fluid contained 46 x 106 red blood-cells per ml, about 2% of which were fetal by the Kleihauer test. This is insufficient to account for more than a marginal elevation of amnioticfluid A.F.P. The effect of fetal bleeding from the previous transplacental amniocentesis could not be estimated, but the absence of adhering cells of morphologies characteristic of neural-tube defects or exomphalos (table II) argued against a fetal abnormality. The fetus was normal. In case A2/7 the presence of a large anterior placenta caused a grossly blood-stained sample. The A.F.P. concentration was 49.4 tg/ml at 19 weeks’ gestation,
the upper limit of normal for this gestation being 24 jjg/ml. In this case fetal blood might have accounted for the raised A.F.P. However, the rapidly adherent cells showed a mixture of cell types: about half were characteristic placental cells with an eccentric nucleus, and half had features indicative of neural-tube defect. The fetus had a large spina bifida. Thus, in the presence of both placental cells and cells from the neural-tube defect all the characteristic cell types are rapidly adherent to the glass surface. Analysis of the morphology of rapidly adhering cells has potential for differential diagnosis of fetal disorders. It may allow recognition of non-neural-tube defects, such as exomphalos, pilonidal sinus, and duodenal atresia, where A.F.P. concentrations are unhelpful. Our experience with a case of exomphalos (A104/6) showed a of neural-tube paucity of some cells characteristic defects. Sutherland et al. ’I 10 described the rapidly adherent cells as macrophages, and we have confirmed that some of these cells do have phagocytic properties, However, the different shapes on glass, the relative migratory properties of the cytoplasm, the relative proportions in different cases, and the vacuolations and inclusions suggest that the cell types have different origins or that different factors in the amniotic fluid (or different concentrations of the same factor) affect cellular shape and adhesion. One of the main arguments in favour of the procedures described in this paper is the ease with which they
may be added to
existing prenatal diagnostic routines. Most laboratories which culture amniotic-fluid cells already estimate total and viable cells in a sample of amniotic fluid. If a coverslip is removed after 20 hours in culture and the adhering cells are stained and counted as a proportion of the viable cells, the laboratory has an early indication of possible fetal abnormality. This procedure has no effect on the time taken to produce a full cytogenetic analysis of the remaining cells." If the amniotic A.F.P. concentration is raised, a differential count of the various adhering cell types can then be used to gain information on the type of fetal abnormality and to eliminate false positives. We thank Dr J. B. Scrimgeour, Dr F. R. Clark, and Dr I. Duthie for making amniotic-fluid samples available. We are grateful for help
provided by Lilias Barron, Heidi Boland, Sandra Brown, Patricia Eason, Paddy Jelen, and Muriel Watt. We are indebted to Prof. H. J. Evans for his help and encouragement during the course of the work. This work was supported in part by a grant from the Medical Research Council to D. J. H. B.
Requests for reprints should be addressed to C. M. G.
REFERENCES 1. 2.
D.
Brock, J. H. Br. med. Bull. 1976, 32, 16. Laurence, K. M., Walker, S. M., Lloyd, M.,
Griffiths,
B. L. Lancet, 1975,
ii, 81. T. M., Seller, M. J. Singer, J. D. 1975, i, 1065. 4. Kunz, J., Schmid, J. ibid. 1976, i, 47. 5. Jandial, V., Thom, H., Gibson, J. Br. med. J. 1976, iii, 22. 6. Weinerg, A. G., Milunsky, A., Harrod, M. J. Lancet, 1975, ii, 496. 7. Brock, D. J. H. Clinica chim. Acta. 1974, 57, 315. 8. Wald, N., Cuckle, H. Lancet, 1976, i, 1292. 9. Sutherland, G. R., Brock, D. J. H., Scimgeour, J. B. ibid, 1973, ii, 1098 10. Sutherland, G. R., Brock, D. J. H., Scrimgeour, J. B. J. med. Genet. 1975. .12, 135. 11. Brock, D. J. H., Sutcliffe, R. G. Lancet, 1972, ii, 197. 12. Brock, D. J. H., Scrimgeour, J. B., Nelson, M. M. Clin. Genet. 1975, 7, 163. 13. Nelson, M. M. in Antenatal Diagnosis of Genetic Disease (edited by A. E. H. Emery). Edinburgh, 1973.
3.
Campbell, S., Pryse-Davies, J., Coltart, ibid.
AMNIOTIC-FLUID CELLS AS AN AID TO THE DIAGNOSIS OF NEURAL-TUBE DEFECTS CHRISTINE M. GOSDEN M.R.C. Clinical and Population Cytogenetics Unit, Western General Hospital, Edinburgh EH4 2XU
J. H. BROCK Department of Human Genetics, University of Edinburgh, Western General Hospital, Edinburgh EH4 2XU D.
In 20 amniotic-fluid samples taken in the second trimester from pregnancies in which the fetus had a neural-tube defect, the proportion of the total viable cells which adhered to glass surfaces after 20 hours in culture ranged from 9 to 100%. In 92 normal amniotic fluids this proportion was less than 6%. Furthermore, the morphology of the rapidly adhering cells was characteristic in spina bifida (8 cases) and anencephaly (12 cases) and distinct from the epithelioidlike cells seen in normal amniotic fluids, including many which were grossly blood-stained. A sample of amniotic fluid from fetal exomphalos and one from a pregnancy in which the placenta had been repeatedly traversed during amniocentesis had proportions of adherent cells in the pathological range, but the morphologies were very different from those seen in neural-tube defects. It is suggested that the techniques described here will be useful adjuncts to amniotic-fluid alpha-fetoprotein determination in the early diagnosis of fetal abnormality, particularly in blood-stained samples.
Summary
Introduction THE early prenatal diagnosis of neural-tube defects through measurement of amniotic-fluid alpha-fetoprotein (A.F.P.) concentrations is well established and widely
used.’ It has three main limitations. The first is in the
diagnosis of some cases of spina bifida, where potentially disabling lesions may be too small to cause more than a marginal elevation in the amniotic-fluid A.F.P. concentration.2 Physical measurements, such as ultrasonar scan or amniography, are usually not helpful in confirming or excluding these abnormalities.3 The second is in elevated A.F.P. values associated with a fetal neural-tube defect from those caused by less serious disorders, such as exomphalospilonidal sinus,5 or duodenal atresia.6 The third problem arises when the amniotic fluid is contaminated by fetal blood, since fetal serum-A.F.p. concentrations are about 150 times those in the amniotic fluid at a corresponding gestation.7 Since A.F.P. determination is now applied routinely to most amniotic-fluid samples, whatever the reason for
distinguishing
30
April 1977
amniocentesis, the possibility of an increasing number of misdiagnoses has become serious." An additional amniotic-fluid parameter which can be used as an adjunct to A.F.P. is urgently needed. Some years ago Sutherland et al. observed that amniotic fluids from pregnancies with fetal neural-tube defects had a high proportion of
rapidly adhering cells which were fetal in origin and had macrophage-like properties. The original "macrophage"-counting technique had serious limitations1o and has not been widely used. We have now modified this procedure and extended it to include a quantitative estimate of the relative proportions of seven types of rapidly adhering cells. The method has promise in confirming A.F.P.-dependent diagnoses of neural-tube defects and in distinguishing these defects from other fetal abnormalities and from fluids contaminated with fetal blood.
Materials and Methods Amniotic Fluid
Amniotic fluids
were
obtained
at
15 to 22 weeks’
by transabdominal amniocentesis, because high values indicated
a
risk of the fetus
having a
gestation
serum-A.F.p.
neural-tube defect.
A.F.P. Determination
concentrations were determined by rocket immunothe method previously described." The upper limit of normal for each week of gestation has already been defined. 12 A.F.P.
electrophoresis by
Determination
of Cell Numbers
Amniotic-fluid cells
were
stained with
trypan-blue diluted
1/5 with amniotic fluids." This mixture was incubated at 37°C for 5 minutes and then counted in Neubauer countingchambers. The total cell number, viable-cell count, and number of red blood cells were recorded, and the cell numbers for total, viable, and blood cells per millilitre of amniotic fluid were calculated. Smears were made from each blood-stained liquor, and these were stained by the Shepherd and Kleihauer methods to determine the proportion of maternal and fetal blood cells.
Plating for 20-hour Cellular Adherence 0.5ml of amniotic fluid was added to 4-5ml of Ham’s F10 medium buffered with Hepes and supplemented with 25% fetal-calf serum, 50 units/ml penicillin, and 50 g/ml streptomycin. Enough cells were taken so that at least 2000 viable cells were available for coverslip counting. 2.5ml of this mixture was added to each of two 50 mm sterile Falcon plastic petri dishes each containing 3 glass coverslips, 19 mm x 35 mm. The dishes were incubated at 37°C for 20 hours without disturbance. The covers lips were removed, rinsed with phosphate-buffered saline, fixed, and stained with one of the following : May-Grunwald-Giemsa, peroxidase, haematoxylm and eosin, cresyl-fast-violet, or gallocyanin. Each coverslip was mounted, and the total number of cells adherent to each coverslip was counted. The area of each coverslip was mea8018
c
920 sured as a proportion of the total area of each dish, and thus the number of viable cells plated per unit area could be calculated. The morphology and staining characteristics of each adherent cell were examined, and a differential count was obtained. The adherence in 20 hours (A2oh) was given by the average number of cells adherent per coverslip expressed as a percentage of the viable cells plated per coverslip.
Classification of Cell Types 1. Long bipolar cells.-These are very extended spindleshaped cells, some of which widen into bulbous ends with pseudopodia or lamelliform membranes. The cytoplasm, especially in the central wide area near the nucleus, often shows prominent vacuolation (see figure, a). 2. Cells with multiple filamentous pseudopodia (processes).-These cells vary in size and may have prominent distinctive feature is the presence of 3-9 filamentous pseudopodia, some of which may be very long (see vacuoles. Their
most
figure, c).
3. Large vacuolated cells.-These resemble amoebocytes derived from blood monocytes or tissue macrophages. They are large and irregular in shape and may adopt a variety of forms. They are usually isolated, although some may be clustered together, and they have prominent vacuoles and may have darkly staining cytoplasmic inclusions (see figure, d). 4. Giant multinucleate cells.-These cells range in size from those containing two nuclei to those with as many as 10-15 nuclei. These cells sometimes have vacuolated cytoplasm
(figure, b). 5. Long triangularfibroblastic cells.-These cells seem to be
polarised and are very long and triangular with a wide ruflling membrane at the base of the triangle. They may have many vacuoles. 6. Cells with small dense nuclei and streak cytoplasm.These resemble the form shown by some lymphocytes. The pale-staining "streak cytoplasm" is caused by the gliding movement of the cell on the glass surface. 7. Cells with an eccentric nucleus.-These cells have dense cytoplasm and cover a smaller area after 20 hours on glass than the other rapidly adherent cell-types. The nucleus is so eccentrically positioned that it appears as a cap at one end. Some of these cells have filamentous pseudopodia, but they do not have prominent vacuoles (figure, e).
Results The proportion of rapidly adhering cells in normal amniotic fluids was less than 6% of the viable cells plated in the 92 samples tested. The calculated proportion of adherent cells was higher when the viable-cell count was low, owing to the errors involved in plating small numbers of cells. The proportion of rapidly adhering cells found in amniotic fluids from pregnancies with neural-tube defects was always greater than that found in normal controls (table i). With one exception (A70/6), the proportion of these cells in anencephaly was greater than 20% and often nearly 100% of the viable cells plated. Though the proportion of adhering cells in spina bifida was lower, it was always at least 50% greater than the upper limit of normal. Many of the amniotic fluids were contaminated with blood of maternal, fetal, or mixed maternal and fetal origin, but, though blood contamination appears to influence the adherence of other cells in certain cases of anencephaly, it does not have this effect in normal amniotic fluids. In amniotic fluids from normal fetuses the few cells which adhere to the coverslips are epithelioid. However, in case A8/6, where the sample was the product of third penetration of a large anterior placenta, the proportion of adherent cells was in the anencephalic range. The cells are classified by morphology into seven types (see table n and figure), and the proportion of each celltype among the adherent cells has been estimated for each abnormal amniotic fluid. Cytoplasmic inclusions and large prominent vacuoles were seen in some of the long bipolar, long triangular fibroblastic, multiple filamentous pseudopodial, and multinucleate cells, as well as in the large vacuolated cells. The proportion of adherent cells with vacuoles is shown in table n. The most characteristic adherent cell-types in amniotic fluids from pregnancies with neural-tube defects are the long bipolar cell (see figure, a), the filamentous pseudopodial cell (figure, c), the large vacuolated cell (figure, d), and the multinucleate cell (figure, b). The long bipo-
Some cell types found in amniotic fluid. a, long bipolar cell with prominent vacuoles; b, giant multinucleate cell with prominent vacuolation; c, cell with multiple filamentous pseudopodia; d, large vacuolated cell; e, small dense cell with eccentrically placed nucleus forming a "cap" at one end (arrowed).
921
lar cell and
multiple-filamentous pseudopodial
cell
were
essentially absent in the one case of exomphalos (A104/6, table n). The proportion of long triangular fibroblastic cells with a wide ruffling membrane was usually less than 10%, except in case A70/6 which, as
already indicated, was an unusual anencephalic with a low proportion of adherent cells and a comparatively low A.F.P. (table i). Cells with an eccentric nucleus in non-vacuolated cytoplasm are comparatively rare but were seen at a low frequency in the case of exomphalos
TABLE I-PROPORTION OF RAPIDLY ADHERING CELLS AND A.F.P. CONCENTRATIONS IN AMNIOTIC FLUIDS FROM CASES OF FETAL ABNORMALITY
’sB=spina bifida, A=anencephaly, ISB=iniencephaly with spina bifida, ASB=anencephaly with spma bifida, o=exomphalos, NAP=3rd transplacental amniocentesis from normal fetus with large anterior placenta, AP=anterior placenta, XX, XY indicate
respectively female and male. fUpper limit of normal in brackets. #=fetal, M=maternal. Normal range of cellular adherence in 20his 0-1-6%. TABLE II-PROPORTION OF CELLS WITH DIFFERENT MORPHOLOGIES IN AMNIOTIC FLUIDS FROM CASES OF FETAL ABNORMALITY
’Abbreviations
as
in table
t.
fTheproportion of vacuolated cells is indicated
as
follows:
+
0-20%,
++
21-40%,
+++
41-60%,
++++
61-80%,+++++ 81-100‘% .
922 and as the predominant cell type in the placental aspiration (case A8/6). In case A2/7, in which there was a large anterior placenta, the fluid contained a mixture of cells with eccentric nuclei and the four cell-types characteristic of a neural-tube defect. The fetus had a large spina bifida. Fluids containing cells with small dense nuclei and streak cytoplasm (which may be derived from
types of white blood-cells) were usually heavily contaminated with blood and were found only in cases of neural-tube defect. In only one case was an A.F.P. concentration outside the normal range observed in a sample where the fetus did not have a neural-tube defect or exomphalos. This sample was heavily contaminated with blood cells, 5% of which were of fetal origin. The proportion of adherent cells was distinctly abnormal (table i). However, the absence of the four cell-types characteristic of neuraltube defects (long bipolar, multiple-filamentous-pseudopodial, vacuolated, and multinucleate) and the presence of cells with eccentric nuclei excluded the possibility of neural-tube defect. some
Discussion The presence of rapidlyadhering cells in amniotic fluids where the fetus has an open neural-tube defect was first noted by Sutherland et awl. A quantitative technique was developed in which cells which became fixed to a glass or plastic surface after 18 hours in culture were expressed as a proportion of the total amnioticfluid cells. The limitation of the method was in the finding of about 5% false-positive counts-amniotic fluids with high proportions of rapidly adhering cells, in which
the outcome of the pregnancy was a normal infant.10, The amended procedure described here, in which adherent cells are expressed as a percentage of viable amniotic-fluid cells, appears to have eliminated this problem. Among 92 normal amniotic fluids, including 19 which were contaminated with varying amounts of maternal and fetal blood, none had more than 6% rapidly adhering cells. None of the 20 cases of neuraltube defect had less than 9% of such cells, and some cases with larger open lesions had more than 90% (table
i). An amniocentesis sample, obtained after a large anterior placenta had been traversed three times and which was heavily contaminated with maternal and fetal blood, showed 47% adherent cells. This suggested to us that examination of the morphology of the adherent cells in addition to an estimation of their proportion might allow more refined discrimination of categories of abnormality. In this case (A8/6) the A.F.P. concentration was 44 flg/ml at 20 weeks’ gestation, the upper limit of normal for this gestation being 17 g/ml. The fluid contained 46 x 106 red blood-cells per ml, about 2% of which were fetal by the Kleihauer test. This is insufficient to account for more than a marginal elevation of amnioticfluid A.F.P. The effect of fetal bleeding from the previous transplacental amniocentesis could not be estimated, but the absence of adhering cells of morphologies characteristic of neural-tube defects or exomphalos (table II) argued against a fetal abnormality. The fetus was normal. In case A2/7 the presence of a large anterior placenta caused a grossly blood-stained sample. The A.F.P. concentration was 49.4 tg/ml at 19 weeks’ gestation,
the upper limit of normal for this gestation being 24 jjg/ml. In this case fetal blood might have accounted for the raised A.F.P. However, the rapidly adherent cells showed a mixture of cell types: about half were characteristic placental cells with an eccentric nucleus, and half had features indicative of neural-tube defect. The fetus had a large spina bifida. Thus, in the presence of both placental cells and cells from the neural-tube defect all the characteristic cell types are rapidly adherent to the glass surface. Analysis of the morphology of rapidly adhering cells has potential for differential diagnosis of fetal disorders. It may allow recognition of non-neural-tube defects, such as exomphalos, pilonidal sinus, and duodenal atresia, where A.F.P. concentrations are unhelpful. Our experience with a case of exomphalos (A104/6) showed a of neural-tube paucity of some cells characteristic defects. Sutherland et al. ’I 10 described the rapidly adherent cells as macrophages, and we have confirmed that some of these cells do have phagocytic properties, However, the different shapes on glass, the relative migratory properties of the cytoplasm, the relative proportions in different cases, and the vacuolations and inclusions suggest that the cell types have different origins or that different factors in the amniotic fluid (or different concentrations of the same factor) affect cellular shape and adhesion. One of the main arguments in favour of the procedures described in this paper is the ease with which they
may be added to
existing prenatal diagnostic routines. Most laboratories which culture amniotic-fluid cells already estimate total and viable cells in a sample of amniotic fluid. If a coverslip is removed after 20 hours in culture and the adhering cells are stained and counted as a proportion of the viable cells, the laboratory has an early indication of possible fetal abnormality. This procedure has no effect on the time taken to produce a full cytogenetic analysis of the remaining cells." If the amniotic A.F.P. concentration is raised, a differential count of the various adhering cell types can then be used to gain information on the type of fetal abnormality and to eliminate false positives. We thank Dr J. B. Scrimgeour, Dr F. R. Clark, and Dr I. Duthie for making amniotic-fluid samples available. We are grateful for help
provided by Lilias Barron, Heidi Boland, Sandra Brown, Patricia Eason, Paddy Jelen, and Muriel Watt. We are indebted to Prof. H. J. Evans for his help and encouragement during the course of the work. This work was supported in part by a grant from the Medical Research Council to D. J. H. B.
Requests for reprints should be addressed to C. M. G.
REFERENCES 1. 2.
D.
Brock, J. H. Br. med. Bull. 1976, 32, 16. Laurence, K. M., Walker, S. M., Lloyd, M.,
Griffiths,
B. L. Lancet, 1975,
ii, 81. T. M., Seller, M. J. Singer, J. D. 1975, i, 1065. 4. Kunz, J., Schmid, J. ibid. 1976, i, 47. 5. Jandial, V., Thom, H., Gibson, J. Br. med. J. 1976, iii, 22. 6. Weinerg, A. G., Milunsky, A., Harrod, M. J. Lancet, 1975, ii, 496. 7. Brock, D. J. H. Clinica chim. Acta. 1974, 57, 315. 8. Wald, N., Cuckle, H. Lancet, 1976, i, 1292. 9. Sutherland, G. R., Brock, D. J. H., Scimgeour, J. B. ibid, 1973, ii, 1098 10. Sutherland, G. R., Brock, D. J. H., Scrimgeour, J. B. J. med. Genet. 1975. .12, 135. 11. Brock, D. J. H., Sutcliffe, R. G. Lancet, 1972, ii, 197. 12. Brock, D. J. H., Scrimgeour, J. B., Nelson, M. M. Clin. Genet. 1975, 7, 163. 13. Nelson, M. M. in Antenatal Diagnosis of Genetic Disease (edited by A. E. H. Emery). Edinburgh, 1973.
3.
Campbell, S., Pryse-Davies, J., Coltart, ibid.