Aneuploidy screening in direct chorionic villus samples by fluorescence in situ hybridisation: the use of commercial probes in a clinical setting

British Journal of Obstetrics and Gynaecology February 2001, Vol. 108, pp. 215±218

SHORT COMMUNICATION

Aneuploidy screening in direct chorionic villus samples by ¯uorescence in situ hybridisation: the use of commercial probes in a clinical setting Claire R. Quilter a, Samantha Holman a, Raiham M. Y. A. AL- Hammadi a, Danae Theodorides a, Rosalind J. Hastings a,b, Joy D.A. Delhanty a,c,* The results of screening for the common aneuploidies involving chromosomes 13, 18, 21, X and Y by ¯orescent in-situ hybridisation (FISH) in direct preparations from 100 chorionic villus samples from pregnancies between 10 and 20 weeks' gestation are reported. Samples prepared using routine methods and analysed with commercially available probes, accurately detected 12 cases of fetal aneuploidy, all referred because of developmental abnormality. Three of the four cases where chromosome abnormality was detected in cultured villi but not by the direct ¯uorescence in situ hybridisation (FISH) assay, were due to con®ned placental mosaicism. No chromosomal anomalies were found in the 20 low risk cases where the referral reason was a familial single gene disorder. We conclude that the FISH assay with commercial probes may act as an accurate and less labour intensive alternative to direct chromosome analysis of chorionic villus samples. In cytogenetically low risk cases its use can obtain a result within the time needed for DNA analysis and avoid the need to set up cultures.

Introduction Karyotyping of chorionic villus samples is carried out in pregnancies at high risk of chromosomal or single gene disorders. In the former, a rapid result by a direct preparation is frequently requested but con®rmatory analysis of cultured mesenchymal tissue is also mandatory. For prenatal diagnosis requiring DNA analysis residual tissue is cultured to provide reassurance regarding the karyotype, although there is a low risk of a chromosomal abnormality. Both direct and culture techniques are very labour intensive and there are recurrent problems due to con®ned placental mosaicism affecting the cytotrophoblast or mesenchymal tissue, either of which may be unrepresentative of fetal tissue. Obtaining a result from cultured chorionic villus samples material may take up to two weeks, considerably longer than the time taken to obtain a DNA result; this can cause dif®culties with counselling and management. With the aim of

a

Clinical Cytogenetics, University College London, UK NE London Regional Cytogenetics Unit, Institute of Neurology, London, UK (Present address) c Department of Obstetrics and Gynaecology, University College London, UK (Present address)

avoiding some of these problems we have applied interphase analysis by ¯orescent in-situ hybridisation (FISH) to direct chorionic villus samples using commercially available DNA probes speci®c for chromosomes 13, 18, 21, X and Y. There are many advantages to this technique. Results are available in 24-48 hours, a considerable saving in time. Since the whole villus is used interphase nuclei from both the cytotrophoblast and the mesenchyme are screened providing an overall view of ploidy status for speci®c chromosomes. Screening for the major aneuploidies is suf®cient in women whose main reason for referral is a single gene disorder. Previous reports of interphase analysis by ¯orescent insitu hybridisation (FISH) of uncultured chorionic villus material using DNA probes from a variety of sources, have provided encouraging results 1±3. Diagnostic accuracy appeared to be high and the expected problem of maternal cell contamination failed to materialise 2. We report the results of the application of readily available commercial probes in a routine clinical setting for rapid screening for aneuploidy directly on chorionic villus samples.

b

* Correspondence: Professor J. D.A. Delhanty, Department of Obstetrics and Gynaecology, University College London, 86-96 Chenies Mews, London WC1E 6HX, UK. q RCOG 2001 British Journal of Obstetrics and Gynaecology PII: S 0306-545 6(00)00045-0

Methods The material totals 100 specimens and includes a pilot study of 29 samples collected from May to July 1998 and a main study group of 71 specimens. The latter comprised a special study group of 39 samples obtained between May and July 1999 together with 32 routine direct chorwww.bjog-elsevier.com

216 SHORT COMMUNICATION

ionic villus samples collected between January and August 1999. All were from ongoing pregnancies between 10 and 20 weeks' gestation. From May to July 1998 and May to July 1999, chorionic villus samples received by the cytogenetics department of University College London were assessed for suitability for inclusion in the pilot and special study groups, the aims of which were to assess the importance of contamination by maternal cells. Samples needed to be . 20mg in size, in order to have suf®cient material for long term culture, as well as two direct preparations, or at least 60mg in the case of DNA disorders. For those samples that were suitable, after setting up long-term cultures, the remaining choronic villus material was divided in half. One aliquot was dissected with ®ne needles under an inverted microscope to remove any maternal or anomalous material, the other aliquot was left undissected. Initially, both aliquots of material were treated with collagenase and colchicine (1mg/mL) in Chang's medium, at 378C for 45 minutes. Routine direct chorionic villus samples were not dissected but were otherwise treated similarly. After centrifugation at 1500 rpm for 5 minutes, pre-warmed 1% sodium citrate was added to the chorionic villus material for 15minutes and incubated at 378C. The sample was centrifuged as before, followed by drop-wise addition of methanol and glacial acetic acid (3:1), with agitation, up to a ®nal volume of 3ml. Fixing was repeated twice more, ending with centrifugation and removal of 2.5ml of ®xative. Two drops of the ®xed suspension were dropped onto either end of a cold, labelled slide, to create two distinct regions of cells. Slides were then air dried in preparation for the FISH procedure. The Vysis Prenatal Detection Kit (AneuVision; Vysis, Downers Grove, Illinois, USA) was used for FISH analysis. This contains labelled unique DNA probes for chromosomes 13 (LS1 13, Spectrum Green, hybridising to 13q14) and 21 (LS1 21 Spectrum Orange, hybridising to 21q22.13 ±21q22.2) as one probe mix, and a second probe mix of labelled a -satellite centromere-speci®c DNA probes for chromosomes 18, X and Y (CEP 18, Spectrum Aqua, CEP X, Spectrum Green, and CEP Y, Spectrum Orange). After denaturation, 7mL of the probe mix 18, X and Y and 7mL of the 13, 21 probe mix were placed on the two distinct regions of cells, covered by a 22 x 50mm coverslip and sealed using rubber cement. Slides were incubated in a sealed dark, damp container at 378C overnight. After removing the coverslip, the slide was placed immediately into 0.4 x SSC/0.3% NP40 at 738C for two minutes, followed by a wash in 2 x SSC/ 0.1% NP40 at room temperature for 30 second. The slide was subsequently put through an alcohol series of 75%, 80% and 100% ethanol for one minute in each and was then air dried in reduced light. Once dry 10mL of DAPI in anti-fade solution was added to each end of the slide for counterstaining and the slide was covered using a 22x50mm coverslip.

For analysis, a Zeiss Axioskop ¯uorescence microscope with a 100 watt lamp and ®lter set speci®c for Spectrum Orange, Spectrum Green and Spectrum Aqua (Vysis) was used. Routinely, counts were made on 30 good quality non-overlapping nuclei for each probe. In cases with a male fetus that were part of the special study group, 50 nuclei were analysed for X, Y and 18 to check for maternal cell contamination. The criteria used for interpretation were as follows: ² A minimum of 25/30 (80%) of nuclei with two signals for normality. ² A minimum of 60% of nuclei with extra or missing signals for aneuploidy. ² Cases with 20-59% of nuclei scored as abnormal and which were likely to be mosaic were referred for further evaluation by counting more cells or repeating the test or waiting for the culture result if still mosaic.

Results Pilot study Of the 29 samples, 20 were male, 9 of which were dissected. The deviant sex ratio is because samples that were male on direct chromosome analysis were preferentially chosen for the pilot study to assess the likely importance of maternal cell contamination. Two of the dissected and 4 of 11 in the non-dissected group had evidence of maternal cell contamination (i.e. 2 X signals in the absence of a Y signal), ranging from one to ®ve nuclei in the former and from one to six in the latter group. Four samples were aneuploid; two were trisomy 13 (one dissected, one not), and one each of trisomy 18 and trisomy 21, neither had been dissected. Interestingly, although all 30 nuclei in the trisomy 21 sample had three signals, the other three trisomies had from 27% - 30% of nuclei with two signals despite the fact that two were male with no evidence of widespread maternal cell contamination. These results may suggest that the samples with trisomies for chromosomes 13 and 18 were mosaic with a normal cell line present, yet the cytogenetic results from cultured material appeared to be nonmosaic. However, overall there were no discrepancies between the FISH and cytogenetic results when the agreed criteria were applied. Main study Referral reasons for the 71 cases in the main study group were as follows: maternal age (24), abnormal ultrasound scan (n ˆ 17), haemoglobinopathies (n ˆ 9), other single gene disorders (n ˆ 11), previous aneuploidy (n ˆ 5), abnormal serum screen (n ˆ 4), oligohydramnios (n ˆ 1). Sixty-three samples gave a normal result for the q RCOG 2001 Br J Obstet Gynaecol 108, pp. 215±218

SHORT COMMUNICATION 217

tested chromosomes by FISH: 40 were male and 23 female. The eight abnormalities detected by FISH included three cases each of 45,X and of trisomy 18 and one each of trisomy 21 and trisomy 13; all were con®rmed by cytogenetic analysis of cultured villi and were referred because of an abnormal ultrasound scan (Table 1). There were four instances where abnormalities of the chromosomes were detected in cultured villi but were not seen by FISH analysis (Table 1). In case 1157, referred because of fetal abnormality, an unbalanced translocation leading to partial monosomy for chromosome 14 was found (45,XX, -13,der (14)t(13;14)(q11;q32.3); this was undetectable with the probes used for FISH. Case 992 was referred because of a previous Down's pregnancy; chromosomes from the direct chorionic villus samples were normal male as were those of a follow up amniotic ¯uid sample, but the slow growing cultured villi showed 22% of cells to be 45,X. Case 1062 was referred for a chorionic villus sample because of a failed amniotic ¯uid result elsewhere; amniocentesis had been performed because of an abnormal serum alpha-fetoprotein result. The chorionic villus sample direct chromosome result was normal but the culture had the karyotype 47, XY, 1i(13)(q10); subsequently fetal blood, placenta and cord were all shown to have normal chromosomes. The fourth case, 2416, referred because of advanced maternal age, showed an additional unidenti®ed marker chromosome in 9% of cells from cultured villi; this was not present in amniotic ¯uid. Later, evidence of hydrocephalus was seen on the ultrasound scan and the fetus was delivered at 36 weeks with ventriculomegaly. The relevance of the marker chromosome seen in a few cultured cells to this anomaly is unknown. No chromosome abnormalities of any kind were detected in the 20 samples where the primary referral reason was for a single gene disorder.

In the special study group of 39 samples that formed part of the main study, for those with normal results on FISH analysis the percentage of cells with the expected number of signals for the chromosomes tested ranged between 98% and 100%. The three trisomic samples in this sub group showed 100% of cells with three signals in the two trisomy 18 cases and 97% with three signals in the case with trisomy 21; all had been dissected prior to processing. To assess maternal cell contamination 14 samples from male pregnancies were analysed from both dissected and non-dissected samples. Only one dissected sample showed a single nucleus with 2 X signals and no Y, whereas the non-dissected sample from the same patient showed ®ve of 50 nuclei with this pattern of signals. In addition, two other nondissected samples showed three and four nuclei with female signals. Overall careful dissection of the specimen prior to processing made a difference in the percentage of female nuclei of 1.6% (0.1% in the dissected samples compared 1.7% in the non-dissected group). Discussion Chromosomal analysis of chorionic tissue is particularly appropriate for high risk pregnancies in which fetal abnormality has been detected. It has a high predictive value, particularly for non-mosaic trisomies 13, 18 and 21 and for numerical sex chromosome anomalies 4. It is evident from our results and those of others 1,2,3that interphase analysis of uncultured chorionic villus samples tissue is in turn an accurate means of detecting these same aneuploidies making it particularly useful when abnormal results are obtained by ultrasound scan. At the opposite end of the spectrum the FISH assay can

Table 1. Detection of chromosomal abnormality by ¯ourescent in situ hybridisation (FISH) assay and by karyotyping of cultured chorionic villi in the main study group of 71 samples. S AFP ˆ serum alpha fetoprotein; FBS ˆ fetal blood sample; CVS ˆ chorionic villus sampling. Case no.

Referral reason

FISH result

Culture result

836 837 992

Fetal abnormality Fetal abnormality Previous Downs

3 £ 18 3 £ 18 normal

1062

Abnormal S AFP

normal

47, XY, 118 47, - -, 118 46,XY/45,X 78% / 22% 47, XY, 1i(13)(q10)

1068 1073 1157

Fetal abnormality Fetal abnormality Fetal abnormality

3 £ 21 3 £ 18 normal

1652 1856 2001 2181 2416

Fetal abnormality Fetal abnormality Fetal abnormality Fetal abnormality Maternal age

1£X 1£X 3 £ 13 1£X normal

a

Fetus delivered at 36 weeks with ventriculomegaly.

q RCOG 2001 Br J Obstet Gynaecol 108, pp. 215±218

47,XY, 121 47,XX, 118 45,XX, -13 der(14) t(13;14)(q11;q32.3) 45, X 45, X 47, - -, 113 45, X 46, - -/47, - - 1mar 29/3

Follow up

amniotic ¯uid 46,XY CVS direct, FBS, cord 46,XY.

amniotic ¯uid 46,± a

218 SHORT COMMUNICATION

also provide an accurate screen for the common aneuploidies in low risk situations when the main reason for referral is a familial single gene disorder. Twelve common aneuploidies were detected in the 100 direct chorionic villus samples tested. Use of the Vysis Prenatal Detection Kit (AneuVision) allowed us to detect all cases of fetal trisomies 13, 18, and 21 and monosomy X. Of the four cases where an abnormality was detected in the cultured chorionic tissue when the FISH assay was normal for the tested chromosomes, one only was of de®nite clinical signi®cance (case 1157) with partial loss of chromosome 14. Clearly, when fetal abnormality has been detected a normal FISH result must be followed by karyotypic analysis of mesenchymal cells. The other three discrepancies are due to con®ned placental mosaicism; in two of these (cases 992 and 2416) abnormalities were present in a small minority of cultured cells. Case 1062 is interesting because effectively non-mosaic tetrasomy 13 was found in cultured tissue yet this abnormality was not found in the FISH assay nor in the karyotyped direct chorionic villus sample, the fetal blood or the postnatal cord and placental samples. In this instance, the FISH assay was a more accurate predictor of fetal karyotype than was the cultured mesenchyme. In view of the close association between chorionic tissue and maternal decidua it might have been anticipated that maternal cell contamination would be serious problem when using a FISH assay on direct chorionic villus samples. Bryndorf et al. 2 did not ®nd this to be a problem when using FISH for the analysis of mesenchymal chorionic villus that had been allowed to attach overnight, providing thorough dissection of villi had taken place. In our experience, although thorough dissection can reduce the number of presumed maternal cells present, routine removal of gross contamination is quite suf®cient to avoid problems that might be associated with maternal cell contamination. A disomic cell line may of course also be detectable in a trisomic conception when mosaicism is present; this will be a common event particularly in cases of trisomies 13 and 18 since the normal chorionic cell line will aid survival 5. Higher levels of disomic cells have the potential to produce FISH results that are at variance with those from cultures, as was the case with two samples from our pilot study. We found no abnormalities in the 20 low risk cases referred solely because of a familial single gene disorder.

Application of the FISH assay has considerable bene®t in this situation; a rapid result is obtained, within the same time as that from the parallel DNA analysis, and the need to set up cultures is avoided. The two-week wait for results from cultured chorionic villus samples can mean that the family is told of a normal DNA result well before the chromosome analysis is complete. A large interphase FISH study 3 on over 2000 chorionic villus samples, with results from mesenchymal tissue only, led to the conclusion that the FISH assay may be more accurate and less labour-demanding than direct chromosome analysis of chorionic villus samples. That study, although large, used FISH probes available only for research purposes. Our much smaller study is a useful indicator since it was carried out in a routine clinical setting, the techniques were straightforward, the whole villus was sampled and use was made of readily available commercial probes. We also can draw the same conclusions as Bryndorf and colleagues. Acknowledgements The authors would like to thank all the patients and staff of the various hospitals that provided samples, and the staff of the Clinical Cytogenetics Unit, UCLH, who are not authors on this paper. References 1. Rao NP, Hayworth R, Cox K, Grass F, Pettenati MJ. Rapid detection of aneuploidy in uncultured chorionic villus cells using ¯uorescence in situ hybridization. Prenat Diagn 1993;13:233±238. 2. Bryndorf T, Christensen B, Xiang Y, Philip J. Prenatal diagnosis by ¯uorescence in situ hybridization on chorionic villus cells: nonsigni®cance of maternal cell contamination. Fetal Diagn Ther 1994;9:73±76. 3. Bryndorf T, Christensen B, Vad M, et al. Prenatal detection of chromosome aneuploidies in uncultured chorionic villus samples by FISH. Am J Hum Genet 1996;59:918±926. 4. Smith K, Lowther G, Maher E, Hourihan T, Wilkinson T, Wolstenholme J. The predictive value of ®ndings of the common aneuploidies, trisomies 13,18 and 21, and numerical sex chromosome abnormalities at CVS: experience from the ACC UK Collaborative Study. Prenat Diagn 1999;19:817±826. 5. Kalousek DK, Barrett I, McGillivray BC. Placental mosaicism and intrauterine survival of trisomies 13 and 18. Am J Hum Genet 1989;44:338±343. Accepted 11 October 2000

q RCOG 2001 Br J Obstet Gynaecol 108, pp. 215±218