Assessment of the separation of X- and Y -bearing sperm on albumin gradients using double-label fluorescence in situ hybridization

Transcription

1 FERTILITY AND STERILITY Copyright" 1994 The American Fertility Society Vol. 61, No.4, April 1994 Printed on acid-free paper in U. 8. A. Assessment of the separation of X- and Y -bearing sperm on albumin gradients using double-label fluorescence in situ hybridization Huai-Xiu Wang, M.D.* Sean P. Flaherty, Ph.D.t Nicholas J. Swann, B.A. Hons. Colin D. Matthews, M.D. Department of Obstetrics and Gynaecology, The University of Adelaide, The Queen Elizabeth Hospital, Woodville, South Australia, Australia Objective: To determine the ratio of X- to Y -bearing human spermatozoa in fractions isolated from discontinuous albumin gradients. Design: The proportions of X- and V-bearing sperm were determined in neat semen samples (control) and in albumin-separated fractions from the same samples. Two albumin methods were used: a two-layer method (experiment 1) and a three-layer method (experiment 2). X- and Y -bearing sperm were identified simultaneously using chromosome-specific DNA probes and fluorescence in situ hybridization. Setting: Hospital-based university department. Participants: Healthy donors with normal semen characteristics. Main Outcome Measures: The proportions of haploid cells (X or Y) and cells with two sex chromosomes (XX, YY, or XV) were determined. Results: Labeling efficiencies were >96% in all samples. Control samples showed a 1:1 ratio of X - to Y -bearing sperm. Fractions isolated on albumin gradients showed a slight, but statistically significant enrichment of X-bearing sperm. This was evident with both albumin methods. Conclusions: Discontinuous albumin gradients do not enrich V-bearing sperm as previously reported. Fertil Steril 1994;61:720-6 Key Words: Human sperm, albumin gradients, fluorescence in situ hybridization, sex selection, X chromosome, Y chromosome The separation of X- and V-bearing sperm has generated interest and controversy since the first report in the 1970s (1). Subsequent studies have suggested that enriched populations of X-bearing sperm can be isolated using Percoll gradients (2) or Received October 18, 1993; revised and accepted December 16,1993. * Present address: Shanxi Provincial People's Hospital, Taiyuan, People's Republic of China. t Reprint requests: Sean Flaherty, Ph.D., Department of Obstetrics and Gynaecology, The University of Adelaide, The Queen Elizabeth Hospital, Woodville, South Australia 5011, Australia (FAX: ). Sephadex columns (3), whereas Y -bearing sperm can be enriched using discontinuous albumin columns (4, 5) or a swim-up procedure (6). Clinical trials have shown altered sex ratios at birth after the application of various sperm selection procedures (2, 4, 7-9), but the efficacy of these methods has been questioned because several studies have been unable to confirm either the laboratory or clinical results (10-17). Part of the controversy may be due to the quinacrine staining method used to detect Y bodies in spermatozoa, because recent evidence has shown that the quinacrine method is unreliable and may not necessarily detect only the Y chromosome (18). 720 Wang et al. Sperm separation on albumin gradients Fertility and Sterility

2 Fluorescence in situ hybridization has been used to identify X - or Y -bearing sperm in human semen (19,20). However, methods that use a single chromosome probe are limited in their application for determining the sex of sperm because one cannot be certain whether the unlabeled cells contain the other chromosome or if the absence of labeling is due to an inability of the probe to enter the sperm nucleus. This can cause highly erroneous results, particularly if the labeling efficiency is low or variable. To overcome this limitation, we recently developed a double-label fluorescence in situ hybridization technique for simultaneously identifying X and Y -bearing human sperm (21). This technology has opened the way for objective assessment of the effects of sperm separation techniques on the ratio of X- to Y -bearing human sperm, and we recently reported that isolation of motile sperm by the swim-up procedure does not alter the 1:1 ratio of X to Y -bearing sperm (17). In this study, we have evaluated whether discontinuous albumin gradients enrich Y -bearing human sperm. The ratio of X- to Y-bearing sperm was examined in neat semen and in fractions isolated from the same semen samples using discontinuous albumin columns. Two albumin methods were used: a two-layer method (4) and a three-layer method (9). The three-layer method is referred to as protocol 3 by Beernink et al. (9). We attempted to duplicate the procedures as closely as possible from the published reports (4, 9). The results clearly demonstrate that there is no enrichment of Y -bearing sperm in the albumin-separated sperm fractions, but there is a slight enrichment of X bearing sperm. Semen Samples MATERIALS AND METHODS Samples were obtained from 14 normal healthy donors who attended the Andrology Laboratory at The Queen Elizabeth Hospital. The samples were produced by masturbation, allowed to liquefy at room temperature for up to 30 minutes, and then analyzed using World Health Organization (WHO) guidelines (22). Sperm morphology was assessed using WHO criteria on smears stained by the modified Papanicolaou technique. The median (range) semen values for the 24 samples used in this study were as follows: volume 3.0 ml (1.5 to 6.0 ml); sperm concentration 75 X 10 6 /ml (30 to 146 X 10 6 ); progressive motility 53% (36% to 68%). Sperm morphology was not assessed in some samples, but all the donors regularly produced samples with >30% normal morphology. A single sample from 12 donors was used to evaluate each albumin method. The sample was divided into two aliquots. One aliquot (control) was prepared immediately for double-label fluorescence in situ hybridization, while a fraction was isolated from the other aliquot using one of the albumin methods and then prepared for double-label fluorescence in situ hybridization. Experiment 1: Two-Layer Albumin Gradients Sperm were isolated using a one-step, two-layer albumin gradient as described by Dmowski et al. (4). The liquefied semen was diluted 1:1 with Tyrode's solution (Commonwealth Serum Laboratories, Melbourne, Victoria, Australia) and then centrifuged at 700 X g for 10 minutes. The supernatant was removed and the pellet was resuspended gently in Tyrode's solution to a concentration of 60 X 10 6 / ml. Albumin gradients were prepared in 8 X 75 mm tubes and consisted of 0.5 ml of 20% (wt/vol) human serum albumin (HSA; A 1653 Fraction V; Sigma Chemical Co., St. Louis, MO) overlaid with 1 ml of 10% HSA. The albumin solutions were prepared using Tyrode's solution. Each gradient was overlaid with 0.5 ml of the washed sperm fraction (three gradients per sample) and incubated at room temperature (20 C). After 60 minutes the top 0.5- ml layer was removed carefully from each gradient and, after an additional 30 minutes, the 10% albumin layers were removed and the 20% layers were centrifuged at 700 X g for 15 minutes. The final sperm pellets were resuspended in Tyrode's solution and diluted to a concentration of 20 X 10 6 /ml. Experiment 2: Three-Layer Albumin Gradients Sperm were isolated using Ericsson's protocol 3 method (9), which consists of two steps and three layers. Semen was diluted 1:1 with Tyrode's solution and then centrifuged for 10 minutes at 300 X g. The sperm pellet was resuspended in Tyrode's solution to a concentration of 30 X 10 6 /ml and then 0.5-mL fractions of this washed-sperm preparation were overlaid on 1.0 ml of 7.5% HSA in 8 X 75 mm tubes (three tubes per sample) and incubated for 60 minutes at room temperature. The 7.5% albumin layer was aspirated and centrifuged for 10 minutes at 300 X g, and each pellet was resuspended in Tyrode's solution to 0.5 ml and then layered onto a Vol. 61, No.4, April 1994 Wang et al. Sperm separation on albumin gradients 721

3 two-step albumin gradient consisting of 0.5 ml of 20% HSA overlaid with 1.0 ml of 12.5% HSA in an 8 X 75 mm tube. After 1.5 hours, the 20% layers were pooled, diluted 1:1 with Tyrode's solution, and centrifuged at 300 X g for 10 minutes. The final sperm pellet was resuspended in Tyrode's solution at a concentration of 20 X 10 6 /ml. Pretreatment of Sperm A modification of our published method (21) was used to pretreat sperm nuclei. Neat semen or the isolated fraction (20 X 10 6 /ml) was mixed with phosphate-buffered saline (PBS, ph 7.2) containing 6 mm ethylenediaminetetraacetic acid (EDT A) and then centrifuged at 175 X g for 10 minutes. The pellet was resuspended to the same sperm concentration and aliquots were treated with 2, 3, or 4 mm dithiothreitol (DTT) in PBS at room temperature for 45 minutes with regular mixing. We found that some samples decondensed better in higher or lower DTT concentrations, so we treated each sample with three concentrations and chose the best preparation for in situ hybridization. This consistently produced hybridization and labeling efficiencies of 96% to 99%. After centrifugation, the pellet was resuspended in PBS and recentrifuged. The supernatant was removed and the sperm were fixed with methanol:glacial acetic acid (3:1), dropped onto slides and air dried. Mitotic Chromosome Spreads Mitotic chromosome spreads from human male peripheral blood lymphocytes were prepared as described previously (17, 21). DNA Probes A biotinylated X chromosome-specific probe (DXZ1) was purchased from Oncor (Gaithersburg, MD). This is an alpha satellite probe that was described originally by Waye and Willard (23). Details of the Y chromosome-specific (HRY) probe were given in our earlier studies (17, 21). HRY was labeled with digoxigenin-dutp by random priming using a DIG DNA labeling kit (Boehringer-Mannheim, Castle Hill, New South Wales, Australia). Fluorescence In Situ Hybridization The protocol for double-label fluorescence in situ hybridization has been described by Han et al. (17, 21). Briefly, a mixture of the X-probe (5 ng) and Y -probe (25 ng) were applied to each slide and hy- bridized to the sperm DNA for 18 hours at 37 C. After thorough washing, the probes were visualized with a multistep immunofluorescence procedure using fluorescein isothiocyanate (FITC) for the X probe and Texas red for the Y -probe. The following steps were used (45 to 60 minutes each): preincubation in 5% nonfat dry milk; 10 ~g/ml FITC-avidin DN (Vector Laboratories, Burlingame, CA); 5 ~g/ ml biotinylated anti-avidin (Vector) and 1 ~g/ml mouse anti -digoxigenin (Boehringer-Mannheim); 10 ~g/ml FITC-avidin DN and 12 ~g/ml Texas red-rabbit anti-mouse immunoglobulin (Ig)G (Jackson ImmunoResearch Laboratories, West Grove, PA); 15 ~g/ml Texas red-goat anti-rabbit IgG (Vector). The slides were washed with Trisbuffered saline (ph 7.5) containing 0.05% Tween 20 and serum (goat or rabbit) between each incubation step. Finally, the slides were mounted for microscopy as outlined by Han et al. (17). Slides were examined at X1,250 with a Leitz microscope (Ernst Leitz, Wetzlar, Germany) equipped with epifluorescence and three filter blocks (A, nuclear counterstain; 12/3, FITC; N-2, Texas red). Slides were coded and 1,000 sperm were scored blind from each preparation by one observer (H.X.W.). Sperm with a single green (FITC) or red (Texas red) spot were classified as haploid X-or Y -bearing sperm, respectively; whereas cells that contained two spots (XX, YY, or XY) were scored separately. Statistical Analysis All statistical analyses were performed using the InStat program (Graph Pad, San Diego, CA). A P value < 0.05 was considered significant. A preliminary analysis demonstrated that data were normally distributed, so differences in X:Y ratios and the frequencies of specific categories (haploid X or Y, XX, YY, XY) between matched controls and Percoll fractions were analyzed by paired t-tests. RESULTS Mitotic chromosome spreads and interphase cells from male lymphocyte preparations were used as positive controls in each fluorescence in situ hybridization procedure. There was a very bright, distinct green spot (FITC) over the centromere of the X chromosome and a more diffuse red spot (Texas red) over the long arm of the Y chromosome. Recoveries from the albumin gradients varied considerably from sample to sample, and poorer 722 Wang et al. Sperm separation on albumin gradients Fertility and Sterility

4 yields occurred in samples with increased viscosity. The median (range) recoveries, expressed as a percentage of the sperm applied to the gradients, were 10.0% (2.5% to 13.0%) for the two-layer method and 8.5% (2.2% to 23.0%) for the three-layer method. Experiment 1: Two-Layer Albumin Gradients A total of 12,000 sperm were scored in the control preparations with a mean labeling efficiency of 97.9% (range 97.1 % to 98.7%). The overall ratio of X- to Y -bearing sperm in the controls was 49.0:48.2 (Table 1). High labeling efficiencies were also obtained in the albumin-separated fractions (mean 98.5%, range 97.3 to 99.5). The overall ratio of X- to Y -bearing sperm in the albumin fractions was 51.3:46.7. While there was some variation between the 12 samples (Table 1), the X:Y ratio in the albumin fractions was found to be significantly different from the controls (P < 0.005). Experiment 2: Three-Layer Albumin Gradients The results are presented in Table 2 and essentially parallel those of Experiment 1. The overall labeling efficiencies were again high in the control (mean 97.4%, range 96.2 to 98.6) and albuminseparated (mean 97.9%, range 96.5 to 99.7) samples. The overall ratios of X- to Y-bearing sperm were 48.8:48.1 (controls) and 50.8:46.7 (albuminseparated fractions). The ratios were significantly different (P < 0.05). Haploid and Aneuploid Sperm The proportions of haploid (X or Y), XX, YY, and XY cells in the control and albumin-separated fractions are shown in Table 3. The proportion of haploid X-bearing sperm was significantly greater in the albumin-separated fractions for both albumin methods. In the two-layer method, there was also a slight but nonetheless significant difference in the proportion of haploid Y -bearing and YY sperm but no significant difference in the XX group. In the three-layer method, there were no significant differences in the proportions of haploid Y -bearing sperm, XX, or YY sperm between the control and albumin fractions. There was a significant reduction in the number of XY cells in the albumin-separated fraction using both methods. DISCUSSION Since the first report by Ericsson et al. (1) that Y -bearing human sperm could be enriched using discontinuous albumin gradients, there have been numerous laboratory and clinical studies that have endeavored to confirm or deny the potential for preconceptional sex selection using the albumin method and other procedures (2-17). Unfortunately, the results of these studies are rather confusing and at times contradictory because of several factors. First, many of the clinical studies are based on small numbers of conceptions (2, 4, 7, 8, 16), although this criticism cannot be levied against the recent report by Beernink et al. (9). Second, compounding factors have not been taken into account in controlled clinical studies, and some factors such as the administration of clomiphene citrate appear to reverse the sex ratio to a predominance of females even when insemination is performed with samples that have been prepared on albumin Table 1 X and Y Labeling Results for the Two-Layer Albumin Method* Neat semen Albumin fraction Donor Labeled X Y Labeled X Y Mean ± SD ± ± ± ± ± ± 1.8 * Values are percentages. Vol. 61, No.4, April 1994 Wang et al. Sperm separation on albumin gradients 723

5 Table 2 X and Y Labeling Results for the Three-Layer Albumin Method* Neat semen Albumin fraction Donor Labeled X Y Labeled X Y Mean ± SD 97.4 ± ± ± ± ± ± 1.5 * Values are percentages. gradients (9). Third, specific methods for determining the ratio of X- to V-bearing sperm have been lacking and many studies have relied on quinacrine staining, which has been shown to be unreliable (18). Finally, studies have not duplicated correctly the previous published procedures, despite claiming to do so. The overall result has been a lack of reproducibility and considerable confusion. In this study, we used a double-label fluorescence in situ hybridization technique (21) to evaluate two albumin methods that were reported to enrich Table 3 Frequencies of Haploid and Aneuploid Cells Before and After Albumin Separation* Experiment 1: two-layer albumin method Haploid X Haploid Y XX YY XY Experiment 2: three-layer albumin method Haploid X Haploid Y XX YY XY * Values are percentages. Neat semen Albumin fraction 51.2t \1 t Albumin group is significantly different from neat semen group, P = Albumin group is significantly different from neat semen group, P < No significant difference between groups. II Albumin group is significantly different from neat semen group, P < \I Albumin group is significantly different from neat semen group, P < Y -bearing sperm. This technique provides simultaneous and conclusive identification of X-and Y bearing sperm using specific DNA probes, and it overcomes the limitations of quinacrine staining and fluorescence in situ hybridization using a single sex chromosome probe. The advantages of this method are discussed in Han et al. (17). Our working hypothesis was that an enriched population of Y -bearing sperm would be retrieved from the albumin gradients. However, no enrichment of Y -bearing sperm was found and most samples showed a slight, but significant enrichment of X-bearing sperm. This result contrasts some studies (1, 4, 5) but is in agreement with others that have evaluated the albumin method by quinacrine staining (10,11) or sperm karyotyping (13,14). The results of this current study and others now cast significant doubt on the efficacy of enriching Y bearing sperm using albumin gradients. Our results also conflict with published clinical studies in which the sex ratio at birth was skewed significantly toward males after insemination with albumin-separated sperm (4, 7-9). These clinical results cannot be disputed in light of the recent report by Beernink et al. (9), but the explanation of why the sex ratio is altered remains unclear. If it is confirmed that ovulation induction with clomiphene citrate shifts the sex ratio to a predominance of females despite sperm preparation on albumin gradients before insemination (9), then presumably it is not the sperm selection procedure per se, but other less obvious factors that are responsible for the altered sex ratios. Certainly this must be consistent with our experimental data. Ericsson et al. (1) originally proposed that albumin 724 Wang et al. Sperm separation on albumin gradients Fertility and Sterility

6 gradients enrich Y -bearing sperm because of their ability to swim faster in more viscous fluids than X bearing sperm. However, we have shown that selection of motile sperm by swim-up does not alter the 1:1 ratio of X- to Y-bearing sperm (17), so it is doubtful if there is any detectable difference in the motility of X and Y -bearing sperm. Because we found that albumin gradients enrich X-bearing sperm, an alternative hypothesis is needed to explain the enrichment on a basis other than motility. It is possible that gravity causes a slight enrichment of X-bearing sperm because of their increased mass. It is interesting that Brandriff et al. (13) also found a higher proportion of X -bearing sperm after karyotyping sperm selected on albumin gradients, although in their study the difference was not statistically significant. In this study we also differentiated between haploid sperm and those with XX, YY, or XY complements to estimate the frequencies of aneuploidy in the control and albumin-separated fractions. The frequencies of most XX and YY groups decreased in the albumin-separated fractions, although the differences were not always statistically significant. There was, however, a significant decrease in the proportion of XY cells in the albumin fractions, and this was anticipated based on the results of our swim-up study (17), as some ofthe XY cells are likely to be immature germ cells or leukocytes, which would not be expected to migrate into the 20% HSA layer. The results of this study also confirm that the ratio of X- to Y-bearing sperm in human semen is 1:1. In this and our previous studies (17, 21), we have now examined 46 donor semen samples using the doublelabel fluorescence in situ hybridization technique and the ratio has been consistently 1:1. This confirms data obtained by sperm karyotyping after penetration of zona-free hamster eggs (13) and from use of the polymerase chain reaction method (24, 25). We obtained labeling efficiencies of >96% in every sample, which is an improvement on our previous studies (17, 20, 21). The unlabeled sperm are either nullisomic or simply remain unlabeled for technical reasons and it is impossible to distinguish the two reasons without using a third autosomal probe. Nevertheless, a high labeling efficiency ensures a minimum bias to the X:Y ratios. We attribute the improvement to more efficient nuclear decondensation, which resulted from treating aliquots of each sample with three different concentrations of DTT and using the aliquot that responded best for in situ hybridization. This overcame the variability between samples that we observed in our earlier studies (17, 20, 21). There are other pretreatment methods that have been ap- plied successfully to human sperm (19) but, in our experience, EDT A followed by DTT is simple and effective. In conclusion, these experiments using double-label fluorescence in situ hybridization clearly demonstrate that discontinuous albumin gradients do not enrich Y -bearing sperm, but in fact provide a slight enrichment of X-bearing sperm. However, to compensate for any possible bias because of an inaccurate duplication of the methods, we have commenced a multicenter, blinded collaborative trial to evaluate sperm sex selection methods that are performed under license as a routine procedure. We believe that the results of this prospective trial will determine with certainty whether laboratory procedures such as albumin gradients alter the sex ratio of human sperm. Acknowledgments. We thank the staff of the Andrology Laboratory at The Queen Elizabeth Hospital for assistance with donor recruitment and semen processing, Judy Ford, Ph.D., of the Genetics Department at The Queen Elizabeth Hospital for supplying the Y chromosome probe, and Jim Wang, Ph.D., for advice on statistics. We also thank other members of the Department of Obstetrics and Gynaecology for assistance, including Tie-Lan Han, Ph.D., for advice on the in situ hybridization procedure and Tony Simula, Ph.D., for help with preparation of the probes. REFERENCES 1. Ericsson RJ, Langevin CN, Nishino M. Isolation of fractions rich in human Y sperm. Nature 1973;246: Iizuka R, Kaneko S, Aoki R, Kobayashi T. Sexing of human sperm by discontinuous Percoll density gradient and its clinical application. Hum Reprod 1987;2: Steeno 0, Adimoelja A, Steeno J. Separation of X- and Y bearing human spermatozoa with the sephadex gel-filtration method. Andrologia 1975;7: Dmowski WP, Gaynor L, Rao R, Lawrence M, Scommegna A. Use of albumin gradients for X and Y sperm separation and clinical experience with male sex preselection. Fertil SteriI1979;31: Quinlivan WLG, Preciado K, Long TL, Sullivan H. Separation of human X and Y spermatozoa by albumin gradients and Sephadex chromatography. Fertil Steri!1982;37: Check JH, Katsoff D. A prospective study to evaluate the efficacy of modified swim-up preparation for male sex selection. Hum Reprod 1993;8: Beernink FJ, Ericsson RJ. Male sex preselection through sperm isolation. Fertil Steril 1982;38: Corson SL, Batzer FR, Alexander NJ, Schlaff S, Otis C. Sex selection by sperm separation and insemination. Fertil Steri! 1984;42: Beernink FJ, Dmowski WP, Ericsson RJ. Sex preselection through albumin separation of sperm. Fertil Steri! 1993;59: Evans JM, Douglas TA, Renton JP. An attempt to separate fractions rich in human Y sperm. Nature 1975;253: Ross A, Robinson JA, Evans HJ. Failure to confirm separa- Vol. 61, No.4. April 1994 Wang et al. Sperm separation on albumin gradients 725

7 tion of X- and V-bearing human sperm using BSA gradients. Nature 1975;253: Schilling E, Lafrenz R, Klobasa F. Failure to separate human x- and y-chromosome bearing spermatozoa by sephadex gel-filtration. Andrologia 1977;10: Brandriff BF, Gordon LA, Haendel S, Singer S, Moore DH II, Gledhill BL. Sex chromosome ratios determined by karyotypic analysis in albumin-isolated human sperm. Fertil Steril 1986;46: Veda K, Yanagimachi R. Sperm chromosome analysis as a new system to test human X- and Y -sperm separation. Gamete Res 1987;17: Beckett TA, Martin RH, Hoar Dr. Assessment of the sephadex technique for selection of X-bearing human sperm by analysis of sperm chromosomes, deoxyribonucleic acid and V-bodies. Fertil SterilI989;52: Jaffe SB, Jewelewicz R, Wahl E, Khatamee MA. A controlled study for gender selection. Fertil SterilI991;56: Han TL, Flaherty SP, Ford JH, Matthews CD. Detection of X- and Y -bearing human spermatozoa after motile sperm isolation by swim-up. Fertil Steril 1993;60: van Kooij RJ, van Oost BA. Determination of sex ratio of spermatozoa with a deoxyribonucleic acid-probe and quinacrine staining: a comparison. Fertil Steril 1992;58: Wyrobek AJ, Alhborn T, Balhorn R, Stanker L, Pinkel D. Fluorescence in situ hybridization to Y chromosomes in decondensed human sperm nuclei. Mol Reprod Dev 1990;27: Han TL, Webb GC, Flaherty SP, Correll A, Matthews CD, Ford JH. Detection of chromosome 17- and X-bearing human spermatozoa using fluorescence in situ hybridization. Mol Reprod Dev 1992;33: Han TL, Ford JH, Webb GC, Flaherty SP, Correll A, Matthews CD. Simultaneous detection of X- and Y -bearing human sperm by double fluorescence in situ hybridization. Mol Reprod Dev 1993;34: World Health Organization. WHO laboratory manual for the examination of human semen and semen-cervical mucus interaction. 2nd ed. Cambridge: The Press Syndicate of the V niversity of Cambridge, 1987: Waye JS, Willard HF. Chromosome-specific alpha satellite DNA: nucleotide sequence analysis of the 2.0 kilobasepair repeat from the human X chromosome. Nucleic Acids Res 1985;13: Cui K-H, Matthews CD. X larger than Y. Nature 1993;366: Lobel SM, Pomponio RJ, Mutter GL. The sex ratio of normal and manipulated human sperm quantitated by the polymerase chain reaction. Fertil Steril 1993;59: Wang et al. Sperm separation on albumin gradients Fertility and Sterility