Strategies to respond to polymerase chain reaction deoxyribonucleic acid amplification failure in a preimplantation genetic diagnosis program

Strategies to respond to polymerase chain reaction deoxyribonucleic acid amplification failure in a preimplantation genetic diagnosis program William Norfolk,

E. Gibbons,

MD,

Susan

A.

Gitlin,

MS,

and

Susan

E. Lanzendorf,

PhD

Virginia

OBJECTIVES: Our purpose was to identify and evaluate practical methods within a preimplantation genetic diagnosis program that will increase the percentage of embryos for which a genetic diagnosis can be obtained, including clinical responses after failure of deoxyribonucleic acid amplification has occurred. STUDY DESIGN: Known human lymphoblast cell lines and human embryo blastomeres were evaluated in a single-cell, nested primer polymerase chain reaction system with primer sequences for the specific locus surrounding the four base pair insertion mutation on exon 11 of 3-hexosaminidase A-Tay-Sachs disease, the AF508 mutation of cystic fibrosis, and the sex-determining region on the Y chromosome. Reamplification polymerase chain reaction with standard polymerase chain reaction and primer extension preamplification was performed in deoxyribonucleic acid preparations after previous polymerase chain reaction amplification attempts had resulted in failure of amplification. RESULTS: The amplification efficiency of Tay-Sachs disease, 51% (97/l 87), was significantly lower than that for cystic fibrosis, 85% (87/l 07), and for the sex-determining region on the Y chromosome, 85% (77/90). Tay-Sachs disease polymerase chain reaction amplification occurred in 51% of one-cell lymphoblasts, 89% of two-cell lymphoblasts, and 94% of samples when more than two cells were processed together. When previous amplification failure had occurred, standard Tay-Sachs disease polymerase chain reaction resulted in an amplification efficiency of 16% (three of 19), whereas primer extension preamplification polymerase chain reaction for Tay-Sachs disease resulted in amplification of 52% (31/59) lymphoblasts and 54% (13/24) of polyspermic human blastomeres. Four of six human blastomeres in which amplification failure occurred in a Tay-Sachs disease preimplantation genetic diagnosis cycle amplified by primer extension preamplification polymerase chain reaction, which increased the diagnostic information obtained from four to six of the seven embryos on which biopsy was performed. CONCLUSIONS: We suggest that practical approaches for consideration within a clinical preimplantation genetic diagnosis program to limit the net effect of amplification failure (i.e., reduced embryo transfer number) include increasing the deoxyribonucleic acid content in the polymerase chain reaction tube by using more than one blastomere and by using primer extension preamplification when the initial attempt at amplification fails. (AM J OBSTET GYNECOL 1995;172:1088-96.) Key words: Preimplantation acid amplification failure

genetic

diagnosis,

polymerase

With expanding knowledge of the genetic basis of disease processes, evolution of techniques in the disciplines of molecular biology, and advanced reproductive techniques, the ability to diagnosis genetic diseases antenatally has broadened to include preimplantation From the Jones Institute for Reproductive Medicine, Department of Obstetrics and Gynecology, Eastern Virginia Medical School. Charles Hunter Award, presented at the Thirteenth Annual Meeting of the American Gynecological and Obstetrical Society, Hot Springs, Virginia, September 8-10, 1994. Reprint requests: William E. Gibbons, MD, Department of Obstetrics and Gynecolog) Eastern Virginia Medical School, 601 Colley, Norfolk, VA 23507-1912. Copyright 0 1995 by Mosby-Year Book, Inc. 0002-9378195 $3.00 + 0 6/6/62732

1088

chain reaction,

deoxyribonucleic

diagnosis.’ Preimplantation genetic diagnosis involves embryo biopsy after in vitro fertilization (IVF) with subsequent evaluation of the genotype by techniques that include polymerase chain reaction or fluorescence in situ hybridization. The polymerase chain reaction is an elegant, sensitive technique that can produce (amplify) thousands or millions of copies of a desired segment of a specific gene.‘. 3 Preimplantation genetic diagnosis imposes rigorous diagnostic criteria on the genetic diagnostic techniques. For conventional antenatal genetic diagnostic procedures, there are hundreds of fetal cells from which to extract the genetic diagnosis rather than one or two cells of a six-to-eight-cell embryo. In contrast, for accurate diagnosis from a single

Volume 172, Number Am J Obstet Gynecol

4, Part 1

cell the practical capacity of the techniques must frequently equal the theoretic capacity. To move from the research phase of preimplantation genetic diagnosis to a standard of care in preventive medicine, practical improvements in diagnostic efficiency (reliability, timeliness, accuracy, and cost) are essential. Successful genetic diagnostic techniques must function within the limits of IVF. Although a missed diagnosis with birth of an affected child could be catastrophic, neither can a negative pregnancy test after extensive preimplantation genetic diagnosis be considered a successful outcome. Thus preimplantation genetic diagnosis can be applied practically only when our methods yield both a high potential for pregnancy and confidence in the accuracy of the genetic testing. Here we offer specific strategies toward these ends. For the polymerase chain reaction techniques used to evaluate single-gene mutations, one of the most common reasons for an inability to obtain the genetic diagnosis of the embryo undergoing biopsy is failure of deoxyribonucleic acid (DNA) amplification of the desired gene locus. Reasons for amplification failure are poorly understood but include variations in the secondary structure of the DNA of a gene that inhibit the ability of an artificially produced, complementary base pair sequence or primer to successfully anneal to its complementary site on the specific gene of interest.4 The purpose of this study was to identify and evaluate practical methods within a preimplantation genetic diagnosis program that will increase the percentage of embryos for which a genetic diagnosis can be obtained, including clinical responses after failure of amplification has occurred. We have observed that some of the variation in amplification efficiency demonstrates a pattern that differs for different genes, or more specifically, amplification regions around the mutation site of different genes. Further, we have determined various strategies to strengthen the clinical capacity to arrive at a genetic diagnosis that would allow for embryo transfer and the opportunity of a successful pregnancy.

Material and methods All procedures were performed under protocols approved by the Eastern Virginia Medical School Institutional Review Board. Source of DNA and DNA preparation. Genomic DNA and single cells were isolated from three sources: (1) lymphoblast cultures of known DNA composition were obtained from the human mutant cell repository (Coriell Institute, Camden, N.J., repository no GM 03441 normal cells, GM 03770-TSD heterozygous cells, GM 03585-TSD homozygous affected or abnormal cells), (2) blastomeres obtained from growtharrested, polyspermic human embryos, or (3) blastomeres obtained from a couple undergoing preim-

Gibbons,

Gitlin,

and

Lanzendorf

1089

plantation genetic diagnosis for Tay-Sachs disease. For some positive control experiments when cell number -and DNA amount should not be a variable, a “genomic” extract was made by pooling DNA from the known lymphoblast cultures. DNA was extracted by use of proteinase K (3 IJ,~of a 10 mg/ml solution added to 500 ~1 of buffer). The optical density of the sample was measured with a spectrophotometer with an ultraviolet light source (Beckman Instruments, Fullerton, Calif.). The sample was then diluted to a working concentration of 0.1 pg of DNA in a 2 l.~l volume.5 Embryo biopsy. Biopsies were performed on a Nikon Diaphot inverted microscope (Nikon, Melville, N.Y.) equipped with Hoffman optics (Modulation Optics, Greenvale, N.Y.) and Narishige manipulators (Narishige USA, Greenvalie, N.Y.). Three sets of Narishige MN-2 manipulators and MO-202 micromanipulators were used to provide control of three pipettes simultaneously. Immobilization of the preembryo with the holding pipette was achieved with a Narishige IM-6 microinjector. Delivery of acidified medium and aspiration of blastomeres was controlled by a Narishige IM200 pneumatic microinjector. Micropipettes were prepared with borosilicate glass (Sutter Instrument, Novato, Calif.; outer diameter 1.0 mm, inner diameter 0.75 mm) with a Sutter P-87 pipette puller (Sutter Instrument) and a Nikon MF-9 microforge. Holding pipettes were fire polished to have an inner diameter between 35 and 40 km. Pipettes for delivery of the acidified medium (acid pipette) and blastomere aspiration (biopsy pipette) were prepared to have inner diameters of 10 and 35 to 40 pm, respectively. Biopsies were performed in prepared micromanipulation dishes containing drops of Dulbecco’s phosphate-buffered saline solution (Life Technologies, Grand Island, N.Y.) supplemented with 10% human serum albumin (Irvine Scientific, Santa Ana, Calif.) and covered with washed, equilibrated mineral oil. The acidified medium was prepared by supplementing the Dulbecco’s phosphate-buffered saline solution with polyvinylpyrrolidone (0.4 gm/lOO ml; ICN Biochemicals, Cleveland; molecular weight 360,000) and decreasing the pH to 2.5. Biopsies were performed by immobilizing each preembryo while an opening was created in the zona pellucida with the release of acidified medium through the acid pipette. The biopsy pipette was then introduced through, the opening, and one to two blastomeres were slowly aspirated. After blastomere removal each preembryo was washed three times and then returned to culture. The excised blastomeres were individually washed three times in Dulbecco’s phosphate-buffered saline solution to remove any contaminating sperm. Blastomeres were then added to a 0.5 ml

1090

Gibbons,

Gitlin,

and

April 1995 Am J Obstet Gynecol

Lanzendorf

polymerase chain reaction for exon 11 site Table I. Oligonucleotide primer sequences used in “standard” of 4 bp insertion mutation of B-hexosaminidase A gene (Tay-Sachs disease), the sex-determining region on Y chromosome, and gene locus around AF508 mutation of cystic fibrosis transmembrane regulator

I Hexosaminidase A Tay-Sachs disease Sex-determining region on Y chromosome Cystic fibrosis

Outer

ptimers

5’

to 3’

CAACAACAGTCTGGTGATGG GCCAGACACAATCATACAGG CAACGTCCAGGATAGAGTGA CCCTGTACCAACCTGTTGTCC GACTTCACTTCTAATGATGAT CTCTTCTAGTTGGCATGC

I

Nested primers 5’ to 3’ AAGGAGCTGGAACTGGTCAC CTCTCAACCACCTTCCCAAT CCAGGAGGCACAGAAATTAC TCCGACGAGGTCGATACTTA TGGGAGAACTGGAGCCTT GCTI-TGATGACGCT-TCTGTAT

Primers used for first 20 polymerase chain reaction cycles are shown in middle panel (outer primers). Internal or nested primers used in final 30 polymerase chain reaction cycles are shown in right panel.

GeneAmp reaction tube (Perkin-Elmer, Norwalk, Conn.) containing 60 ~1 of water while being visualized under a dissecting microscope, to assure that the cells were successfully transferred. Two microliters of the last medium rinse drop of each blastomere was added to another tube to serve as a contamination control. All samples were then frozen-thawed twice in liquid nitrogen and then boiled for 10 minutes at 100” C to lyse the blastomeres. All samples were stored at -20” C until use. Primer design. Primers were designed with Primer Designer software (S&E Software, Stateline, Pa.) by use of selection criteria as outlined by Saiki.4 Table I lists the outer and inner (nested) primer pairs for the following genes: (1) the specific locus surrounding the four base pair insertion mutation on exon 11 of B-hexosaminidase A-Tay-Sachs disease, (2) specific locus for the AF508 mutation of cystic fibrosis transmembrane conductance regulator, which is a 3 bp deletion causing a loss of a phenylalanine residue at amino acid position 508, and (3) the sex-determining region on the Y chromosome. Standard polymerase chain reaction method Reagents. A polymerase chain reaction master mix, which consisted of polymerase chain reaction buffer (1.5 mm magnesium chloride, 10 mmol/L Tris-hydrochloric acid at pH 8.3, 50 mmol/L potassium chloride, and 0.01% [wt/vol] gelatin), 200 umol/L deoxyribonucleoside triphosphate (USB, Cleveland), 0.2 pmol/L primers (Genosys, The Woodlands, Tex.), and 0.5 unit of Tuq polymerase (Perkin-Elmer), was used. Tay-Sachs disease. The reagent concentrations and conditions were performed as previously described and optimized by Morsey et a1.,6 except that a “nested” primer (second amplification step with a second set of primers that are internal to the first primer set) procedure was used. The initial 100X reaction volume containing DNA and the master mix was placed into a DNA thermal cycler (Perkin Elmer Cetus, Norwalk, Conn.).

After an initial denaturation of 94” C for 3 minutes, 20 cycles were carried out with 94” C denaturation, 61” C annealing, and 72” C extension steps of 1 minute each. Two microliters of this amplified product was removed and added to a reaction mix of nested primers. The above thermal cycler protocol was followed for 30 cycles, except for a reduction in the annealing temperature to 54” C. Total time for amplification was approximately 5 hours. In addition to the tubes containing DNA from the samples being tested, reaction tubes containing the master mix reagents plus water and oil without added DNA were prepared and run simultaneously with the samples through the polymerase chain reaction process as negative controls against possible contamination from extraneous DNA. Because these controls do not contain any added DNA, no amplification signal should be seen. Amplification in these tubes indicates a contaminating DNA source, which could come from anywhere. Evidence of contamination would invalidate an experimental set. In addition to these controls, aliquots of the third cell rinse (after the lymphoblast was separated from the cell culture or after embryo biopsy) were also placed in reaction tubes and polymerase chain reaction was performed along with the samples as another contamination control. Cysticfibrosis. The polymerase chain reaction reagents for cystic fibrosis were similar to those for Tay-Sachs disease, except that primer sets specific for cystic fibrosis were used. The conditions were performed under the protocol used by Handyside et al.’ except that the concentration of the primers was 0.4 pmol/L and the concentration of the Taq polymerase was 0.5 unit. The annealing and extension temperatures and times for the initial (20 cycles) and nested (30 cycles) polymerase chain reaction cycles were 40” C and 45 secondsl72” C and 90 seconds and 50” C and 45 secondsl72” C and 90 seconds, respectively. Sex-determining region on the Y chromosome. The initial

Volume 172, Number Am J Obstet Gym01

Gibbons, Gitlin, and Lanzendorf

4, Part 1

EMBRYO i -

N

4

+N iA

-

EMBRYO 4

2

+N +A

R -

Normal

1091

R H - pGEM

No Amplification

Fig. 1. Ten percent polyacrylamide gel of heteroduplex mixing experiment of primer extension preamplification polymerase chain reaction DNA-reamplification products from two human bastomeres. Primer sequences were for specific locus surrounding the four base pair insertion mutation on exon 11 of P-hexosaminidase may-Sachs disease. -, Negative master mix (contamination) control, amplification of buffer without added DNA, N, normal, homozygous genomic DNA; R, amplification of final blastomere rinse droplet (contamination control); H, heferozygous genomic DNA for Tay-Sachs disease; 11, lane containing unknown DNA only; +N, lane containing unknown DNA plus known normal homozygous DNA, +A, lane containing unknown DNA plus known homozygous affected DNA; PGEM, lane containing reference size markers; Dx, diagnosis.

and nested primers specific for the sex-determining region on the Y chromosome are shown in Table I. Experiments were run with lymphoblast cultures with a known male genotype. After an initial denaturation of 94” C for 3 minutes, 20 cycles were carried out with 94” C denaturation, 61” C annealing, and 72” C extension steps of 1 minute each. Two microliters of this amplified product was removed and added to a reaction mix of nested primers. The above thermal cycler protocol was followed for 30 cycles, except for a reduction in the a?nealing temperature to 54” C.’ Heteroduplex formation. The size difference (four base pairs) between the amplified DNA product of the normal or abnormal gene is too small to differentiate on a short-duration polyacrylamide gel. When the experimental conditions required discrimination between the normal and abnormal genes for Tay-Sachs disease and cystic fibrosis, mixing experiments were performed to produce a polymerase chain reaction artifact - heteroduplex formation. In the heterozygous condition one allele-has the deletion/insertion, whereas the other allele does not. During the simultaneous amplification of thousands of copies of these two alleles the similarity of their complementary single strands allows them to anneal except for the segment where there are extra (insertion) or missing (deletion) base pairs, which theoretically forms a “loop.” During electrophoresis the “loop” causes the complex to run slower on the gel than the “homozygous” complexes, resulting in a characteristic band pattern (i.e., heteroduplex formation) (Fig. 1).

To detect a difference in the normal versus the affected conditions, the unknown DNA sample was “mixed” with previously amplified product obtained from a known homozygous normal cell and separately with a homozygous affected DNA source (10 yl of amplified product from both the known and unknown DNA samples were placed in a reaction tube). The mixture was denatured at 95” C for 5 minutes, and then the temperature was lowered (65” C for 5 minutes then stored at 4’ C until use) to allow for reannealing of the products. The amplified product of the sample cell with an unknown DNA status was run on three lanes of the gels, alone and one each with the known homozygous normal and homozygous affected/mutant DNA products. The pattern of heteroduplex formation will allow a diagnosis of the unknown sample. If the unknown sample run by itself formed the heteroduplex banding pattern, then the unknown DNA sample contained both a normal and abnormal allele and therefore the “cell” was a heterozygotelcarrier. If the unknown lane alone formed a single band, then the unknown was either homozygous nornial or homozygous affectediabnormal. If the sample mixed with the known homozygous normal DNA formed a heteroduplex pattern, the unknown DNA was homozygous affected. If the heteroduplex formation occurred with mixing of the unknown with the known abnormal DNA, then the “diagnosis” for the unknown was homozygous normal (Fig. 1). Gel electrophoresis. Twenty microliters of amplified product was mixe$ with bromophenol blue dye and loaded onto a 10% polyacrylamide gel. Gel electro-

1092

Gibbons,

Gitlin,

and

Lanzendotf

phoresis was run at 175 V (Fisher Biotech power supply, Pittsburgh) for 90 minutes. The gel was stained with 0.5 mg/ml of ethidium bromide for 15 minutes, destained with water for 5 minutes, and then photographed under ultraviolet transillumination. The time required for the PCR process from initiation of the amplification procedure through to fluorescence of the gel and diagnosis is about 8 hours. Sensitivity experiments. Repetitive polymerase chain reaction experiments were performed on lymphoblasts from known cell lines containing normal and mutated cystic fibrosis and Tay-Sachs disease genes. The polymerase chain reaction experiments were performed from which the DNA source was either one cell, two cells, or more than two cells. Reamplification experiments Reamplijcation procedure by standard polymerase chain reaction. Reamplification refers to submitting a DNA sample in which a previous attempt at polymerase chain reaction amplification had resulted in failure (i.e., the inability to identify an amplified gene product after polymerase chain reaction and electrophoresis). For reamplification testing the initial reaction tube was retrieved from its storage at -20” C. The same master mix described previously for an initial polymerase chain reaction cycle containing primers, Tag polymerase, buffer, etc., was added directly to the tube, and polymerase chain reaction was performed for Tay-Sachs disease or cystic fibrosis as described above for the initial, de novo amplification procedure. Primer extension preampli$cation method. If instead of the initial set of specific gene primers a random mixture of 15-base oligonucleotides is used, it is estimated that at least 78% of the genomic sequences in a single haploid cell can be copied no less than 30 times8 Samples analyzed de novo were collected into a lysis buffer containing 200 mmol/L potassium hydroxide and 50 mmol/L dithiothreitol, heated for 10 minutes at 65” C, then neutralized with a solution of 900 mmol/L Tris-hydrochloric acid (pH 8.3), 300 mmol/L potassium chloride, and 200 mmol/L hydrochloric acid. To the lysed and neutralized sample was added 5 l~,l of a 400 ymol/L solution of random primers (Operon Technologies, Alameda, Calif.), 6 ~1 of 10x K+-free polymerase chain reaction buffer (25 mmol/L magnesium chloride/gelatin [ 1 mg/ml]/lOO mmol/L Tris-hydrochloric acid, pH 8.3), 3 l.~l of a mixture of the four deoxyribonucleoside triphosphates (each at 2 mmol/L), and 1 l.rl of Taq polymerase (Perkin Elmer Cetus, 5 units), and brought to 60 l~,l with water. Fifty primer-extension cycles were carried out, consisting of a l-minute denaturation step at 92” C, a 2-minute annealing step at 37” C, a programmed ramping step of 10 seconds per degree to 55” C, and a 4-minute incubation at 55” C for polymerase extension.

April 1995 Am J Obstet Gynecol

For samples that were subjected to reamplification with primer extension preamplification, the initial reaction tube from the first polymerase chain reaction was stored at -20” after use. Because the reaction volume was 100 ~1 minus the 2 ~1 used for the nested primer polymerase chain reaction cycles, the primer extension preamplification reagent mixture was adjusted from the published protocol to reflect the increased volumes (i.e., the concentrations remained the same). Primer extension preamplification was performed overnight, and then 5 l~,l of product was used for the template for a regular Tay-Sachs disease amplification protocol. Primer extension preamplification adds about 10 hours to the diagnosis process, in addition to the 8 hours for the post-primer extension preamplification polymerase chain reaction. Statistical methods. The amplification results for the experiments for the different genes were compared by x2 analysis. Further, because the data were obtained in multiple polymerase chain reaction experiments (38 separate experiments for Tay-Sachs disease, 22 for cystic fibrosis, etc.) to evaluate whether variability of amplification between different individual experiments could explain part of the variation noted between genes, the amplification-positive proportions were compared between genes by Kolmogorov-Smirnov testing. Results De novo amplification of known lymphoblast populations. The efficiency of single-cell amplification of the Tay-Sachs disease exon 11 locus by polymerase chain reaction was 51%. Amplification or diagnosis was achieved in 97 of 187 cells evaluated. For cystic fibrosis the amplification efficiency was 85% (87/107). For evaluation of the sex-determining region on the Ychromosome locus the amplification efficiency was 85% (77190). The amplification efficiency for Tay-Sachs disease was statistically lower than for both cystic fibrosis (p = 0.0095) or sex-determining region on the Y chromosome (p < 0.0118). Sensitivity evaluation: cell number versus amplification. When polymerase chain reaction was performed with Tay-Sachs disease primers but with increasing numbers of cells, the following amplification efficiencies were observed: one cell, 51% (971187); two cells, 89% (16118); three cells, 100% (17117); four cells, 89% (8/g); five cells, 87.5% (7/8). Combining the amplification efficiencies for the experiments where more than two cells were amplified together demonstrated an amplification rate of 94% (32134). Reamplification experiments Repeat amplification attempt with standard polymerase chain reaction. When amplification was attempted again with Tay-Sachs disease primers after a previous failure

Volume 172, Number Am J Obstet Gym01

Gibbons, Gitlin, and Lanzendorf

4, Part 1

Table II. Results of polymerase chain reaction amplification of site of 4 bp insertion mutation of the p-hexosaminidase A gene (Tay-Sachs disease) in seven human embryos obtained for preimplantation genetic diagnosis

Embryo

code No. 1

Yes/normal

No Yeslhomozygous No* No Yes/normal* No Yes/normal No* No

5 6 7

11

Primer extension preamplzjkation-reamplijcation status and genetic finding

Initial amplzjication status and genetic finding

2 3 4

of exon

1093

No affected Yes/normal Yes/normal Yes/normal No Yes/normal

Results of initial attempt at gene amplification are shown with genetic findings (middle panel). Results of attempts at reamplification of DNA from biopsied blastomeres in which initial amplification attempt failed are shown in right panel. *Designates more than one blastomere biopsied and tested.

in an attempt to amplify the Tay-Sachs disease gene, amplification occurred in only three of 19 attempts (16%). This is in contrast to evaluations in which cystic fibrosis primers were used in the standard polymerase chain reaction after failed Tay-Sachs disease polymerase chain reaction amplification in which successful cystic fibrosis amplification was observed in 18 of 25 (72%) cells (vs 85% for cystic fibrosis polymerase chain reaction de novo, above). Repeat amplijication with primer extension preamplification. In de novo Tay-Sachs disease amplification failures primer extension preamplification Tay-Sachs disease polymerase chain reaction amplification occurred in 52% (31159) of single lymphoblasts and 54% (13/24) of single blastomeres from polyspermic embryos. As a primer extension preamplification control initial de novo amplification (without a documented previous failure) for Tay-Sachs disease was observed in 48% (16/33) of single-cell lymphoblast experiments. Thus these preliminary experiments would suggest that primer extension preamplification allows a “diagnosis” for Tay-Sachs disease in about half of cases where there has been a previous amplification failure and that this efficiency is not different from the primer extension preamplification rate noted de novo. Six biopsied blastomeres plus three negative controls from a couple undergoing preimplantation genetic diagnosis for Tay-Sachs disease in which failure of amplification occurred underwent primer extension preamplification with subsequent reamplification with TaySachs disease primers. Amplification occurred in four of six blastomeres (none of the controls demonstrated contaminating DNA amplification) (Fig. 1). The significance of the reamplification testing is indicated in Table II. At the initial preimplantation diagnosis, biopsy of seven normally fertilized embryos was per-

formed (with more than one blastomere obtained and tested in embryos 4, 5, and 7) with amplification and genetic diagnostic data obtained on four of the seven, which allowed for the transfer of three embryos (and a resulting singleton pregnancy). At the time of primer extension preamplification reamplification, diagnostic information was obtained on two additional embryos. These embryos were cryopreserved and could be transferred at some point in the future. The net effect of the primer extension preamplification reamplification is that information that could be used to determine clinical management was obtained in six of the seven embryos.

Comment Because the “practical” sensitivity of preimplantation genetic diagnostic techniques approach the theoretic sensitivity of the technique, failure of gene amplification is not a rare event.’ Knowledge of the rate of amplification failure is needed to provide for and obtain informed consent. Attempts to enhance amplification efficiencies are continuing in laboratories performing these procedures.‘, I0 Because of the limitations imposed on the preimplantation genetic diagnosis process resulting from variations in oocyte recruitment, fertilization rate, final embryo number, amplification efficiency, and time restraints, strategies are needed to maximize the number of embryos available for transfer. This is because one of the most important variables influencing pregnancy is the number of embryos available for transfer.“. ” When our clinical preimplantation genetic diagnosis program was initiating patient cycles, it was our intent to obtain a second blastomere, when possible, to increase the opportunity of reaching a clinical genetic diagnosis. Our plan was to amplify the blastomeres

1094

Gibbons,

Gitlin,

and Lanzendorf

separately (even in different thermocyclers) to “place our eggs in separate baskets” so that contamination or equipment malfunction would not impede the opportunity to maintain the transfer “receptivity window” for a fresh transfer or require an embryo to undergo another biopsy attempt. With a continuing pregnancy resulting from our first cycle, amplification efficiency was not a significant problem. However, when amplification was observed in only one of eight embryos in our second patient with Tay-Sachs disease, the need to explore other strategies became critical. We have tested the Tay-Sachs disease amplification difficulty from the standpoint of sensitivity by evaluating the amplification efficiencies when the DNA from one, two, or more cells underwent polymerase chain reaction. When two blastomeres are amplified together, the amplification efficiency improved from 50% to nearly 90%. There was not an appreciable gain in efficiency when more than two cells were amplified together (This is important because of the limited number of blastomeres in the embryo.) These findings suggests the following. Because the amplification rate improved when more DNA was present, sensitivity played at least a partial role in the single-cell amplification rate for Tay-Sachs disease. Second, analyzing two blastomeres simultaneously would seem an appropriate clinical strategy in a preimplantation genetic diagnosis program, because the probability of a useful end point (i.e., a clinical diagnosis) appears to be slightly greater by combining two blastomeres than by running them separately (two cells, 85%+ vs one-cell and one cell at 50% each, which would produce a 75% chance of diagnosis for an embryo). These findings are compatible with the theoretic calculations for preimplantation genetic diagnosis performed by Navidi and Arnheim.13 The difference in the amplification efficiencies between the different genes suggests that there is some factor such as secondary gene structure that could antagonize primer attachment or polymerase extension, that is, lowering Tay-Sachs disease amplification efficiency.4 Some of the efforts to improve amplification not detailed were to reduce the stringency of the initial polymerase chain reaction cycles. This lowers the specificity of the process in that the primers would be able to bind to more than just a specific gene, but this does not have to affect the final accuracy of the process if coupled with a “nested” primer polymerase chain reaction system. This could be done by lowering the annealing temperature, which would allow more misfits to occur between the primer and the DNA annealing sites, which the polymerase would begin to copy or amplify. Another strategy is to shift the location on the gene where the primer would anneal and, hopefully, move away from the structural area creating the amplification inadequacy. Another approach would be to decrease

April 1995 Am J Obstet Gynecol

the number of base pairs in the primers. Although this would also increase the possibility that more genes would match the primer sequence, it might also allow a smaller sequence access to bind to the desired DNA binding locus. These permutations were attempted without appreciable benefit (data not shown). Primer extension preamplification represents a variation of the above themes. With primer extension preamplification the intent is to add every possible combination of 15 base pair primers (rather than the 20 base pair primers identified above that are specific to certain gene loci) to attach or anneal to every gene. Furthermore, primer extension preamplification is performed at an annealing temperature that is lower than more gene-specific polymerase chain reaction reactions. Other investigators are now reporting experience with primer extension preamplification within the framework of preimplantation genetic diagnosis.14 Recently Xu et al.” reported on the use of primer extension preamplification to detect multiple genetic loci by use of blastomeres from growth-arrested human embryos. These investigators also evaluated methods to reduce the time required for the process and were successful in reducing the time by 3 hours. When lack of amplification has occurred in a standard polymerase chajn reaction protocol, reamplification with the same primer or gene process was associated with a relatively poor outcome. We observed amplification in only three of 19 samples. Previous amplification with resultant residual master mix components did not prevent subsequent amplification of a different gene. After Tay-Sachs disease amplification failure, 72% of specimens were amplification positive for cystic fibrosis, which was little different (p = 0.7215) than the de novo cystic fibrosis amplification rate of 85%. Reasons for amplification failure include (1) the microscopic size of the cell may result in its not getting transferred into the reaction tube, (2) the blastomere may not contain a nucleus or have a complete complement of chromosomes or genes, (3) DNA structural inhibition, and (4) insufficient polymerase chain reaction sensitivity.+ I3 The large percentage of amplification noted with the cystic fibrosis primers would seem to eliminate the cause being either the failure to transfer the cell or the lack of a cell nucleus. For previous amplification failure, primer extension preamplification resulsted in successful amplification in about half the samples. Why then, if primer extension preamplification is so successful when other techniques have failed, would we not just perform primer extension preamplification de novo? Primer extension preamplification, when used as an initial procedure, in our experience was not more successful than standard polymerase chain reaction. Further, the additional 10 hours required results in important stresses to the

Volume 172, Number Am J Obstet Gynecol

4, Part 1

Gibbons,

limited window for embryo transfer. However, if there is a need to evaluate a large number of genes or multiple gene loci within the same gene when minimal DNA is present, primer extension preamplification represents a logical alternative. Our preliminary experience with amplification failures after Tay-Sachs disease polymerase chain reaction in blastomeres from our preimplantation genetic diagnosis program resulted in a similar primer extension preamplification reamplification rate (four of six), as noted in our lymphoblast experiments. Importantly, the effect of primer extension preamplification was that of extending the number of embryos in which a genetic diagnosis was produced. This could increase the number of embryos that would be available for transfer, thereby affecting the success of the clinical process. An additional benefit obtained from the ability to reamplify the initial blastomere biopsy specimen is that it eliminates the necessity to rebiopsy the embryo and possibly further traumatize it. To our knowledge, our report is the first on the use of primer extension preamplification on previous polymerase chain reaction amplification failures. In a few short years the evolution of genetic techniques paired with advances in assisted reproductive techniques has allowed extension of therapeutic options to the prevention of genetic diseases. However, these techniques create their own difficulties. One major difficulty is proceeding through the entire counselingovulation induction-oocyte retrieval-embryo biopsypolymerase chain reaction process only to obtain inadequate genetic information because of failure of the amplification process. In addition to the general inefficiencies of the process, there also appear to be variations

in

polymerase

chain

reaction

amplification

effi-

ciencies between different genes and gene loci, which can adversely affect the clinical outcome. We suggest that practical approaches for consideration within a clinical limit

preimplantation the

net

effect

genetic

diagnosis

of amplification

failure

program (i.e.,

to

at amplification

and

Lanzendorf

1095

4. Saiki FK. The design and optimization of the PCR. In: Erlich HA, ed. PCR technology: principle and applications for DNA amplification. New York: Stockton Press, 1989: 7-16. 5. Higuchi R. Simple and rapid preparation of samples for PCR. In: Erlich HA, ed. PCR technology: principle and applications for DNA amplification. New York: Stockton Press, 1989:31-8. 6. Morsey M, Takeuchi K, Kaufmann R, Veeck L, Hodgen G, Beebe S. Preclinical models for human embryo biopsy and genetic diagnosis, II: polymerase chain reaction amplification of deoxyribonucleic acid from single lymphoblasts and blastomeres with mutation detection. Fertil Steril 1992;57:431-8. 7. Gitlin SA, Lanzendorf SE, Kaufmann RA, Gibbons WE. Modification of SRY target sequence length in a simultaneous coamplification polymerase chain reaction system to improve resolution [Abstract]. In: Proceedings of the fortieth annual meeting of the Society for Gynecologic Investigation, Toronto, Ontario, Canada, March 31-Apri13, 1993. Toronto: Society for Gynecologic Investigation, 1993:87. 8. Zhang L, Xiangfeng C, Schmitt K, Hubert R, Navidi W, Arnheim N. Whole genome amplification from a single cell: implications for genetic analysis. Proc Nat1 Acad Sci U S A 1992;89:5847-51. 9. Liu J, Lissens W, Devroey P, Van Steirteghem A, Liebaers I. Polymerase chain reaction analysis of the cystic fibrosis delta F508 mutation in human blastomeres following oocyte injection of a single sperm from a carrier. Prenat Diagn 1993;13:873-80. 10. Van den Veyver IB, Chong SS, Kristjansson K, Snabes MC, Moise KJ, Hughes MR. Molecular analysis of human platelet antigen system 1 antigen on prevention of alloimmune thrombocytopenia. AM J OBSTET GYNECOL 1994;170: 807-12. 11. Edwards RG, Fishel SB, Cohen J, et al. Factors influencing the success of in vitro fertilization for alleviating human infertility. J IVF ET 1984;1:3-23. 12. Garcia J, Acosta AA, Andrews MC, et al. In vitro fertilization in Norfolk, Virginia, 1980-1983. J lVF ET 1984;l: 24-8. 13. Navidi W, Arnheim N. Using PCR in preimplantation aenetic disease diasmosis. Hum Reurod 1991:6:836-49. 14. kristjansson K, Ch&g SS, Van den ‘Veyver IB, ‘Subramanian S, Snabes MC, Hughes MR. Preimplantation single cell analysis of dystrophin gene deletions using whole genome amplification. Nat Genet 1994;6:19-23. 15. Xu K, Tang Y, Crifo JA, Rosenwaks Z, Cohen J. Primer extension preamplification for detection of multiple genetic loci from single human blastomeres. Hum Reprod 1993;8:2206-10.

reduced

embryo transfer number) include increasing the DNA content in the polymerase chain reaction reaction tube by using more than one blastomere and using primer extension preamplification when the initial de novo attempt

Gitlin,

fails.

REFERENCES 1. Handyside AH, Lesko JG, Tarin JJ, Winston RML, Hughes MR. Birth of a normal girl after in vitro fertilization and preimplantation diagnostic testing for cystic fibrosis. N Engl J Med 1992;327:905-9. 2. Scharf SJ, Horn CT, Erlich HA. Direct cloning and sequence analysis of enzymatically amplified genomic sequences. Science 1986;233:1076-8. 3. Mullis K, Faloona F, Scharf S, Saiki R, Horn G, Erlich H. Specific enzymatic amplification of DNA in vitro: the polymerase chain reaction. Cold Spring Harb Symp Quant Biol 1986;51:263-73.

Discussion DR. SHERMAN ELLM, Houston, Texas. I have two questions. First, one of the complexities in using polymerase chain reaction in single-cell assays is dealing with compound heterozygotes, for example, if each parent carries different mutations in the cystic fibrosis gene. Have you developed any strategies to overcome this complexity? The second question is in regard to primer extension preamplification. As you indicated, about 15% to 20% of the genome is not amplified. Are you certain that you are amplifying the sequence of interest? DR. GIBBONS.

Certainly

trying

to confirm

the

presence

or absence of the specific target sequence by subsequent polymerase chain reaction cycles is standard, but in actuality your concerns are right. There is a concern

1096

Gibbons,

Gitlin,

and Lanzendorf

that only one allele of the two alleles that may be present will be amplified. So that results in some of the uncertainty of the process and partially explains why we have the patients come to Norfolk for a 2-day orientation to make them aware of the concerns and our capacity to diagnose before they initiate and agree to participate. DR. JOHN E. BUSTER, Houston, Texas. You indicated that getting more cells would be helpful in resolving some of these problems. I agree with that. What strategy have you looked at to get more cells? I don’t know that we have to be locked into a single cell diagnosis. DR. GIBBONS. It is our strategy to attempt to perform a biopsy on two cells. Sometimes this can be easily performed. There are times, however, when it appears that the attempt to try to go for the second blastomere would be potentially damaging to the embryo. One other thing we can try to obtain is an accurate diagnosis within the strict time restraints. So we typically will perform a biopsy of the blastomere starting at the six-cell stage. If the six-cell stage has been reached by the evening of day 2, we will initiate the biopsy attempt. If the six-to-

April 1998 Am J Obstet Gynecol

eight-cell stage has not been reached by the evening of day 2, then we will wait and perform the biopsy the following morning. We’re examining some coculture systems to see if coculturing our blastomeres with some type of epithelial cell line will improve the number of blastomeres that reach the six-to-eight-cell stage rapidly and we can obtain information that would result in a higher cell number or improved development. So we’re looking at some other strategies to try to improve the numbers, but part of it is limited by technique itself. DR. BUSTER. One approach that we have been looking at in Houston is to perform a biopsy at the blastocyst stage where anywhere from 10 to 30 cells are taken from trophectoderm. This avoids the inner cell mass, and therefore the embryonic parts are not injured. Another way might be just to coculture to an even more advanced stage. Morulas are difficult to biopsy but blastocysts are not, and that might be a solution; but perhaps the technique you’re describing that provides more cells might allow us to bring this technique to a really high level of efficiency to make it clinically practical.