Optimal polymerase chain reaction amplification for preimplantation diagnosis in cystic fibrosis ((Delta)F508)

BMJ 1995; 311 doi: (Published 26 August 1995) Cite this as: BMJ 1995;311:536
  1. Ke-Hui Cui, scientist in charge of preimplantation diagnosisa,
  2. Eric A Haan, director of departmentb,
  3. Ling-Jia Wang, scientista,
  4. Colin D Matthews, professora
  1. a Department of Obstetrics and Gynaecology, University of Adelaide, Queen Elizabeth Hospital, Woodville, Adelaide, SA5011, Australia
  2. b Department of Medical Genetics, Women's and Children's Hospital, North Adelaide, SA 5006
  1. Correspondence to: Dr Cui.
  • Accepted 12 June 1995


Objective: To evaluate direct polymerase chain reaction amplification of mutation on single embryo cells for the routine preimplantation diagnosis of cystic fibrosis.

>Design: Direct polymerase chain reaction amplification of mutation was performed to identify the cystic fibrosis F508 mutation in human blood DNA, single lymphocytes, embryos, and embryo cells obtained by biopsy. Preimplantation diagnosis was performed for a couple who were heterozygous carriers of the F508 mutation.

Setting: Laboratory for preimplantation diagnosis in a reproductive medicine unit.

Main outcome measure: Correct diagnosis of homozygous normal, heterozygous, and homozygous abnormal DNA of the cystic fibrosis F508 mutation.

Results: 45 blood samples (18 homozygous normal, 17 heterozygous, and 10 homozygous abnormal) and 204 single lymphocytes from known sources showed 100% amplification and were diagnosed correctly. 17 human embryos and 52 normal nucleated embryo cells obtained by single cell embryo biopsy also showed 100% amplification. After a miscarriage of the initial pregnancy (diagnosed at preimplantation to be homozygous normal) in the heterozygous carrier couple, fetal tissue was confirmed to be homozygous normal.

Conclusion: Direct polymerase chain reaction amplification of mutation is a simple, fast, reliable test for the common cystic fibrosis mutation in blood DNA and single cells and should be applicable to routine programmes of general screening, maternal blood examination, and preimplantation diagnosis.

Key messages

  • Key messages

  • Misdiagnoses in sex linked diseases and cystic fibrosis are mainly due to limitations of methods

  • A basic experimental design with 100% correct results when testing blood DNA, single lymphocytes, embryos, and single embryo cells combined with good clinical practice and success in an animal model were the key factors in good outcome of this study

  • Applying direct polymerase chain reaction amplification of mutation in general screening, preimplantation, and prenatal diagnosis should lead to a substantial reduction in cystic fibrosis in the future


Cystic fibrosis is the most prevalent severe inherited disease in white people with an incidence of about one in 2500 births and a carrier frequency of about one in 25.1 The disorder is caused by over 200 mutations in a gene called the cystic fibrosis transmembrane regulator, which is on chromosome 7q31.2 The commonest mutation is the deletion of an amino acid (phenylalanine) at position 508 of the gene,3 which is responsible for 70% of cystic fibrosis cases.

Preimplantation diagnosis is a technique of prenatal diagnosis that avoids medical abortion. Precision of diagnosis--that is, as close to 100% accuracy as possible--is crucial4 and is particularly demanding technically, given that a single copy of the gene must be amplified successfully from a single embryo cell obtained by single cell biopsy. Precision of preimplantation diagnosis has been achieved in sex determination by polymerase chain reaction.5 6 Polymerase chain reaction is a method by which a targeted gene is greatly amplified by a polymerase enzyme with two short leading segments of the gene (primers) and other reagents by many cycles of suitable temperature changes in a thermal cycler. Precision of diagnosis is wholly dependent on the method used. Any inappropriate method is easily detected by basic experiments7 8 9 10 11 and will be shown directly by misdiagnosis.12 A technique that achieves near 100% precision of diagnosis would provide a scientific basis for routine clinical practice in preimplantation diagnosis.13 14 15 16

This paper evaluates direct polymerase chain reaction amplification of mutation in single embryo cells to detect the common cystic fibrosis mutation F508.



Human blood DNA (which contains plenty of genes) was used to determine the ability of our own designed primers to detect the F508 mutation in cystic fibrosis. DNA from the whole blood of 18 homozygous normal, 17 heterozygous, and 10 homozygous abnormal subjects was extracted and prepared for amplification.17 All polymerase chain reaction mixtures were freshly combined and transferred to 0.5 ml tubes. Each tube contained 18 μl sterile distilled water; 4 μl 10x reaction buffer (500 mmol potassium chloride and 100 mmol TRIS-hydrochloric acid per litre, pH 8.3); 1.5 μl of each 10mM deoxynucleotide with either 1.6 μl (0.2 μmol/ml) of each normal gene primer (5'-GGCACCATTAAAGAAAATATCATCTTTG-3' and 5'-AGCTTCTTAAAGCATAGGTCATGTG-3'--designated as N tubes with 1 μl phenol red 0.1 mg/ml water) or 1.6 μl of each F508 mutation gene primer (5'-CTGGCACCATTAAAGAAAATATCATTG-3' and 5'-AGCTTCTTAAAGCATAGGTCATGTG-3'--designated as F tubes); 0.2 μl ampli Taq DNA polymerase (5 U/μl); 6 μl 25mM magnesium chloride; and 4 μl (or 1.5 μg) individual human blood DNA in a final volume of 40 μl. DNA was denatured for six minutes at 94°C followed by 30 cycles of 94°C for one minute, 65°C for one minute, and 72°C for two minutes and a final extension at 72°C for 10 minutes. Amplified DNA results and marker (that is, DNA size control) were examined.


Single lymphocytes were used to test further the precision, sensitivity, and stability of amplification with the designed primers. Sixty eight homozygous normal single lymphocytes (from two homozygous normal males), 78 heterozygous single lymphocytes (from three heterozygous females), and 58 homozygous abnormal single lymphocytes (from two homozygous abnormal males) were diluted and individually transferred to 0.5 ml tubes containing 10 μl buffer (10 mmol TRIS-hydrochloric acid, 50 mmol potassium chloride, and 2.5 mmol magnesium chloride per litre, pH 8.0). After DNA denaturing, freshly made mixture (30 μl) containing 4 μl 10x reaction buffer, 6 μl combined deoxynucleotides, 0.2 μl Taq polymerase, 6 μl magnesium chloride, and 1.6 μl of each outer primer (5'-GCATAGCAGAGTACCTGAAACAGGA-3' and 5'-GACGTTTGTCTCACTAATGAGTGAAC-3’) was transferred to the heat treated tubes containing the single lymphocytes.

The DNA was denatured for six minutes at 94°C followed by 30 cycles of amplification of the common (normal and mutant) sequences covering the mutation points (bases CTT) and temperature processed similar to that for blood DNA amplification. Another batch of N tubes (to detect the normal gene) and F tubes (to detect the mutant gene) were prepared with 40 μl mixture containing the inner normal or mutation gene primers identical with those used in blood testing and placed in 1 μl of first amplification product for 30 cycles of reamplification to identify separately the normal and mutant genes.


Four to eight cell human embryos from 31 couples were used. The embryos were secured by a holding pipette. A drilling pipette was approximated to the zona and acid Tyrode's solution (pH 2.3) gently expelled to dissolve the zona pellucida (the outer layer of the embryo). A single normal nucleated embryo cell was aspirated from each embryo with a biopsy pipette. All aspirated single embryo cells were transferred separately to 0.5 ml tubes for two stages of 30 cycles of amplification (as for single lymphocyte testing). Fifty two normal nucleated embryo cells obtained by biopsy from 52 embryos and a further 17 whole embryos were subjected to polymerase chain reaction amplification.

Preimplantation diagnosis of cystic fibrosis was performed for a couple who were heterozygous carriers of the F508 mutation. briefly, using well established in vitro fertilisation techniques we recovered 18 oocytes and generated 16 embryos by fertilisation. Single cell embryo biopsy was performed on 14 embryos at the 6-10 cell stage. These 14 single normal nucleated embryo cells were subjected to DNA amplification.


Ethical approval was obtained for all aspects of the study.



All N tue solutions (with normal gene primers) used to detect the normal cystic fibrosis gene fragments were red (owing to the phenol red) and easily differentiated from the colourless F tube solutions (with mutant gene primers), which were to detect the F508 mutation. Blood DNA from the 18 homozygous normal subjects showed normal gene fragments only (that is, positive amplification from N tubes, or N bands positive). All blood DNA samples from the 17 heterozygous subjects were positive for both normal and mutation gene fragments (N and F bands positive). All blood DNA samples from the 10 homozygous abnormal subjects were positive for mutation gene fragments (F bands positive) and negative for normal gene fragments (N bands negative). All the bands were bright and clear.

Fig 1
Fig 1

Results of blood DNA amplification in 12 of 45 subjects tested for F508 mutation. NN1 to NN4 denote first to fourth homozygous normal subjects, NF1 to NF4 first to fourth heterozygous carriers, and FF1 to FF4 first to fourth homozygous abnormal subjects. M represents marker pUC 19 and B the blank. N represents normal gene and F mutant gene


All 204 single lymphocytes tested showed 100% amplification and were correctly identified. The 68 single lymphocytes from homozygous normal sources all showed normal gene fragments (N bands) only . All 78 heterozygous single lymphocytes were positive for both normal and mutation gene fragments (N and F bands positive). All 58 homozygous abnormal single lymphocytes were positive for mutation gene fragments (F bands positive) only. Weak common DNA bands (self control bands) derived from the initial set of primers were present.

Fig 2
Fig 2

Results of DNA amplification in 12 of 204 single lymphocytes tested for F508 mutation. NN denotes homozygous normal, NF heterozygous abnormal, and FF homozygous abnormal


Seventeen human embryos and 52 normal nucleated embryo cells obtained by single cell biopsy from a further 52 human embryos showed 100% polymerase chain reaction amplification. All were homozygous normal. The bands were bright and clear.

Fig 3
Fig 3

Results of DNA amplification in seven of 17 single human embryos tested for F508 mutation. All were homozygous normal. (Notations as in figs 1 and 2)

Fig 4
Fig 4

Results of polymerase chain reaction amplification in eight of 52 normal nucleated embryo cells obtained by single cell biopsy from 52 human embryos tested from F508. All were homozygous normal. (Notations as in figs 1 and 2)

The results in the couple having clinical preimplantation diagnosis (August 1993) weretwo homozygous normal, seven heterozygous, and five homozygous abnormal embryos with 100%polymerase chain reaction amplification. Two homozygous normal embryos were transferred to the mother's uterus and five heterozygous embryos frozen. Pregnancy was established andconfirmed by rising human chorionic gonadotrophin concentrations. At seven weeks, however, ultrasound examination indicated a failing pregnancy and uterine curettage was performed. Fetal tissue was confirmed to be homozygous normal.


At present the most favourable material for use in preimplantation diagnosis is singleembryo cells obtained by single cell biopsy.16 Gene analysis of single embryo cells should be simple, fast, sensitive, stable, and precise.A simple technique can help avoid technical misdiagnosis and a rapid diagnosis can allow embryo transfer to be performed in the same in vitro fertilisation cycle. Sensitivity is especially important for gene amplification from the DNA of a single embryo cell, and polymerase chain reaction amplification is the most suitable technique. The stability of the technique allows the visualisation of bright, identical, and clear amplification products, thus avoiding the need for frequent changes of primers or reaction conditions. Precision of the diagnosis is the end point and the key to preimplantation diagnosis. Without precision, confirmatory procedures for prenatal diagnosis may be required. Furthermore, some couples object to aborting a fetus that they know to be genetically abnormal despite a wrong diagnosis of normality.

Low amplification rates in sex determination18 and incomplete amplification in the heterozygous condition--for example, resulting from “allele drop out,” when one of the normal or mutant genes cannot be amplified--are closely related to a high fetal misdiagnosis rate. (Misdiagnosis is defined here as the actual fetal genetic condition not being consistent with the results of preimplantation diagnosis.) However, though the rate of amplification with the direct polymerase chain reaction is 100%, the possibility of misdiagnosis as a result of technical mistakes--for example, wrong labelling of tubes--or contamination, chromosomal abnormality, or gene mutation could still occur and should be explained to patients. The risk of misdiagnosis can be almost eliminated by strict procedures.

This study followed studies in a mice model (39 mice pups born with correct sex diagnosis) which included safety analyses5 and a study of human sex determination,6 in which 100% amplification and correct diagnoses were achieved. Apart from experimental design,5 6 factors such as amplification conditions, quality of reagents (such as 10x buffer and magnesium chloride), quality control of the thermal cycler, single cell transfer, and even conscientiousness and skills of the operator are closely related to the amplification rate and precision of diagnosis. In human preimplantation diagnosis, selecting embryo cells with normal nuclei (rather than those degenerating or degenerated cells with nuclear morphological change) for single cell embryo biopsy is also crucial for 100% amplification.19

When DNA amplification is applied to blood or tissue samples many copies of the target sequence are present, facilitating a reliable and correct diagnosis. By contrast, sex determination in preimplantation diagnosis requires amplification of a single male cell which contains only one copy of the Y linked testis determining gene.6 Precise diagnosis from a fragment of the single copy target of this gene in a single cell is clearly more difficult. Furthermore, precise diagnosis of cystic fibrosis in preimplantation diagnosis requires amplification of a single copy target to detect a difference of only three bases and is more difficult than detecting a difference of more than 100 bases to sex an embryo. Thus the conditions of amplification for cystic fibrosis need to be more rigorous than those in sex determination. In these experiments conditions allowed 100% amplification and precision in the diagnosis of the cystic fibrosis mutation with single embryo cells.


The risk of fetal misdiagnosis due to contamination in the diagnosis of X linked diseases is very small.6 Cystic fibrosis is as example of an autosomal recessive mutation disease in which misdiagnosis due to contamination can be analysed . Ifonly homozygous normal embryos are to be transferred without loss of the embryo cell (or DNA), then no fetal misdiagnosis will occur by contamination. There are four reasons. Firstly, when a normal gene contaminates a homozygous abnormal embryo the DNA result is a heterozygous embryo. This embryo would not be transferred. Secondly, when a normal gene contaminates a homozygous normal or heterozygous embryo (about 1% in our study) there is no alteration of the DNA diagnosis. Thirdly, when a homozygous abnormal gene contaminates a homozygous abnormal or heterozygous embryo there is also no alteration of the DNA diagnosis. Fourthly, when a homozygous abnormal gene contaminates a homozygous normal embryo, the DNA misdiagnosis would occur but the embryo would be wasted and not transferred.

Fig 5
Fig 5

Theoretical consequences of contamination in preimplantation diagnosis of cystic fibrosis (F508)

If only homozygous normal embryos are to be transferred with loss of the embryo cell (or DNA), or if heterozygous embryos are to be transferred, then in the 1% of cases in which contamination might occur there would be homozygous normal, heterozygous, and affectedbabies in a ratio of 1:2:1 when both parents are heterozygous carriers. It should be important to set up five or 10 blank control tubes to identify the amount of contamination inthe same batch of samples for further calculation (to be detailed in a future report). Ifno contamination is found in the blanks of that batch and the cell contains a normal nucleus, then the diagnosis is also very precise, allowing heterozygous embryos to be confidently transferred.


Several different methods have been examined for the diagnosis of cystic fibrosis. Some primers were designed towards locus CS.7, which is located only close to rather than directly covering the mutation points of cystic fibrosis.20 This method was later abandoned.

A method of direct electrophoresis of 95/98 bp (mutant gene/normal gene) bands by size difference and another method of heteroduplex formation (namely, normal--mutant gene combination) yielded an amplification rate of 84% in single human polar body and embryo cell experiments but with unclear and unstable bands.21 22 Until recently the reason for heteroduplex rather than homoduplex formation (that is, normal-normal or mutant-mutant gene combination) was not thoroughly understood. The method was originally published (in 1975) for location of the position of the mutation and for sequencing23 but only around half to three quarters of all mutations could be detected in the β globin gene.24 25 The fidelity of the method was adversely influenced by the poor discrimination (or separation) rate (at the same separation temperature) of the normal and mutant homoduplexes of the cystic fibrosis gene, even with blood DNA.26 27 In single cell experiments for the diagnosis of cystic fibrosis a correct diagnosis rate of only 80% (48/60)28 rather than 100%29 was reported.

Under these conditions the heteroduplex method has been used for preimplantation diagnosis in humans but the amplification rate was only 75% (12/16) with nucleated embryo cells.29 It is clear that failure of amplification was due to failure of the technique and failure to differentiate normal and abnormal nuclei. Hence the likelihood of misdiagnosis was very high. Two misdiagnoses were reported with the same heteroduplex method for cystic fibrosis. The primers used29 were also not suitable for single cell amplification (unpublished data). Thus a reported birth29 does not mean that these three basic technical problems have been solved. Routine use of the two cell embryo biopsy18 29 to patch these shortcomings of the diagnostic method is also inappropriate owing to safety problems. The use of “a series of pregnancies” (human fetal life) first rather than basic scientific experiments “to assess the method”29 (or design) raises important ethical questions and has been the subject of severe and increasing criticism.4 9 13 14 15 30

In preimplantation diagnosis of cystic fibrosis polar body biopsy and double biopsy (polar body biopsy followed by embryo biopsy in the same embryo) have been reported.22 It was recognised that polar body biopsy “will not make it possible to establish genetic diagnosis of the embryo” because (a) paternal alleles cannot be analysed, (b) the test might not be efficient owing to loci crossing over, and (c) sex determination is not possible.22

This paper was presented at the 12th annual scientific meeting of the Fertility Society of Australia, Sydney, 2-6 November 1993. We thank Dr Phillip Morris, Mr Paul Nelson, and Mr William Carey, of the department of chemical pathology, Women's and Children's Hospital, for blood samples and the in vitro fertilisation team at Queen Elizabeth Hospital for their help.


  • Funding The study was funded from the budget of the University of Adelaide's department of obstetrics and gynaecology.

  • Conflict of interest None.


View Abstract