Preimplantation HLA Testing

Context Preimplantation genetic diagnosis (PGD) has become an option for couples for whom termination of an affected pregnancy identified by traditional prenatal diagnosis is unacceptable and is applicable to indications beyond those of prenatal diagnosis, such as HLA matching to affected siblings to provide stem cell transplantation.

Objective To describe preimplantation HLA typing, not involving identification of a causative gene, for couples who had children with bone marrow disorders at need for HLA-matched stem cell transplantation.

Design, Setting, and Participants HLA matching procedures conducted at a single site during 2002-2003 in an in vitro fertilization program for 9 couples with children affected by acute lymphoid leukemia, acute myeloid leukemia, or Diamond-Blackfan anemia requiring HLA-matched stem cell transplantation. In 13 clinical cycles, DNA in single blastomeres removed from 8-cell embryos following in vitro fertilization was analyzed for HLA genes simultaneously with analysis for short tandem repeats in the HLA region to select and transfer only those embryos that were HLA matched to affected siblings.

Main Outcome Measures Results of HLA matching and pregnancy outcome.

Results As a result of testing a total of 199 embryos, 45 (23%) HLA-matched embryos were selected, of which 28 were transferred in 12 clinical cycles, resulting in 5 singleton pregnancies and birth of 5 HLA-matched healthy children.

Conclusion This is the first known experience of preimplantation HLA typing performed without PGD for a causative gene, providing couples with a realistic option of having HLA-matched offspring to serve as potential donors of stem cells for their affected siblings.

Preimplantation genetic diagnosis (PGD) has become available as an alternative to prenatal diagnosis in order to avoid the risk for pregnancy termination, because PGD allows selection of unaffected embryos before a pregnancy is established. Despite the need for ovarian stimulation and in vitro fertilization (IVF) to be part of the procedure, PGD has become an acceptable method for avoiding the birth of children with genetic disorders.^1-3 Introduced more than a decade ago, PGD has been applied in thousands of clinical cycles.⁴

At present, more than 100 different genetic conditions are indications for PGD, including a few novel indications for which traditional prenatal diagnosis has never been used.^5-9 This includes preimplantation HLA matching combined with PGD, not only to allow couples to have an unaffected child, but also to select a potential donor progeny for stem cell transplantation.⁸ The approach has previously been applied to avoid the birth of a child with Fanconi anemia (FA), which is a severe autosomal recessive disorder characterized by inherited bone marrow failure requiring bone marrow or cord blood transplantation from an HLA-matched sibling.⁸ This approach resulted in the birth of an HLA-matched child free of FA whose cord blood stem cells were transplanted to the affected sibling with FA, resulting in a successful hematopoietic reconstitution. HLA matching would not be considered appropriate for prenatal diagnosis because of the potential for termination of pregnancy, which could not be justified for the reason of HLA incompatibility.

We describe the first clinical experience of preimplantation HLA matching without PGD of a causative gene, demonstrating the feasibility of this novel approach for stem cell transplantation in siblings with bone marrow failure.

The preimplantation testing procedures were conducted at the Reproductive Genetics Institute (RGI), Chicago, Ill. The RGI was established in 1989 and in 1994 was designated as a Pan American Health Organization/World Health Organization Collaborating Center for Prevention of Genetic Disorders. The study was approved by the RGI institutional review board, which is composed of individuals who are independent of the RGI and which includes 2 obstetricians; 1 clinical geneticist physician; 1 internist; 1 anesthesiologist; 1 individual with a master's degree in genetic counseling; 1 medical laboratory technologist; 2 individuals with doctorate degrees in public health and psychology, respectively; 3 individuals with master's degrees in public relations, industrial relations, and nursing, respectively; and a church pastor. Of these, 6 are women. The institutional review board approval covers protocol and consent forms, which allows publication of the preimplantation genetic diagnosis data presented herein. All sets of parents provided written informed consent for preimplantation genetic diagnosis testing.

Thirteen IVF cycles (initial cycles did not have the desired outcome for 4 couples) were initiated in 2002-2003 at the RGI for 9 couples from the United States and England who had a child affected with acute lymphoid leukemia or acute myeloid leukemia (11 cycles), or Diamond-Blackfan anemia (DBA) (2 cycles). These conditions require bone marrow transplantation and also have been successfully treated by cord blood transplantation.¹⁰ Although mutation testing also may be required in combination with HLA typing in DBA, which can be caused by mutations in the gene encoding ribosomal protein S19 on chromosome 19 (19q13.2) or in another gene mapped to chromosome 8 (8p23.3-p22), the majority of DBA cases are sporadic with no mutation detected,¹¹ such as in both cases included in this study (testing for the mutation was performed elsewhere). Each of the couples desired to have another child—a desire separate from the hope of having another child who could potentially serve as a donor of stem cells for the affected sibling, for whom an acceptable HLA match had not been found. As there is no institutional setting for identification of couples at need for preimplantation HLA testing, all couples requested the test on their own, having learned of the availability of the technique through their physicians or the Internet.

A standard protocol for IVF, which includes a psychological evaluation performed by a psychologist at the RGI, was used in combination with a micromanipulation procedure to remove single blastomeres from the cleaving embryos at day 3, as described elsewhere.¹² HLA genes from the blastomeres were tested simultaneously with short tandem repeats in the HLA region¹³ using a multiplex heminested polymerase chain reaction (PCR) system^12,14 involving only closely linked polymorphic short tandem repeat markers located throughout the HLA region, as shown in Figure 1 (HLA genes: HLA-A1, -A2, -A3, -A24; -A32; HLA-C4, -C6; HLA-B27, -B18, -B35, -B51, -B57; HLA-DRB1*1, -DRB1*7, -DRB1*10, and -DRB1*11. Short tandem repeats: D6S461; D6S276*; D6S299; D6S464*; D6S105; D6S306*; D6S1624*; D6S258; D6S248*; MOG a,b,c,d; RF; D6S265; D6S510; MIB; MICA; TNF a,b,c,d; 62; 82-1; 9N-2; D6S273*; DN; LH1; DQ-CAR II; DQ-CAR; G51152; D6S2447; TAP1; Ring 3CA; D6S439*; D6S291; and D6S426). Figure 1 presents positions of closely linked short tandem repeats throughout the HLA region ordered from telomere (top) to centromere (bottom), allowing accurate HLA typing and identification of possible recombinations, which may lead to misdiagnosis. An example of typing for a marker is given in Figure 2. For each family we selected heterozygous alleles and markers not shared by the parents. Such markers provide the information about the origin of chromosome 6. A haplotype analysis for father, mother, and the affected child was performed for each family prior to preimplantation HLA typing. This allowed avoidance of misdiagnosis due to preferential amplification and allele drop out exceeding 10% in PCR analysis of single blastomeres,^12,14 potential recombination within the HLA region, and a possible aneuploidy or uniparental disomy of chromosome 6, which may also affect the diagnostic accuracy of HLA typing of the embryo. The multiplex nature of the first round of PCR analysis required a similar annealing temperature for the outside primers. Thirty cycles of PCR were performed with a denaturation step at 95°C for 20 seconds, annealing at 62°C to 50°C for a minute, and elongation at 72°C for 30 seconds. Twenty minutes of incubation at 96°C were performed before starting cycling. After cycling, 10 minutes of elongation were performed at 72°C. Annealing temperature for the second round was programmed at 55°C (details of the method have been published^8,12).The applied strategy provided a 100% HLA match, because the selected embryos had the same paternal and maternal chromosome 6 as the affected siblings.

Two to 3 embryos determined to be HLA compatible with the affected sibling were transferred; the other HLA-matched embryos were frozen for future availability. As per the parents' written informed consent, some of the HLA-incompatible embryos not chosen for freezing, and those with inconclusive results, underwent PCR analysis of the whole embryo to corroborate the diagnosis based on blastomere analysis. A follow-up assessment of the HLA match was conducted in the established pregnancies using chorionic villus sampling or amniocentesis. Since the leukemias and the DBA in the affected siblings were sporadic and not associated with chromosome 6, there was no reason for concern that the HLA-matched children would be at risk for the same disease as the affected siblings.

A total of 199 embryos were tested for the 9 couples in 13 clinical cycles performed by the time the data were submitted for publication (15 embryos per cycle on the average), of which 45 (23%) were determined to be HLA matched to the affected siblings. The HLA-matched embryos were available for transfer in all but 1 cycle, for which there were no matched embryos. Overall, 28 HLA-matched embryos were transferred in 12 clinical cycles, resulting in 5 singleton pregnancies (42%) and births of 5 HLA-matched children. Of course, some couples require multiple cycles of IVF to achieve pregnancy. These results suggest that testing of approximately 10 embryos per cycle allows for selection of a sufficient number of the HLA-matched embryos for transfer to achieve a pregnancy and birth of an HLA-matched progeny.

The usefulness of detection of recombination within the HLA region is demonstrated in Table 1, describing the results of HLA typing of 1 of the cycles resulting in the birth of a child HLA matched to a sibling with acute lymphoid leukemia. Of 10 embryos tested simultaneously for 11 alleles within the relevant HLA region in this family, recombination between the alleles for the D6S426 and Ring 3CA markers was observed in embryos 4 and 9, and between D6S276* and D6S510 in embryo 7. Of the remaining 7 embryos, 3 were fully matched (embryos 2, 6, and 8), while the other 4 were HLA incompatible with the affected sibling, as seen from the haplotypes of the mother, father, and affected child.

Figure 3 presents the results of preimplantation HLA matching for DBA. One embryo with maternal recombination (embryo 8) and another with both paternal and maternal recombination (embryo 16) (1 in the allele for the Ring 3CA marker and the other in the allele for D6S426) were detected in testing of 16 embryos (examples of different HLA typing results are shown). In addition, there was another embryo with trisomy 6 (embryo 5) with an extra maternal chromosome 6, making this and the other 2 above also unacceptable for transfer. However, 5 embryos appeared to be HLA matched, of which 2 were transferred, resulting in the birth of an HLA-matched child.

The relevance of aneuploidy testing for chromosome 6 is seen from the results of HLA typing in another cycle, resulting in birth of an infant who was HLA matched to the sibling with acute lymphoid leukemia (Figure 4). Two of 10 embryos tested in this case (examples of different HLA typing results are shown) appeared to have only maternally derived chromosomes 6; 1 had only 1 maternal chromosome (embryo 1), and the other had 2 maternal chromosomes, representing uniparental maternal disomy of chromosome 6 (embryo 2). In addition, recombination between D6S291 and class II HLA alleles was evident in embryo 7, which was also unacceptable for transfer. Of the remaining 7 embryos, only 2 (only embryo 4 is shown) were HLA matched to the affected sibling and transferred, resulting in the birth of an HLA-matched child.

Of only 7 embryos available for testing in another case of DBA, 4 were HLA nonmatched, including the one with recombination between the alleles for the TAP1 and Ring 3CA markers (embryo 8). The remaining 3 HLA-matched embryos (embryos 2, 3, and 9) were transferred, resulting in a singleton pregnancy and birth of an HLA-matched child (Table 2). In the remaining cycle performed for acute lymphoid leukemia and resulting in the birth of an HLA-matched child, 6 HLA-matched embryos were selected from 11 tested, with no evidence of recombination or aneuploidy.

The 5 children born as a result of preimplantation HLA matching (mean age, approximately 1 year) are healthy and are comparable to more than 1000 children born after preimplantation genetic diagnosis procedures, including more than 400 children born after preimplantation genetic diagnosis procedures at our center. The mean birth weight percentile was 47% and the mean birth length percentile was 57%. The 5 children born after preimplantation genetic diagnosis procedures in the current study were confirmed to be HLA matched (per HLA typing of blood) to their affected siblings. One sibling with DBA received transplantation and is no longer red blood cell transfusion dependent, whereas the others are in preparation for transplantation or are in remission.

The data herein show the potential feasibility of preimplantation HLA matching for couples having a child affected with a bone marrow disorder, who may wish to have another child as a potential HLA-matched donor of stem cells for transplantation to the affected sibling. As our data show, HLA-matched embryos were selected and transferred in all but 1 cycle, resulting in pregnancies and birth of HLA-matched children in 42% of the transferred cycles, much higher than the pregnancy rate observed in women undergoing IVF.¹⁵ Despite the relatively high rate of preferential amplification and allele drop out observed in PCR analysis of single blastomeres and potential for recombination within the HLA region described in our material, and a high rate of mosaicism for aneuploidies at the cleavage stage,¹⁶ the introduced approach appeared to be highly accurate in the selection of HLA-matched embryos for transfer. This approach involves a multiplex PCR analysis involving testing for HLA alleles together with short tandem repeat markers within HLA and flanking regions, allowing the avoidance of misdiagnosis due to allele drop out, aneuploidy, or recombination of HLA alleles. Recombination in the HLA region was observed in 4.3% of blastomeres tested, suggesting the existence of a few areas in the HLA region that are prone to a higher recombination rate, which should be detected in order to avoid the risk for misdiagnosis (Verlinsky et al, unpublished data, September 2003). On the other hand, aneuploidy of chromosome 6 was detected in 6.4% of blastomeres, including trisomy (2.2%) and monosomy 6 (4.2%) (Verlinsky et al, unpublished data, September 2003).

As mentioned, preimplantation HLA matching has previously been used together with PGD for FA to select and transfer unaffected embryos, avoiding affected pregnancy and also producing a potential donor progeny for stem cell transplantation for the affected sibling with FA.⁸ In contrast, no testing for a causative gene was performed in the present series of couples, the sole objective being to identify the HLA-matched embryos. Although still highly controversial and even not allowed in some countries, this appeared to be a reasonable option for couples, because only a limited number of the embryos resulting from a hormonal hyperstimulation in IVF are actually selected for transfer anyway. Therefore, instead of selecting embryos for transfer based on morphologic criteria, those unaffected embryos representing an HLA match are selected. There are important ethical issues involving the selection of embryos with different normal parameters such as HLA type; however, preimplantation HLA typing allows for the avoidance of the ethical dilemma of therapeutic cloning.^17,18 The decision about preimplantation HLA typing is presently left with individuals and clinicians, as it is still expensive (approximately $3000 on average for preimplantation HLA typing alone [the cost of IVF can exceed $10 000 in the United States and is covered by insurance in only some states; HLA testing is not covered]) and not obligatorily covered by health insurers.

Our results also demonstrate the prospect for the application of this approach to other conditions requiring an HLA-compatible donor for stem cell transplantation. This provides a realistic option for those couples who are parents of an affected child, who desire to have another child, and who—as a wholly separate issue—hope that the subsequent child might serve as a donor of stem cells for the affected sibling. In addition to sporadic forms of DBA, as well as leukemia, the method may potentially be applied to other conditions; for example, the method might be used by parents who have unsuccessfully sought an HLA-compatible donor for a child with other types of cancer. These expanding indications make preimplantation testing a complement to traditional prenatal diagnosis, allowing parents to avoid inherited conditions and pregnancy termination. At the same time, the evidence suggests that it may now be possible for a pregnancy to have genetic characteristics that may be beneficial for affected individuals in the family.