HLA-B typing by reference strand mediated conformation analysis using a capillary-Based semiautomated genetic analyzer

HLA-B Typing by Reference Strand Mediated Conformation Analysis Using a Capillary-Based Semiautomated Genetic Analyzer David Turner, Sandra Akpe, Juliette Brown, Colin Brown, Alastair McWhinnie, Alejandro Madrigal, and Cristina Navarrete ABSTRACT: The application of reference strand conformation analysis (RSCA) to HLA-A typing using the ABI PRISM 310 capillary based genetic analyzer has recently been described. This study outlines the development and validation of capillary RSCA for HLA-B typing. Mobility values for 93 HLA-B alleles were defined following electrophoresis of known controls through the system. Three fluorescently labelled references, labelled with three different dyes can be electrophoresed simultaneously. The technique was validated by comparing results from 296 cord blood donors with those obtained using reverse SSO. Following capillary RSCA 14.5% of samples required confirmatory typing, compared with a repeat rate of 5.1%

ABBREVIATIONS RSCA reference strand conformation analysis FLR fluorescent labelled reference HLA human leukocyte antigens PCR polymerase chain reaction

INTRODUCTION The HLA genes are the most polymorphic loci known in the human genome and current typing techniques such as sequence-specific primer (SSP) amplification [1] and hybridization with sequence-specific oligonucleotide From the Department of Histocompatibility and Immunogenetics (D.T., S.A., J.B., C.B., C.N.), North London Centre, National Blood Service, London; and The Anthony Nolan Research Institute (A.McW., A.M.), The Royal Free Hospital, London, United Kingdom. Address reprint requests to: Dr. David Turner, Department of Histocompatibility and Immunogenetics, North London Centre, National Blood Service, Colindale Ave, London NW9 5BG, United Kingdom; Tel: ⫹44 (0) 20-8258-2820; Fax: ⫹44 (0) 20-8258-2973; E-mail: david.turner@ nbs.nhs.uk Received October 9, 2000; revised December 12, 2000; accepted January 3, 2001. Human Immunology 62, 414 – 418 (2001) © American Society for Histocompatibility and Immunogenetics, 2001 Published by Elsevier Science Inc.

following reverse SSO. In samples where no other typing was necessary there was 100% correlation between the two methods. Capillary RSCA for HLA-B typing is quick, easy to implement, and with the introduction of new FLRs and gel matrices has the potential to evolve into a high resolution typing method. Human Immunology 62, 414 – 418 (2001). © American Society for Histocompatibility and Immunogenetics, 2001. Published by Elsevier Science Inc. KEYWORDS: RSCA; capillary electrophoresis; donor typing

SBT SSO SSP

sequence based typing sequence-specific oligonucleotide sequence-specific primer

(SSO) probes [2] require large numbers of reagents for high resolution typing. The method of reference strand mediated conformation analysis (RSCA), formerly known as double-strand conformation analysis [3] relies upon the separation of homo- and heteroduplexes in a gel matrix. Heteoduplexes are generated by the annealing of polymerase chain reaction (PCR) product generated from different human leukocyte antigens (HLA) alleles to a common, fluorescently labelled reference (FLR) PCR product. Sequence differences between the sample and the FLR lead to the production of homo- and heteroduplexes following denaturation and cooling. As particular allele/FLR heteroduplexes migrate different distances relative to the homoduplex following electrophoresis, mea0198-8859/01/$–see front matter PII S0198-8859(01)00213-0

HLA-B Typing by RSCA

415

TABLE 1 Mobility values using B*4201, B*4403, and B*5801 FLRs HLA-B

B*4201

SD

B*4403

SD

B*5801

SD

B*0801 B*67011 B*3701 B*3901/5/11 B*4102 B*8101 B*0705 B*0702 B*0707 B*1401 B*1402 B*0802 B*1801 B*3505 B*3512 B*3517 B*1802 B*4801 B*4101 B*3502 B*4803 B*3513 B*3503 B*3501 B*3906 B*3508 B*1507 B*1511 B*1530 B*4409 B*1510 B*1520 B*1532/3 B*5501 B*1301 B*1501 B*5502 B*1518 B*1521 B*1502 B*5601 B*1512 B*5401 B*4601 B*3802 B*1519 B*1503 B*4002 B*4003 B*4004 B*40011 B*40012 B*3801 B*4005 B*7801 B*4703 B*2708 B*4701

536 561 576 578 583 583 589 593 593 597 603 607 611 611 612 612 613 614 615 616 617 617 618 620 622 625 626 626 629 630 633 635 636 636 636 637 637 637 637 639 639 643 646 648 649 649 650 652 655 660 663 665 667 674 681 681 683 684

1.3 3.2 2.5 1.9 2.9 2.3 2.4 2.4 1.0 1.7 1.9 0.9 1.6 4.4 1.5 2.3 2.1 2.4 1.8 1.8 4.0 1.4 2.0 2.0 2.1 2.8 2.2 2.1 1.3 4.4 2.2 1.4 — 2.9 2.7 2.3 2.2 4.3 0.9 3.2 2.5 — 2.9 1.1 3.1 2.4 1.7 3.3 2.9 5.4 2.5 2.5 2.4 2.7 2.9 — 3.1 3.7

695 695 670 668 655 721 729 727 729 691 690 626 719 636 627 629 722 674 714 635 675 634 633 632 722 636 735 745 741 622 743 620 730 785 589 735 779 747 731 713 763 753 809 796 597 747 716 636 630 636 655 659 598 663 751 617 690 574

3.0 2.6 2.7 3.7 1.8 3.9 3.7 4.4 6.6 5.6 2.9 1.8 3.4 2.1 1.5 1.3 3.3 4.3 2.3 2.3 1.7 4.2 3.0 3.0 4.0 4.7 4.1 3.9 1.2 0.7 1.9 3.6 — 3.4 0.5 3.7 5.7 3.0 1.5 5.3 3.4 — 2.7 1.4 1.7 8.8 2.7 3.3 2.1 2.1 2.2 3.4 3.1 4.0 4.5 — 5.0 2.9

794 800 631 749 800 820 825 824 828 769 773 651 729 686 708 690 714 751 861 692 774 690 693 691 800 689 737 765 749 794 755 673 735 857 601 743 848 759 718 713 840 752 870 768 617 749 733 790 781 774 801 792 625 799 755 670 849 665

4.3 3.7 6.0 4.6 6.3 3.7 2.2 3.9 — 6.7 4.9 3.6 5.7 — — 0.1 — 6.6 6.5 2.4 — 2.7 4.0 4.5 5.3 2.9 2.6 5.9 — — 5.1 8.3 — 6.6 4.0 4.8 8.1 4.1 0.8 3.0 2.2 — 1.1 5.9 3.8 12.4 1.9 6.0 2.8 — 6.5 5.4 2.7 5.2 10.5 — 3.7 4.5 (Continued)

TABLE 1 (Continued) HLA-B

B*4201

SD

B*4403

SD

B*5801

SD

B*1504 B*2703/5/10 B*2706 B*2704 B*4402 B*4408 B*4006 B*5301 B*5108 B*1523 B*1524 B*1535 B*4403 B*5901 B*1513 B*2702 B*52011 B*5802 B*5801 B*5102 B*5703 B*1517 B*5701 B*51011 B*5107 B*1302 B*5702 B*52012 B*4501 B*5002 B*5001 B*7301 B*4901 B*1516 B*4201

684 688 688 692 702 707 709 715 716 726 729 738 741 746 747 763 764 765 765 765 766 771 774 775 776 778 779 786 865 912 913 972 1039 1061 —

4.7 3.9 2.0 3.7 2.7 5.4 3.4 4.5 2.0 3.9 3.5 1.2 3.5 4.4 4.4 4.1 4.0 6.0 2.5 7.3 3.7 2.5 2.8 2.8 2.8 3.4 5.3 2.2 4.4 1.8 4.9 4.5 5.9 11.9 —

815 611 609 610 545 552 691 558 699 646 657 737 — 655 624 607 597 592 581 654 656 661 652 660 657 627 661 649 1012 986 985 1035 881 938 738

2.1 2.5 2.4 2.3 1.9 0.7 3.0 4.9 1.7 2.2 3.2 2.3 — 2.9 2.7 3.2 2.0 2.3 2.0 6.1 1.8 2.7 3.1 2.7 6.9 2.0 5.1 2.1 4.4 5.4 4.7 11.5 1.8 3.5 3.1

802 722 698 703 658 649 836 554 653 617 615 742 623 652 570 711 611 — — 615 565 556 561 618 615 636 567 612 1055 1022 1029 1146 850 813 822

0.4 1.6 6.0 11.8 2.5 5.5 3.7 4.1 1.0 0.7 2.1 2.1 2.8 9.3 1.4 4.9 2.1 — — 1.2 1.2 0.8 1.7 2.7 5.3 3.6 2.4 — 4.5 3.5 9.6 0.7 2.4 2.4 2.3

Alleles are shown in ascending order according to the mobility of the heteroduplex generated between the allele and B*4201 FLR. FLR ⫽ fluorescently labeled references; SD ⫽ standard deviation.

surement of mobility values allows alleles to be assigned. The use of an automated genetic analysis platform allows all fragments to migrate the same distance to the point of detection, thereby reducing inter-run variability that has hindered efforts to develop other conformational techniques as typing methods. RSCA has been used for high resolution class I typing, class I and class II matching for bone marrow donor selection, and for typing and matching HLA-DPB1 locus [4 – 6]. Recently, we have described an RSCA method for HLA-A typing using a capillary based ABI PRISM 310 system (Applied Biosystems, Kelvin Close, Birchwood Science Park North, Warrington, Cheshire) [7]. In our laboratory HLA-B typing for cord blood and bone marrow volunteer samples is performed using a commercial reverse dot blot kit, RELI™ (Dynal, Wirral, UK), with ambiguities being resolved by PCR-SSP. In

416

this study we have developed and validated an RSCA method for the typing of the HLA-B locus on a semiautomated capillary based genetic analyzer. To validate the technique, 296 cord blood samples were typed by HLA-B RSCA that had previously been typed by a combination of RELI™ and SSP. MATERIALS AND METHODS DNA Typed DNA reference samples supplied by United Kingdom Transplant Support Service Authority (UKTSSA) were used to establish points of migration for different HLA-B alleles and provided homozygote samples for testing potential FLRs. Alleles that were not represented in the UKTSSA panel were found in DNA from the Terasaki cell exchange [8]. Three hundred sixteen cord blood samples were typed by HLA-B RSCA and RELI typing to compare both methods. PCR Reactions HLA-B locus specific PCR amplifications were performed using the following primers 5⬘BRSCA: 5⬘-GG GAGGAGCGAGGGGACCGCAG-3⬘ [9] and 3⬘BR SCA: 5⬘-TGTTGGTCCCAATTGTCTCCCCTCCTTG 3⬘. 50-␮l reactions contained 67 mM Tris, 16.6 mM NH4SO4, 0.01% Tween 20, 2.0 mM MgCl2, 0.25 mM of each dNTP, 0.25 ␮M each primer and 2.5 U Taq polymerase. For amplification of the FLRs the 5⬘ primer was labelled with either the fluorescent TET, FAM or HEX dyes (Cruachem, UK). The following cycling conditions were used for the PCR: 96°C for 4 min: 33 cycles of 96°C for 30 seconds, 65°C for 45 seconds, 72°C for 60 seconds, and 72°C for 10 min. Heteroduplex Formation and Sample Preparation FAM labelled PCR product from a homozygote B*4201 DNA (RSH), and TET labelled PCR product from a homozygote B*4403 DNA (PIT-DUT) were used as FLRs in initial experiments. HEX labelled PCR products produced from the homozygote cell lines LWAGS (B*1402), C1R (B*3503), KAS011 (B*3701), SWEIG (B*4002), KT2 (B*5101), AMAI (B*5301), and WT49 (B*5801) were tried as a third FLR to split alleles that were not sufficiently separated by the B*4201 and B*4403 FLRs. Heteroduplexes were generated as previously described [7], with the exception that the annealing was performed under the following conditions: 96°C for 5 min, max ramp to 55°C, 55°C for 5 min, max ramp to 4°C. For each test sample, 3.0 ␮l of each FLR generated homo/heteroduplex mix was combined with 0.5 ␮l TAMRA labelled Genescan 2500 ladder (PE Biosystems, Warrington, UK) and 10.0 ␮l of dH2O (19.5 ␮l in

D. Turner et al.

TABLE 2 Use of HLA-B capillary RSCA to type 296 cord blood samples n Concordance with RELI™ Samples requiring repeat typing by alternative method In samples requiring repeat typing RSCA unable to distinguish B*5701 and B*1517 Heteroduplex peaks below threshold of detection by ABI 310 Alleles present in sample not previously seen by RSCA

Percent

253 43

85.5 14.5

30/43

69.8

8/43

18.6

5/43

11.6

total). This mixture was electrophoresed on the ABI PRISM 310 Genetic Analyzer using a 61 cm capillary under the following conditions: injection time ⫽ 15 seconds, injection voltage ⫽ 15 kV, run voltage ⫽ 15 kV, collection time ⫽ 26 min, and heatplate temperature ⫽ 30°C. Control samples with known HLA-B alleles were electrophoresed under the conditions described to provide mobility values for alleles and establish a standard table. RESULTS A total of 93 HLA-B alleles were electrophoresed through the system and mobility values noted. The majority of alleles were run more than twice (94%) and standard deviations calculated for each allele/FLR combination. Table 1 shows the mobility values for the HLA-B alleles, plus standard deviations using the B*4201, B*4403 and B*5801 FLRs. These mobilities were obtained when the following values were used for the fragments of the TAMRA 2500 ladder (PE Biosystems): 488 bp ⫽ 60, 845 bp ⫽ 490, 1133 bp ⫽ 588, 1199 bp ⫽ 625, 1740 bp ⫽ 761, 2180 bp ⫽ 873, and 2499 bp ⫽ 951. In the blind study samples were tested initially using the FLRs B*4201 and B*4403. It was observed that the B*4201 and B*4403 FLRs did not allow discrimination of the following alleles: B*51011/B*5701/B*1517, B*3701/B*3905, and B*5201/B*2702. Seven different homozygote DNA samples were used to select a third reference that would discriminate these alleles. HEX labelled B*5801 allowed discrimination of the B*3701/ B*3905 and B*5201/B*2702 ambiguities, but despite being able to split B*5101 and B*5701 it could not separate B*5701 and B*1517. Three hundred sixteen cord blood samples were amplified using the HLA-B RSCA primers, of which 20 failed in the PCR amplification due to low DNA concentration. Therefore 296 samples were subsequently

HLA-B Typing by RSCA

417

FIGURE 1 Percentage frequency of HLA alleles defined in 283 cord blood samples using HLA-B capillary RSCA.Using the FLRs (B*4201, B*4403, and B*5801), some groups of alleles had overlapping mobility values, therefore only allowing definition to the specificity level (e.g., B*35). In other cases sufficient differences in mobility values between the common alleles of a specificity allowed discrimination (e.g., B*4402 and B*4403, B*4001 and B*4002, and B*4006). Accurate high resolution typing by RSCA will require electrophoresis of all known HLA-B alleles through the system to ensure that no alleles have overlapping mobility values. In 30 samples B*5701 and B*1517 could not be discriminated.

typing (5.1%). Of these, 10/15 had weak positive bands because of suboptimal PCR and in the remaining samples the following types were not discriminated; two cases of B*3501/7/10/11/21 with either B*5201 or B*5107; one case of B*3801 with either B*5501/2/5 or B*5601; one case of B*3501/7/11/23 and B*4901 or B*5001 and B*5301; and one case of B*4402 and B*5001/2 or B*4403/7 and B*4501. The RSCA method is able to resolve these ambiguities.

typed using HLA-B capillary RSCA. Samples were assigned a type after comparison of mobility values against standard tables (Figure 1). The results were compared with those previously obtained using the RELI™ SSO kit. RELI™ SSO typing was performed according to the manufacturers instructions (Dynal). The HLA-B RELI™ strip consists of 56 probes and allows the potential discrimination of 192 HLA-B alleles depending upon the heterozygote combination. Of the 296 samples typed, there was complete concordance between HLA-B capillary RSCA and RELI™ in 253/296 samples. Forty-three samples (14.5%) required confirmatory typing by another method to confirm the type (Table 2). For five samples requiring confirmation, the types had not been electrophoresed through the system previously (B*1507, B*1510, B*1517, B*5107, and B*5108). In 30 cases the HLA-B capillary RSCA was not able to distinguish between B*5701 and B*1517, even with three FLRs as outlined above. Finally, in eight cases the heteroduplex peaks were below the threshold for assignment of mobility values by the instrument; this is related to the amount of PCR product mixed with the FLR, therefore these samples could be considered as PCR failures rather than failures of the technique. In the same 296 samples 15/296 samples required further typing following RELI

DISCUSSION Recently, a large number of novel HLA typing methods have been developed. Many of these, such as sequence based typing (SBT), are becoming established in routine laboratories and some, such as reverse dot blot methodologies, are commercially available. In the future, technologies using oligonucleotide arrays immobilized on glass slides may provide simple methods for accurate high resolution HLA typing [10]. RSCA is a new typing technique that provides high throughput medium resolution typing that can be semiautomated on genetic analyzer platforms. The method described in this study for HLA-B typing, and previously for HLA-A typing [7] on the ABI PRISM 310, has a number of advantages when compared with other systems. First, the method is easy to implement as standard reagents are used at all stages (PE Biosystems). The hands-on time is reduced compared with other techniques; a technician is required only for sample preparation and results interpretation. Because four dyes can be analyzed simultaneously in the ABI PRISM 310, there is the potential to run three FLRs at the same time, thus increasing throughput. Currently, 48 HLA-A or HLA-B samples can be run in 36 h. Developments in the dyes used may, in the future, allow five dyes to be analyzed at the same time. This would

418

enable the simultaneous running of HLA-A and HLA-B RSCA (two FLRs per locus). Capillary RSCA is also inexpensive once the hardware is purchased. The same hardware can be used for SBT, so investment can be justified by the acquisition of two HLA-typing systems. Also, other fragment analysis methods that have relevance to tissue typing laboratories can be developed on the capillary based system. These include variable number tandem repeat analysis, used to assess chimerism post-bone marrow transplant, and single nucleotide polymorphism analysis. Capillary based methods for RSCA analysis are perhaps more applicable to routine HLA typing than methods that require the use of acrylamide slab gels. Developments in hardware mean that 16 and 96 capillary systems are now available that could provide high throughput HLA typing by RSCA. Although the resolution obtained by HLA-B capillary RSCA was not as high as some alternative methods, it is of the level usually sought by laboratories involved in high throughput typing for bone marrow and cord blood registries. Improvements in gel composition that are currently being investigated may improve resolution by giving better separation of heteroduplexes. Also, for true high resolution typing by RSCA, all known alleles would have to be run through the system, to ensure that all alleles have unique mobilities with different FLRs. The strength of the HLA-B RSCA method described in this study may be in reducing the number of samples that have to be tested by expensive high resolution methods by removing samples with more frequently occuring HLA-B types. The method described also allowed the discrimination of certain allele combinations that are not resolved by other techniques.

ACKNOWLEDGMENT

RSCA is a patent pending procedure, developed by the Anthony Nolan Research Institute with Pel-Freez威 Clinical Systems (Milwaukee, WI, USA) as the exclusive licensee.

D. Turner et al.

REFERENCES 1. Welsh K, Bunce M: Molecular typing for the MHC with PCR-SSP. Rev Immunogenet 1:157, 1999. 2. Cao K, Chopek M, Ferna`ndez-Vin˜a MA: High and intermediate resolution DNA typing systems for HLA-A, B, C genes by hybridisation with sequence-specific oligonucleotide probes (SSOP). Rev Immunogenet 1:177, 1999. 3. Argu¨ello JR, Little AM, Pay AL, Avakian H, Gallardo D, Rojas I, Marsh SGE, Goldman JM, Madrigal JA: Mutation detection and typing of polymorphic loci through double-strand conformation analysis. Nat Genet 18:192, 1998. 4. Argu¨ello JR, Little A-M, Bohan E, Goldman JM, Marsh SGE, Madrigal JA: High resolution HLA class I typing by reference strand mediated conformation analysis (RSCA). Tissue Antigens 52:57, 1998. 5. Argu¨ello, JR, Little A-M, Boham E, Gallardo D, O’Shea J, Dodi IA, Goldman JM, Madrigal JA: A high resolution HLA class I and class II matching method for bone marrow donor selection. Bone Marrow Transplant 22:527, 1998. 6. Ramon D, Argu¨ello JR, Cox S, McWhinnie A, Little A-M, Marsh SGE, Madrigal JA: Application of RSCA for the typing and matching of HLA- DPB1. Hum Immunol 59:734, 1998. 7. Turner DM, Poles A, Brown J, Argu¨ello JR, Madrigal JA, Navarrete CV: HLA-A typing by reference strand-mediated conformation analysis (RSCA) using a capillary-based semi-automated genetic analyser. Tissue Antigens 54:400, 1999. 8. Lau M, Terasaki P, Park M: International cell exchange. In Terasaki P, Cecka J (eds): Clinical Transplantation. Los Angeles: UCLA Tissue Typing Laboratory, 1994. 9. Cereb N, Maye P, Lee S, Kong Y, Yang SY: Locus-specific amplification of HLA class I genes from genomic DNA: locus specific sequences in the first and third introns of HLA-A, -B, and -C alleles. Tissue Antigens 45:1, 1995. 10. Guo Z, Hood L, Petersdorf EW: Oligonucleotide arrays for high resolution HLA typing. Rev Immunogenet 1:220, 1999.