Saturday 4 May
1991
No 8749
ORIGINAL ARTICLES
PCR-fingerprinting for selection of HLA matched unrelated
marrow
HLA typing contributes to the delays that occur in the search for HLA-matched unrelated marrow donors, and that result in poor patient survival. A new DNA technique for testing DR match between patient and unrelated marrow donors has been assessed. The technique is based on the formation of heteroduplexes between heterologous amplified coding and non-coding DNA sequences during the final annealing stage of the polymerase chain reaction (PCR), and different HLA-DR/Dw types give unique banding patterns ("PCR fingerprints") on non-denaturing polyacrylamide gel electrophoresis. HLA-DR matching is by visual comparison of patients’ with donors’ fingerprints. Identity can be confirmed by mixing donor and recipient DNA before the final stage of the PCR ("DNA crossmatching"). In an assessment of the technique in 53 unrelated HLA-A and HLA-B matched patient-donor pairs, 42 pairs gave the same results with PCR fingerprinting as with DNA-RFLP analysis. In the 11 other pairs DR/Dw mismatches were detected by PCR fingerprinting but not by the standard DNA-RFLP method; PCR-SSO typing with selected sequence-specific oligonucleotides (SSO) confirmed that mismatches were due to different subtypes of DR4. PCR fingerprinting might thus accelerate the selection of unrelated marrow donors by simplifying the logistics of the donor search. Introduction A proportion of patients with leukaemia or marrow failure can be cured by bone marrow transplantation (BMT) from
donors
HLA-matched unrelated donors1 but delay can reduce chance of success. The time taken to find suitable unrelated donors can therefore influence transplant results.2 The 3 average time taken to find an unrelated donor is 6 months. One important reason for the long interval is the time taken to establish HLA-DR identity between patient and donor. Routine DR typing methods are based on serology, DNA restriction fragment length polymorphisms (DNARFLP),4 and the mixed lymphocyte reaction (MLR).5 DR serological tests require only 6-8 h, but certain specificities are difficult to identify because of a lack of suitable alloantisera. DNA-RFLP typing is more accurate than DR serological techniques, but takes 2 weeks to complete. Furthermore, several specificities, particularly the subtypes of the common DR4 antigen, can be defmed only by supplementing DNA-RFLP typing by use of the polymerase chain reaction (PCR) and sequence specific oligonucleotide (PCR-SSO) probes,9 PCR-RFLP typing,lO or DNA sequencing. At present the final step for assessing DR identity is the MLR, which is difficult to perform and interpret, and is of unknown clinical importance.1 Our recent discovery that PCR amplification products of the DRB gene give unique banding patterns termed "PCR fingerprints" when electrophoresed on non-denaturing polyacrylamide gels offers a simple basis for matching HLA-DR/Dw.11 PCR fingerprints are derived from the variable number of HLA-DRB gene and pseudogene copies in different HLA haplotypes.12,13 During the final annealing stage of the PCR, the single-stranded DNA products of these genes associate, forming both "homoduplexes" between the complementary strands from individual DRB
ADDRESS, UK Transplant Service, Southmead Road, Bristol BS10 5ND (T M. Clay, BSc, J. L Bidwell, PhD, M R. Howard, MB, Prof B A Bradley, MB) Correspondence to Prof B A. Bradley.
1050
HLA-DR MATCHING OF PATIENTS AND DONORS BASED ON
DNA-RFLP TYPE, PCR FINGERPRINT AND DNA CROSSMATCH
Kingdom Transplant Service, Bristol, by use of methods described elsewhere.’
PCR
fingerprinting
Genomic DNA was extracted" from peripheral blood and subjected to DRB gene second exon PCR fingerprinting." Essentially, samples were amplified in a PCR mix containing 10 III of 10 x PCR buffer (100 mmol/1 "tris"-HCI, pH8’3, 500 mmollil KC1, 15 mml/1 Mgd, 0-01% w/v gelatin [Sigma], autoclaved before use), 0-5 amol/1 final concentration of the primers GH46 and GH50, 200 nmol/1 final concentration of each
deoxyribonucleotide triphosphate (dATP, dCTP, dGTP, dTTP), and sterile double-distilled water to a volume of 85 Ill. Taq polymerase (’Amplitaq’, Perkin Elmer Cetus, distributed by ILS Ltd, London) was diluted 1 in 10 in 1 x PCR buffer. PCR mix and diluted amplitaq were irradiated for 17 min with a combination of 254 and 300 nm UV lamps to disrupt any contaminating double-stranded DNA.18 10 ul of target DNA at a concentration of 0’ 1 pg/pl was added to 85 pl of PCR mix, denatured at 94°C for 5 min, and then cooled on ice for 5 min. 2-5 units of diluted amplitaq were added and the samples overlaid with 50 ui of high-meldngpoint bath oil (Sigma). Samples were subjected to 35 cycles at 94°C for 60 s, 65°C for 90 s, and 72°C for 120 s, on a Perkin Elmer Cetus thermal cycler. 10 III of PCR products were electrophoresed on 12% non-
denaturing polyacrylamide gels (PAG) prepared with ’Protogel’ gel Diagnostics, Manville, New Jersey, USA). was done in 1 x TBE ("t-ris"-borate edeticacid)19 at Electrophoresis
mix (National
200 volts for 90 min. Gels were stained with ethidium bromide (0’5 )J.g/ul in lxTBE) and photographed on a UV transilluminator.
DNA
crossmatching
For each DNA crossmatch, 10 pl from each of two PCR products mixed and subjected to heat denaturation at 94°C for 60 s, followed by incubation at 65°C for 90 s and at 72°C for 120 s. The latter two incubation steps were designed to simulate the final stages of the PCR thermal cycling protocol used in PCR fingerprinting.’’ DNA crossmatch samples were electrophoresed as described above. were
-denotes not done *Denotes RFLP or SSO defined mismatch tDenotes discordance between RFLP matching result and PCR fmgerpnnt matching result
genes and "heteroduplexes" between heterologous sequences from different DRB genes. Heteroduplexes move more slowly than homoduplexes during electrophoresis because of regions of mismatched sequence. Variation in the number, size, and position of these mismatched regions influences the extent of retardation. The PCR fingerprints that result are directly visualised with ultraviolet light. The PCR fingerprint produced from DNA of DR heterozygous individuals can be reproduced by mixing the PCR products
Results 53 patient-donor pairs were investigated by PCR fingerprinting. HLA-DR matching by fingerprinting was compared with matching by DNA-RFLP. In 42 out of 53 pairs the results of the two techniques were in agreement, in the other 11 PCR fingerprinting detected DR mismatches not apparent with DNA-RFLP. All 11 mismatched pairs
from two corresponding homozygous DR types. HLA-DR matching can be confirmed in the "DNA crossmatch", whereby patient and donor DNA are amplified separately in the PCR and mixed before the final PCR cycle. If additional bands absent from an individual’s fingerprint are seen in the crossmatch a DR mismatch is present.14 We have studied the potential of PCR fingerprinting as a means of accelerating identification of matched unrelated donors. With the cooperation of clinicians participating in the International Marrow Unrelated Search and Transplant (IMUST) Study15 we have investigated 24 patients and their potential donors, a total of 53 patient-donor pairs.
Methods and materials Patients and donors
Peripheral blood samples from 24 patients and their HLA-A,B matched actual and potential unrelated donors were provided by 18 haematology departments in the United Kingdom. Donors were selected from the British Bone Marrow and Platelet Donor Panel and the Anthony Nolan panel. Samples were collected as part of the IMUST Study protocol.ls DNA-RFLP typing for HLA-DR genes was done in the molecular genetics laboratory, United
Fig 1--PCR fingerprints
in 4
patient-donor pairs.
Patients 6, 7 and 12 have PCR fingerprints identical to those of their respective donors Patient 13 is mismatched with its donor
1051
-in the mismatched pair, not only do the fingerprints give different profiles, but also in the crossmatch the number, intensity, and distribution of the bands have changed with respect to the unmixed samples.
Fig 2-PCR fingerprints from HLA-DR identical but
Discussion
patient and donor who are have different Dw subtypes of DR4. a
Patent 9 DR 72,4
(Dw4), and donor 11s DR F,4(Dw 14.1 ).In DR4 positive heterozygotes, the different Dw subtypes of DR4 can be distinguished by PCR fingerprinting were
DR4
positive. PCR-SSO probing confirmed that the
mismatches detected by PCR fingerprinting were due to the presence of different subtypes of DR4. In 35 pairs judged to be a DR match on fingerprints, DNA crossmatching was performed to confirm DR identity, and confirmation was obtained in all 35 (table). Examples of PCR fingerprinting from 4 pairs showing HLA-DR identity and 1 an HLA-DR mismatch are illustrated in figure 1. Examples of a dOllur-reclpiem pair mismatched for the Dw subtypes of DR4 are illustrated in fig 2; this mismatch was undetected by DNA-RFLP typing. Examples of crossmatch results are illustrated in figure 3
We have evaluated a novel DNA matching technique in 53 unrelated patient-donor pairs. The results were compared with HLA-DR matching by DNA-RFLP, supplemented by PCR-SSO probing for the subtypes of DR4. PCR fingerprinting, confirmed with DNA crossmatching, was more informative than DNA-RFLP alone and equivalent to DNA-RFLP supplemented by PCR-SSO probing. In the DR4 positive pairs PCR fingerprinting alone could discriminate between the different subtypes. We are aware that PCR fingerprinting in its present form may lack discrimination between individuals homozygous for some DR4 subtypes." PCR fingerprinting is a rapid technique, with results being available less than 8 h after DNA isolation, compared with the 2 weeks necessary with DNA-RFLP. The method requires less equipment and technical skill than does conventional HLA-DR typing. It is also cost-effective because radioisotopes, Southern blotting, restriction enzymes, cDNA probes, and oligonucleotide probes are unnecessary.
Primarily because of long delays in finding HLA matched unrelated donors 50% of patients receiving unrelated donor BMT have advanced disease by the time of transplantation and therefore poor chance of survival 22 We are confident that HLA-DR matching by PCR fingerprinting will improve the efficiency of donor searches, and therefore the probability of patient survival after unrelated donor
transplants. We thank Mr Nigel Wood (UKTS) for synthesising the oligonucleotide probes and primers for this study; the following clinicians and staff from the transfusion and immunology services in their hospitals for initiating unrelated donor searches and providing samples for patients included in this study-Dr C. Bailey (Leeds); Dr B. Boughton (Birmingham); Dr J. Chessels, Dr J. M. Hows, Dr J M. Goldman and Dr J. Barret (London); Dr R. Collin (Chesterfield); Dr H. Daly (Truro); Dr S. Durrant (Leicester); Dr M. Joyner and Dr R. Lee (Exeter); Dr P. Darbyshire (Birmingham) ; Prof D. Gardner-Medwin and Dr S. Proctor (Newcastle upon Tyne); Dr K. Hunt (Bradford); Dr A. Oakhill and Dr S. J. Cornish (Bristol); Dr T. Phaure (Stafford); Dr G. Scott (Bristol); Dr J. Whittaker (Cardiff); and Dr D. Winford (Sheffield); Dr Richard Goffin, Mr Andrew Penhaligon, and Miss
Maria Truman of the British Bone Marrow and Platelet Donor Panel for their advice and help; Mrs E. Bidwell and Mrs J. Evans (UKTS) for DNA-RFLP typing; and the Leukaemia Research Fund and EC Concerted Action Programme (grant MR4*-0216-S) for their support.
REFERENCES Bradley BA. The use of unrelated marrow donors for transplantation. Br J Haematol 1990; 76: 1-6. 2. Howard MR, Hows JM, Gore SM, et al. Unrelated donor marrow transplantation between 1977 and 1987 at four centres in the United Kingdom. Transplantation 1990; 49: 547-53. 3. Ruutu T, Goldman JM. A Nordic Registry for volunteer marrow donors? Bone Marrow Transplantation 1990; 5: 273-79. 4. Bidwell JL, Bidwell EA, Savage DA, Middleton D, Klouda PT, Bradley BA. A DNA-RFLP typing system that positively identifies serologically well-defined and ill-defined HLA DR and DQ alleles, including DRw10. Transplantation 1988; 45: 640-46. 5. Bach FH, Hirschorn K. Lymphocyte interaction: a potential histocompatibility test in vitro. Science 1964; 143: 813-14. 6. Wood N, Joysey V, Bidwell J, Klouda P. HLA-DR typing of renal allograft patients and donors by DNA-RFLP: correlation with pre-transplant serotyping. Transplant Proc 1990; 22: 2287. 1. Hows JM,
Fig 3-DNA crossmatch results. (a) 3 patient-donor pairs with negative DNA crossmatch results. The crossmatch result is obtained by mixing the two samples before the final stages of the PCR All 3 patients have PCR fingerprints identical to their respective donors, and crossmatchlng yields a pattern identical to the individual patient and donor fingerprints (b) An example of a positive DNA crossmatch result Individuals A, who is HLA-DR2(w5)-Dw2,DR4-Dw4, and B, who is HLA-DR3(w17)-Dw24.DR4-Dw4, have different PCR fingerprints, and DNA crossmatchmg gives a pattern differing from both individual PCR fingerprints in tenns of number, intensity, and dlstllbutior. of the bands
1052
7.
Mytilineos J, Scherer S, Opelz G. Comparison of RFLP-DR Beta and serological HLA-DR typing in 1500 individuals. Transplantation 1990;
50: 870-73. 8. Middleton D, Savage DA, Cullen C, Martin J. Discrepancies in serological tissue typing revealed by DNA techniques. Transplant Int
1988; 1:
161.
Clay TR, Howard MR, Bidwell JL, et al. PCR-SSO typing for DR4-Dw subtypes: application to unrelated bone marrow transplant donor selection. Eur J Immunogenet 1991; 18: 97-104. 10. Uryu N, Maeda M, Ota M, Tsuji K, Inoko H. A simple and rapid method for HLA-DRB and DQB typing by digestion of PCR-amplified DNA with allele specific restriction endonucleases. Tissue Antigens 1990; 35: 9.
13. Andersson G, Larhammar D, Widmark E, Servenius B, Peterson PA, Rask L. Class II genes of the human major histocompatibility complex. J Immunol 1987; 262: 8748-58.
Clay TM, Bidwell JL. HLA-DR/Dw matching by PCR fingerprinting: the origin of PCR fingerprints and further applications Eur J Immunogenet 1991; 18: 147-53. 15. Bradley BA, Gore SM, Howard MR, Hows JM. International Marrow Unrelated Search and Transplant (IMUST) Study. Bone Marrow Transplantation 1989; 4 (suppl 2): 44. 16. Miller S, Dykes D, Polesky H. A simple salting out procedure for extracting DNA from human nucleated cells. Nucl Acids Res 1988; 16: 14. Wood NAP,
215.
20-31.
11. Bidwell JL, Hui KM. Human HLA/DR-Dw allotype matching by analysis of HLA-DRB gene PCR product polymorphism (PCR "Fingerprints"). Technique 1990; 2: 93-100. 12. Böhme J, Andersson M, Andersson G, Möller E, Peterson PA, Rask L. HLA-DR&bgr; genes vary in number between different DR specificities whereas the number of DQ&bgr; genes is constant. J Immunol 1985; 135: 2149-55.
S, Long C, Erlich H. Sequence analysis of HLA-DR and HLA-DQ loci from three pemphigus vulgaris patients. Hum Immunol
17. Scharf
1988; 22: 61-69. S, Sommer SS. Shedding light on PCR contamination. Nature 1990; 343: 27. 19. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory Press, 1989. New York. 18. Sarkar
Clustering of features of von Hippel-Lindau syndrome: evidence for a complex genetic locus
Hippel-Lindau syndrome (HLS), an autosomaldominant inherited disease, was studied in 92 affected subjects from 29 kindreds. In an initial survey to identify HLS gene carriers, all patients treated at the University of Freiburg for angiomatosis retinae (22), haemangioblastoma of the central nervous system (CNS) (63), and phaeochromocytoma (54) were examined as potential HLS gene carriers. HLS was diagnosed in 86% of the patients with angiomatosis retinae, 19% of the patients with haemangioblastoma of the CNS, and 19% of the patients with phaeochromocytoma. Based on these and on an additional 49 newly diagnosed cases (24 by clinical examination and 25 by pedigree analysis), the calculated prevalence of the disease in the district of Freiburg, Germany, with a population of 1·909 million is 1/38 951. There was a striking tendency for familial clustering of HLS features in affected kindreds. Both angiomatosis retinae and haemangioblastoma of the CNS occurred in most families, whereas renal lesions and/or pancreatic cysts and phaeochromatocytoma were mutually exclusive. This finding suggests that HLS is caused by different mutations within a complex genetic locus, or additional genetic lesions, which cooperate with the HLS gene on chromosome 3p. The data point to a linear sequence of features as follows: phaeochromocytoma, angiomatosis retinae, haemangioblastoma of the CNS, renal lesions, pancreatic cysts, and epididymal von
by the principal features of angiomatosis retinae, haemangioblastoma of the central nervous system (CNS), renal and cancer, cysts pancreatic cysts, and phaeochromocytoma, epididymal cystadenoma.1 Although there are more than 500 published cases,2 the prevalence of the syndrome has not been investigated, nor have the various manifestation patterns of the common lesions been analysed. We here present data on a large series of HLS kindreds and affected descendants.
Patients and methods initial survey, all patients who had presented at the University of Freiburg with symptomatic angiomatosis retinae from 1962-89, a haemangioblastoma of the CNS between 1976 and 1989, or phaeochromocytoma (1971-89) were examined as potential HLS gene carriers. 22 patients had angiomatosis retinae, 63 had haemangioblastomas of the CNS, and 54 had a phaeochromocytoma. All patients underwent extensive pedigree analysis for HLS. 115 of these patients and 75 of their first-degree relatives were assessed by a standard diagnostic programme for additional features of the syndrome. The screening programme consisted of a physical, neurological, and ophthalmological examination, ultrasonography of the abdomen, and a 24-h catecholamine assay. Most of the screened individuals were also studied by computed tomography or magnetic resonance imaging of the skull and of the abdomen and by metaiodobenzylguanidine scintigraphy for lesions of the adrenal glands and the sympathetic trunk. HLS was diagnosed according to the criteria of Melmon and Rosen1 and Neumann3-ie, detection of one of the principal HLS features in addition to a first-degree relative with angiomatosis retinae or haemangioblastoma of the CNS. For 6 HLS patients without affected family members it was not possible to establish unequivocally whether they represented new mutations. In
an
cystadenoma. Introduction von Hippel-Lindau syndrome (HLS) is an inherited autosomal-dominant cancer-prone disorder characterised
ADDRESSES. Department of Medicine, Albert-LudwigsUniversität, Freiburg im Breisgau, Germany (H P H Neumann, MD) and Department of Pathology, University of Zürich, Zürich, Switzerland (O. D. Wiestler, MD) Correspondence to Dr H P H Neumann