How Rare are Rare HLA Alleles

J.-M. Tiercy (1), V. Dubois (2), L. Gebuhrer (2)

1. Transplantation Immunology Unit, National Reference Laboratory for Histocompatibility, University Hospital, Geneva, Switzerland
2. Laboratoire d'Histocompatibilité, EFS Rhône-Alpes Site de Lyon, Lyon, France.

Due to the continuous discovery of new HLA alleles, tissue typing at the allele level is becoming an increasingly difficult task for the clinical laboratories involved in the unrelated bone marrow transplantation setting. In the routine histocompatibility laboratory, high resolution typing is usually achieved by the classical PCR-SSP or PCR-SSOP techniques, mostly using commercial kits. These methods however need to be constantly updated in order to cope with the most recently sequenced variants. The delays in updating the kits might be considered as an obstacle to optimal high resolution HLA typing. It also argues in favour of using sequence-based typing (SBT) for resolving the actual 1183 HLA-A, -B, -Cw, -DRB1/B3/B4/B5, -DQB1, and -DPB1 alleles (IMGT/HLA database, January 29, 2001). The technology might however require a rather large investment in terms of equipment to be compatible with large scale typing in the routine laboratory, and still results in a number of ambiguities when sequencing is performed without prior separation of both alleles. Furthermore no comparative evaluation of the cost using these different techniques is yet available.

Some data from the literature suggest that most recently sequenced HLA alleles are indeed rare (1-3). In order to quantify the occurrence of the supposedly rare HLA variants we have screened our high resolution typing data from the last 2 years (Jan. 1998 to Oct. 2000). All records of patients and bone marrow donors from BMDW Registry that were analysed at the allele level (according to the last HLA Nomenclature Report or to more recent updates), using either SBT, PCR-SSP (Dynal, Genovision), or PCR-SSOP (Innogenetics) were reviewed. A total of 5350 analyses at HLA-A, -B, -Cw, DRB1/B3/B4/B5 and -DQB1 loci were recorded in our two laboratories. Because of low number of tests DPB1 was not included in the survey.

The list of all alleles that were identified is shown in Table 1. Allelic distributions in a given serotype were comparable to those previously published (e.g. ref. 4-7). For example, out of 305 HLA-A2 subtyping tests, no other alleles than A*0201-0207 were identified (level of resolution 0201-0235: Assignment of A*0201 based on PCR-SSP A2 subtyping kit, Jan. 1998-Oct. 2000, would not allow the discrimination of A*0240, 0242, 0245, and 0246 from 0201. However all SBT analyses performed so far for patient and bone marrow donors have confirmed A*0201 assignments.). Subtyping analysis (B*4402-4418) of 195 B44 individuals revealed the presence of only 4 alleles: B*4402-4405. Out of 75 Cw3 positive samples, only Cw*0302, 0303, and 0304 alleles could be detected (resolution: Cw*0302-0312). From 103 Cw*05 and 100 Cw*06 samples only 0501 (resolution 0501-0503) and 0602 (resolution: 0602-0603) alleles, respectively, could be detected. The same holds true for DR: from 392 DR11 individuals, 389 had one of the 4 common subtypes 1101-1104, and the remaining 3 donors were respectively 1108, 1114, and 1127 (resolution: 1101-1135). In DR4 and DR13 patients or donors no other allele was detected besides the common 0401-0408 (resolution: 0401-0433) and 1301-1305 (resolution: 1301-1334) alleles. What is the probability that rare alleles have been mistyped as the most common ones because primers/probes do not detect all relevant polymorphisms? It is likely to be low since new HLA variants have frequently been discovered because of unusual probe hybridisation (8) or primer amplification (9) patterns.

From the 5350 tests, a total of 186 A, B, Cw, DRB1/B3/B4/B5, DQB1 alleles were identified representing 13-38% of known alleles at a given locus (average 17.1%). Therefore >80% of all alleles assigned by the HLA Nomenclature Committee was not identified among this large group of patients and donors. It may be argued that this results from the fact that a very homogenous population of patients and donors was analysed. This is not the case since patients were of different geographic origins, including several Mediterranean and East European countries, although most of the patients were of Caucasoid origin. Furthermore the potential donors were recruited from several bone marrow registries in Europe and the US. Excluding patients for whom no potential A,B,DR-compatible (serology or generic DNA typing) donors are available in BMDW certainly contributed to decrease the occurrence of rare alleles. Similarly by having A,B,DR phenotypes that occurred more than once, donors from the BMDW Registry are expected to have a higher frequency of conserved haplotypes. For example, a higher frequency of A2-B44-DR4 haplotype in our patient/donor panel will increase the relative occurrence of A*0201, B*4402, Cw*0501, and DRB1*0401 alleles in the pooled data, and thereby decrease the relative occurrence of rare alleles within the A2, B44 or DR4 serotypes. Nonetheless our data suggest that alleles that have been assigned during the last couple of years should not be expected to have a major impact in the unrelated bone marrow donor search procedure. An increase in the number of non-Caucasoid patients and donors is likely to correlate with the occurrence of a larger number of alleles, although we think that still a majority of HLA alleles will not be found. This is supported by a recent high resolution PCR-SSOP locus -A analysis of about 700 individuals from 9 populations world wide (7). This study revealed the presence of a total of 48 different HLA-A alleles, i.e. 23.1% of currently known HLA-A alleles, a ratio slightly above that reported in this analysis. Reports from other laboratories will be informative, particularly from those using SBT for unrelated bone marrow donor matching.

Acknowledgements: We are grateful to M.-P. Labonne, I. Favre-Victoire, M. Bujan, B. Kervaire, P. Roux-Chabbey, and S. Stadelmann for their technical support.

References

1. Hurley et al. Bone Marrow Transpl. 2000; 25: 385-93.
2. Knipper et al. Hum. Immunol. 2000; 61: 605-14.
3. Darke et al. Tissue Antigens 2000; 56: 467-9.
4. Hsu et al. Hum. Immunol. 1999; 60. 159-67.
5. Gauchat-Feiss et al. Transplantation 1995; 60: 869-73.
6. Mickelson et al. Hum. Immunol. 2000; 61: 92-100.
7. Middleton et al. Hum. Immunol. 2000; 61: 1048-52.
8. Steiner et al. Tissue Antigens 2000; 56: 551-2.
9. Elsner & Blasczyk, Tissue Antigens 2000; 56: 371-5.

Table 1: List of HLA class I and II alleles identified in 5350 tests performed in 250 patients and 1200 unrelated donors from the BMDW Registry over a period of 22 months a), b).

a) The following class I and II alleles were found only once: A*0102, 2407, 2416, 6810, 3004, 3402, 8001, B*0703-0706, 1523, 1524, 1803, 2709, 2710, 3505, 3510, 3702, 5105, 5602, 5703, 5802, 7301, Cw*0302, 0707, 1801, DRB1*1108, 1114, 1127, 1304, 1416, 0810, and DRB5*0203.

b) This analysis did not consider rare variants from QC samples or from blood donors referred to our laboratories because of difficult antigen assignment.

Table 2: Number of HLA alleles detected in our patient/donor group compared to total number of HLA alleles defined at each locus.

© The European Federation for Immunogenetics
Page Created 10 July, 2002, Last Modified 1 August, 2006