Biology of Blood and Marrow Transplantation
Volume 12, Issue 1, Supplement 1 , Pages 28-33, January 2006

Advances in HLA: Practical Implications for Selecting Adult Donors and Cord Blood Units

Georgetown University Medical Center, Blood and Marrow Transplant Program of the University of Minnesota, and National Marrow Donor Program, Minneapolis, Minnesota

Article Outline

 

Back to Article Outline

Introduction 

HLA plays a critical role in hematopoietic cell transplantation, but a full understanding of this role is constantly in flux because of changing technologies and burgeoning amounts of clinical transplantation results. Today the clinician considering unrelated donor transplantation must also consider stem cell source: an adult donor or an umbilical cord blood unit (CBU). In the context of cord blood, the task is further complicated by the strong interaction of cell dose.

Each individual carries from 10 to 12 genes that encode the HLA-A, HLA-B, HLA-C, HLA-DR, HLA-DQ, and HLA-DP molecules. Most of these genes are highly polymorphic, ranging from 13 (HLA-DRB4) to 699 (HLA-B) alleles at a locus [1]. Most of the allelic differences result in changes in the HLA protein sequences, thus producing potential targets for allorecognition [2]. The frequency of alleles differs in different populations: some are common in most populations, and others are found predominantly in only 1 or a few populations. The same is true for HLA haplotypes—the collection of multiple HLA alleles within a short region of chromosome 6—which are typically inherited as a single block of genetic information. Table 1 provides examples.

Table 1. Examples of Differences in Allele and Haplotype Frequencies in Different Populations
Population GroupAverage Frequency
B⁎4601 allele frequency [26]
Black5 in 10 000 black individuals
White4 in 10 000 white individuals
Asian567 in 10 000 Asian individuals
Native American0 in 10 000 Native Americans
Haplotype frequency⁎
A⁎0101, B⁎0801, DRB1⁎03011 in 15 whites, 1 in 75 blacks
A⁎3001, B⁎4201, DRB1⁎03021 in 48 blacks, almost never in whites

⁎ Unpublished data.

DNA-based testing methods provide an accurate and powerful approach to identify HLA diversity, and these are replacing serologic testing [3, 4]. DNA testing detects specific nucleotide differences that distinguish alleles or sets of alleles. The ability to distinguish among particular alleles at a locus (ie, the resolution) depends on the methodology, the set of oligonucleotide reagents used in the assay, or both. The choice of assay depends on the purpose of the typing, eg, initial testing of newly recruited volunteers or CBUs for a registry (low to intermediate resolution) compared with testing of transplantation candidates and a few selected potential donors or CBUs (high or allele-level resolution).

Back to Article Outline

HLA implications for selecting adult unrelated donors 

Since 1987, the National Marrow Donor Program (NMDP) has facilitated hematopoietic cell transplantations from unrelated donors. Until 1999, almost all of these transplantations used donor bone marrow, whereas today more than half use mobilized peripheral blood stem cells (PBSCs). Most recently, more than 45000 CBUs have also been added to the NMDP donor registry, and another approximately 140000 CBUs are available worldwide [5].

In 1995, the NMDP initiated a project to identify all HLA alleles carried by transplant recipients and their donors and to associate the level of matching with outcome [6]. The data generated from this ongoing study, combined with those from independent investigations, have provided considerable insight into the role of HLA in unrelated donor bone marrow transplantation. NMDP investigators authored a set of recommendations for selection of unrelated adult donors in 2004 [7]. The recommendations focused on survival, as opposed to other outcomes such as engraftment and graft-versus-host disease (GVHD), and included considerations from a large study in Japan and from studies of the Fred Hutchinson Cancer Research Center (FHCRC) group [8, 9, 10, 11, 12]. Although these 3 sets of studies differed in their designs and final conclusions, they suggested that matching for HLA-C, in addition to HLA-A, -B, and -DRB1, is important for overall survival (Table 2). In addition, the studies suggested that matching at the allele (or high-resolution) level conferred additional benefit over matching at an antigen level (ie, information available through serologic typing or low- or intermediate-level molecular typing). The NMDP study further concluded that mismatches between alleles (eg, 0401 versus 0404) were less detrimental than mismatches between antigens (eg, 0401 versus 1301).

Table 2. Effect of HLA Mismatching on Survival
Mismatched HLA Locus
StudyABCDRB1DQ
NMDP [6]DecreasedDecreasedDecreasedDecreasedNone
FHCRC [10]Merged A, B, and C decreasedMerged DRB1 and DQ decreased
JMDP [8]DecreasedDecreasedNoneMerged DRB1 and DQ none

JMDP indicates Japan Marrow Donor Program.

These results led the NMDP Histocompatibility Committee to recommend allele-level typing of the potential recipient at HLA-A, -B, -C, and -DRB1 for the purposes of searching and matching [7]. An ideal donor would clearly be one whose alleles match the recipient at HLA-A, -B, -C, and -DRB1. The NMDP data failed to show any additional benefit for matching at HLA-DQ or HLA-DP, although HLA-DQ linkages are valuable in selecting partially typed donors for more detailed evaluation. In the absence of an 8/8 allele–matched donor, the NMDP recommended minimizing the number of allele-level mismatches in total without preferentially weighing any particular locus, eg, HLA-DRB1 as more important than HLA-A, -B, or -C. If antigen mismatching is unavoidable, then the number of additional allele-level mismatches should still be minimized. What remained unknown, however, was how to balance allele mismatches against antigen mismatches (ie, what number of allele mismatches equals an antigen mismatch).

It is important to note that the effect of a single-allele mismatch (HLA-A, -B, -C, -DRB1, -DQB1, or -DPB1) may vary with the underlying diagnosis. In a recent publication on 948 donor-recipient pairs at the FHCRC, it was found that a single-allele mismatch conferred a higher risk of death, but only for low-risk patients, defined as CML within 2 years of diagnosis [13]. In contrast, a single-allele mismatch had no effect on survival among higher-risk patients, ie, those with more advanced CML, acute leukemia, or myelodysplastic syndrome.

Notably, in contrast to the NMDP study, the FHCRC study found that among the low-risk patients, a single-allele mismatch had the same negative effect as a single-antigen mismatch [13]. Furthermore, their results suggested that when multiple mismatches were unavoidable, HLA-DQ mismatches were more detrimental. Although the effect of HLA-DQ in the setting of multiple mismatches remains to be confirmed, these results clearly illustrate the complexity of the HLA matching.

A few points are worth emphasizing with respect to HLA and adult donor selection. First, essentially all the available information pertains to bone marrow transplantation. It is unknown whether the results will pertain to recipients of PBSC grafts. Because PBSCs differ from bone marrow in the numbers of CD34+ cells and lymphocyte content, which may account for differences in clinical “behavior” (eg, more rapid neutrophil engraftment and altered patterns of GVHD), it is possible that lessons learned for marrow may not apply to PBSCs. Second, the influence of reduced-intensity conditioning regimens on the principles of HLA matching is also largely unknown. Finally, the potential role of killer inhibitory receptors and their ligands in relation to unrelated donor selection is unknown. When the use of T cell–replete grafts from adult unrelated donors is considered, there are no data to justify selection of donors who are HLA-C mismatched outside the context of a research study.

Another difficult decision for physicians is whether to wait for a better matched donor or CBU to be recruited when no suitable donor or cord currently exists for a patient. Whether waiting is a reasonable strategy (assuming, of course, that the patient’s condition is stable) depends on the patient’s HLA type. If a white patient with slightly less common haplotypes has no matched donor, there is a modest likelihood that 1 of the 20000 new donors recruited each month may be the match. However, if the patient is of multiple race heritages (eg, Asian and Caribbean black) or has rare alleles, the probability of ever finding a perfectly matched unrelated donor is negligible. Waiting for a better donor is always a matter of balancing the risks involved, but understanding the patient’s haplotypes and having access to detailed allele and haplotype frequency tables by race could provide some direction on a case-by-case basis. The enhancements to the NMDP matching algorithm described below will be especially helpful in providing probability estimates that are useful for making these types of decisions.

Back to Article Outline

HLA implications for selecting CBUs 

Far less information is available to clarify the precise role of HLA in the setting of umbilical cord blood (UCB) transplantation. Overall, there have been fewer UCB transplantations (approximately 6000 worldwide), most have been children with acute leukemia and nonmalignant disease, data on HLA matching are less complete in terms of the availability of high-resolution typing, and the effect of HLA is confounded by the interaction of nucleated cell dose. Most reports have evaluated the effect of mismatching at the intermediate-resolution antigen level for HLA-A and -B and the allele level for -DRB1. HLA-C data and allele-level class I data are rarely available. Nevertheless, Rubinstein et al. [14], in their 1998 report on 562 CBU transplantations, observed that better matching at HLA-A and -B (antigen level) or -DRB1 (high resolution) predicted improved survival. In this milestone study, recipients with 0 or 1 HLA mismatches had a lower risk of nonrelapse events (death, autologous recovery, or retransplantation) than those with ≥2 mismatches.

Similarly, Wagner et al. [15] observed an independent effect of HLA match and cell dose on 2-year survival in 102 UCB recipients who underwent transplantation at a single institution. In contrast, Gluckman et al. [16], who analyzed Eurocord data on 550 UCB recipients who underwent transplantation for hematologic malignancies, reported that increasing degrees of HLA disparity had no effect on transplant-related mortality or overall survival. Notably, in this study, greater HLA disparity was associated with reduced neutrophil and platelet engraftment, an increased risk of acute GVHD, and a lower risk of relapse.

Two recent publications reported on adults who underwent UCB transplantation [17, 18]. In both studies, HLA matching was again reported at the serologic level for HLA-A and -B and at the allele level for HLA-DRB1. In the European Group for Blood and Marrow Transplantation/Eurocord study, all 584 bone marrow recipients were HLA matched, and 92 (94%) of 98 UCB recipients were mismatched [17]. Despite this difference, transplant-related mortality, relapse, and overall survival were comparable between the 2 groups. In the US study, 450 adults received bone marrow, and 150 received UCB (77% were mismatched at 2 HLA loci) [18]. In contrast to the Eurocord study, survival after UCB transplantations was inferior to that after marrow transplantation from an HLA-matched adult donor and was comparable to that after a 1 antigen/allele–mismatched adult donor. Taken together, both these studies suggest that with respect to survival, UCB is less HLA restricted than adult bone marrow.

Most recently, the Institute of Medicine commissioned a study of UCB outcomes that pooled data contributed by the New York Blood Center, the NMDP, and the National Heart, Lung and Blood Institute Cord Blood Transplantation study [5]. A total of 755 cases representing transplantations of children and adults in the United States were included. Matching for HLA-A and -B was evaluated at the low to intermediate level, and HLA-DRB1 was evaluated at the allele level. The Institute of Medicine study showed a clear effect of HLA matching on survival, with 6/6 matches faring better than 5/6 or 4/6 matches. The magnitude of this effect, however, was most apparent at lower cell doses (<2.5 × 107 total nucleated cells per kilogram). The Institute of Medicine report is the first major study to emphasize the effect of cell dose relative to HLA match with substantial numbers of patients. The conclusion is that HLA mismatch cannot be evaluated without concomitant consideration of cell dose. The data suggest that the greater the HLA disparity (0 versus 1 versus 2), the greater the effect of cell dose on survival. A higher cell dose can compensate, at least in part, for the negative effect of HLA disparity.

At this point, the precise role of HLA matching in the setting of UCB transplantation is less clear than that with unrelated bone marrow transplantation. Although some studies have suggested that better matching at the antigen level improves survival or reduces transplant-related mortality, others have not made that association. Cell dose and differences in patient populations and treatments confound these analyses. Still, the overall sense is that when considering intermediate-level HLA-A and -B typing and allele-level HLA-DRB1, a 5/6 or 6/6 match is preferred over a 4/6 match. Higher cell doses (>5 × 107 total nucleated cells per kilogram), however, may overcome the negative effect of HLA mismatches [5].

Finally, few data are available to address the allele-level influences of HLA in UCB transplantation. The US COBLT study recently reported on 32 young children (<4 years old) who received transplants for treatment of various leukemias [19]. In this small cohort, allele-level HLA data revealed an association between matching and survival that was not evident in the analysis of lower-resolution typing data.

Back to Article Outline

Improving the value of HLA data: strategies for increasing the available information 

The HLA types of adult volunteers and CBUs are listed within a registry database. Although increasingly transplant centers use HLA allele-level matching in their search for adult donors and CBUs, the specific alleles carried by each potential donor/CBU appearing on a search report usually are not known. For the purposes of creating a searchable registry, donor/CBU HLA types are almost always defined at low to intermediate resolution. This is because the HLA typing strategy must balance the resolution of HLA typing, which costs increasingly more with increasing resolution, against the desire to increase the total numbers of available donors and CBUs. New strategies, however, make it feasible to glean additional information from even low- and intermediate-resolution typing data. The techniques for selecting donors and CBUs for testing with the greatest likelihood of a high-resolution match are outlined in the remaining sections.

The Probability of Finding HLA-Identical Donors in the Registry 

Within the NMDP adult donor registry, the probability of identifying at least 1 antigen level–matched (HLA-A and -B) and potential allele level–matched (HLA-DRB1) matched donor varies within different racial and ethnic populations. Probabilities for the various populations are as follows: white, 88%; black, 60%; Hispanics, 81%; Asian/Pacific Islanders, 78%; and American Indians/Alaska Natives, 82% [20]. Without consideration of cell dose, the probability of finding at least 1 antigen-level HLA-A and -B and potential allele -DRB1 CBU that is a 5/6, or better, match within the NMDP Cord Registry is 80% for whites, 49% for blacks, 70% for Hispanics, 65% for Asian/Pacific Islanders, and 80% for American Indians/Alaska Natives. Once typed at a higher resolution, how many of these potential donors and CBUs will be allele matches for the patient? The probability of finding an allele match for specific patients varies dramatically. Patients with common haplotypes will find many allele-matched donors. This is because a donor or CBU that shares 6 antigens with 1 of these patients has a very high likelihood that the antigen-encoding genes are arrayed along the same common haplotypes possessed by the patient. The alternative result, namely, that the donor/CBU genes are arrayed along less common haplotypes that just happen to give the same 6 antigens, is statistically (and often dramatically) less likely. So when 1 of these donors/CBUs is typed at allele-level resolution, all of the alleles match, even those (eg, HLA-C) that were never previously typed. In contrast, some patients have fairly common HLA alleles but uncommon haplotypes. In these cases, it can be extremely difficult to find matched donors. For other patients with rare alleles, matching is very difficult unless the allele is found in a conserved haplotype (unpublished data).

The HLA diversity within a registry and the likelihood of matched donors can be evaluated with mathematical algorithms [21, 22]. These techniques make it possible to estimate the probability that certain low- or intermediate-resolution HLA typings represent specific known alleles. In other words, by generating tables of allele-level, multiple-locus haplotype frequency data, the logic can predict the HLA-A, -B, and -DRB1 alleles without performing the high-resolution typing.

The NMDP Search Algorithm in 2005 

Within the NMDP registry, HLA assignments of patients and donors (adults and CBUs) are converted to a common nomenclature called search determinants to allow comparison between the HLA types [23]. Once potential matches are identified, the algorithm converts the patient and donor search determinants back into their original nomenclature and prioritizes matching relationships on the search report. Donors/CBUs with previously defined HLA alleles identical to the patient’s alleles appear first on the search report, followed by donors/CBUs with lower-resolution DNA assignment matches, followed by donors with serologic typing (no CBUs are typed by serology methods) that indicates a potential match, followed by mismatched donors and CBUs. On the basis of this search report, the transplant center, working with HLA experts, must predict which donors/CBUs might ultimately be allele matched once typed at a higher resolution. If there are many potential matches, only a few can be tested because of constraints in time and patient resources. Choosing the most likely matched pairs requires insight into the reactivities of the HLA typing reagents used to perform the initial typing and knowledge of HLA population genetics. The NMDP is refining its search algorithm to use this information.

HLA DNA Data Behind the Scenes 

To provide information on the reactivities of the HLA typing reagents for each adult donor, the NMDP has developed methods to collect HLA nucleotide polymorphisms detected as present or absent in most new recruits [24]. This allows the registry to convert the HLA typing of a potential donor to a list of alternative genotypes, thus conserving valuable information about the allele pairs that can physically exist in the donor. This is information that is normally lost in the translation between the HLA typing laboratory and the reporting of the result to the registry. The nucleotide polymorphisms can be used later to determine whether a donor might carry a newly described allele. This is often essential information required to facilitate searches for patients who have many potential donors but for whom the probability of those donors carrying the patient’s alleles is very low.

Incorporating Population Genetics into the Algorithm 

One limitation in applying population genetics to donor and CBU searching is the relatively small number of individuals or CBUs typed for HLA at high resolution. The problem is amplified within racial and ethnic subgroups, in which even fewer data are available. To address this issue, biostatisticians and geneticists working with NMDP have developed a more robust algorithm to predict allele-level haplotypes (unpublished data). The new algorithm will allow the use of the varying-resolution HLA data in the 5.5 million–donor NMDP registry to predict allele and allele-level haplotype frequencies by race and ethnicity. The registry has also used the high-resolution data obtained in its donor-recipient pair typing project to add to the haplotype data required for the improved algorithm [25]. The composite probability predictions can be applied to all donors, including CBUs. For every patient’s search, the potential for each donor or CBU to be an allele-level match with the patient can be calculated. The integration of the nucleotide polymorphism data will ultimately increase the power to predict whether a newly described allele could be present but was undetected in an older HLA typing. With this new logic, donors and CBUs will be sorted on search reports according to their calculated overall probability for allele matching with the patient at 6/6, 5/6, or 4/6 (for CBU) for HLA-A, -B, and -DRB1.

A Step Forward to Speed the Search for a Matched Donor 

More than half (58%) of transplantations today are for acutely ill patients for whom search time is limited. Clearly, one of the profound advantages of UCB as a source of stem cells is its immediate availability. In some cases, a more mismatched UCB transplantation (eg, 4/6) will be performed even if the cell dose is borderline because of the length of time it takes to identify and clear a volunteer adult donor even when he or she is HLA matched. To assist these patients and expedite the search process, the NMDP is building a more powerful search algorithm. The new refinement will be especially helpful when physicians are faced with long lists of potential adult or CBU donors but have resources and time to evaluate only a few from that list. The goal is to provide more information on the search report to indicate the likelihood of each potential donor carrying the same alleles as the patient to streamline the donor/CBU selection process and reduce the chance of overlooking an optimal match.

Back to Article Outline

References 

  1. Marsh SG , Albert ED , Bodmer WF , et al.   Nomenclature for factors of the HLA system, 2004 . Tissue Antigens . 2005;65:301–369
  2. Whitelegg A , Barber LD . The structural basis of T-cell allorecognition . Tissue Antigens . 2004;63:101–108
  3. Noreen HJ , Yu N , Setterholm M , et al.   Validation of DNA-based HLA-A and HLA-B testing of volunteers for a bone marrow registry through parallel testing with serology . Tissue Antigens . 2001;57:221–229
  4. Setterholm M , Maiers M , Coleman L , et al.   HLA typing of volunteers for a hematopoietic stem cell donor registry (ensuring and maintaining quality of the NMDP registry) . ASHI Q . 2005;29:50–52
  5. In:  Meyer EA ,  Hanna K ,  Gebbie K editor. Cord Blood (Establishing a National Hematopoietic Stem Cell Bank Program) . Washington, DC: The National Academies Press; 2005;
  6. Flomenberg N , Baxter-Lowe LA , Confer D , et al.   Impact of HLA class I and class II high-resolution matching on outcomes of unrelated donor bone marrow transplantation (HLA-C mismatching is associated with a strong adverse effect on transplantation outcome) . Blood . 2004;104:1923–1930
  7. Hurley CK , Baxter Lowe LA , Logan B , et al.   National Marrow Donor Program HLA-matching guidelines for unrelated marrow transplants . Biol Blood Marrow Transplant . 2003;9:610–615
  8. Morishima Y , Sasazuki T , Inoki H , et al.   The clinical significance of human leukocyte antigen (HLA) allele compatibility in patients receiving a marrow transplant from serologically HLA-A, HLA-B, and HLA-DR matched unrelated donor . Blood . 2002;99:4200–4206
  9. Sasazuki T , Juji T , Morishima Y , et al.   Effect of matching class I HLA alleles on clinical outcome after transplantation of hematopoietic stem cells from an unrelated donor . N Engl J Med. . 1998;339:1177–1185
  10. Petersdorf EW , Gooley TA , Anasetti C , et al.   Optimizing outcome after unrelated marrow transplantation by comprehensive matching of HLA class I and II alleles in the donor and recipient . Blood . 1998;92:3515–3520
  11. Petersdorf EW , Gooley T , Malkki M , et al.   The biological significance of HLA-DP gene variation in haematopoietic cell transplantation . Br J Haematol. . 2001;112:988–994
  12. Petersodorf EW , Hansen JA , Martin PJ , et al.   Major-histocompatibility-complex class I alleles and antigens in hematopoietic cell transplantation . N Engl J Med. . 2001;345:1794–1800
  13. Petersdorf EW , Anasetti C , Martin PJ , et al.   Limits of HLA mismatching in unrelated hematopoietic cell transplantation . Blood . 2004;104:2976–2980
  14. Rubinstein P , Carrier C , Scaradavou A , et al.   Outcomes among 562 recipients of placental-blood transplants from unrelated donors . N Engl J Med. . 1998;339:1565–1577
  15. Wagner JE , Barker JN , DeFor T , et al.   Transplantation of unrelated donor umbilical cord blood in 102 patients with malignant and non-malignant diseases (influence of CD34 cell dose and HLA disparity on treatment-related mortality and survival) . Blood . 2002;100:1611–1618
  16. Gluckman E , Rocha V , Arcese W , et al.   Factors associated with outcomes of unrelated cord blood transplant (guidelines for donor choice) . Exp Hematol. . 2004;32:397–407
  17. Rocha V , Labopin M , Sanz G , et al.   Transplants of umbilical-cord blood or bone marrow from unrelated donors in adults with leukemia . N Engl J Med. . 2004;351:2276–2285
  18. Laughlin MJ , Eapen M , Rubinstein P , et al.   Outcomes after transplantation of cord blood or bone marrow from unrelated donors in adults with leukemia . N Engl J Med. . 2004;351:2265–2275
  19. Wall DA , Carter SL , Kernan NA , et al.   Busulfan/melphalan/antithymocyte globulin followed by unrelated donor cord blood transplantation for treatment of infant leukemia and leukemia in young children (the Cord Blood Transplantation Study (COBLT) experience) . Biol Blood Marrow Transplant. . 2005;11:637–646
  20. Hurley CK , Fernandez VM , Setterholm M . Maximizing optimal hematopoietic stem cell donor selection from registries of unrelated adult volunteers . Tissue Antigens . 2003;61:415–424
  21. Muller CR , Ehninger G , Goldman SF . Gene and haplotype frequencies for the loci HLA-A, HLA-B, and HLA-DR based on over 13,000 German blood donors . Hum Immunol. . 2003;64:137–151
  22. Beatty PG , Mori M , Milford E . Impact of racial genetic polymorphism on the probability of finding an HLA-matched donor . Transplantation . 1995;60:778–783
  23. Hurley CK , Setterholm M , Lau M , et al.   Hematopoietic stem cell donor registry strategies for assigning search determinants and matching relationships . Bone Marrow Transplant . 2004;33:443–450
  24. Maiers M , Hurley CK , Perlee L , et al.   Maintaining updated DNA-based HLA assignments in the National Marrow Donor Program Bone Marrow Registry . Rev Immunogenet. . 2000;2:449–460
  25. Klitz W , Maiers M , Spellman S , et al.   New HLA haplotype frequency reference standards (high-resolution and large sample typing of HLA DR-DQ haplotypes in a sample of European Americans) . Tissue Antigens . 2003;62:296–307
  26. Marsh SGE , Parham P , Barber LD . The HLA Factsbook . San Diego, CA: Academic; 2000;

PII: S1083-8791(05)00689-0

doi:10.1016/j.bbmt.2005.10.005

Biology of Blood and Marrow Transplantation
Volume 12, Issue 1, Supplement 1 , Pages 28-33, January 2006

Article Tools