Helper and cytotoxic T cell precursor frequencies are not ... - Nature

Bone Marrow Transplantation, (1998) 22, 1049–1055  1998 Stockton Press All rights reserved 0268–3369/98 $12.00 http://www.stockton-press.co.uk/bmt

Helper and cytotoxic T cell precursor frequencies are not predictive for development of acute graft-versus-host disease after partially T cell-depleted HLA-identical sibling BMT A van der Meer1, WA Allebes1, CEM Voorter2, EM van den Berg-Loonen2, AV Schattenberg3, TJ de Witte3 and I Joosten1 1

Blood Transfusion Service, and 3Division of Haematology, University Hospital Nijmegen, Nijmegen; and 2Tissue Typing Laboratory, University Hospital Maastricht, The Netherlands

Summary: Despite the use of partially T cell-depleted grafts, 20% of the recipients of an HLA-identical sibling marrow graft develop aGVHD ⭓II. This indicates that the current method for selecting a sibling donor, ie serological typing for HLA-A, B and DR, and a mixed lymphocyte culture (MLC) or molecular typing for HLADRB/DQB, is not predictive for aGVHD. In order to optimise our selection procedure, we retrospectively analysed patients who developed aGVHD ⭓II by means of sequencing based typing for HLA-DPB and frequency analysis of alloreactive helper and cytotoxic T lymphocyte precursors (HTLp-f and CTLp-f). Patients who did not develop aGVHD or developed aGVHD grade I served as controls. Retrospective typing for HLA-DPB revealed only a single disparity in the group with aGVHD ⭓II, indicating that mismatches for antigens other than HLA are the major cause of aGVHD in these patients. Furthermore, in our patient group, neither HTLp-f nor CTLp-f were predictive for development of aGVHD indicating that these assays in their current set-up are insufficiently sensitive to predict aGVHD in BMT with a partially T cell-depleted graft. We conclude, that HLA-identical siblings can be identified by means of serological typing for HLA-A and B and intermediate resolution molecular typing for DRB and DQB, but that for the prediction of aGVHD cellular tests with higher sensitivity and specificity as compared to the currently used HTLp-f and CTLp-f assays need to be developed. Keywords: bone marrow transplantation; HLA-identical sibling; cytotoxic T cell precursor frequency; helper T cell precursor frequency; HLA-DP

Alloreactive T cells present in the allogeneic marrow graft are the initiators of acute graft-versus-host disease (aGVHD) after allogeneic bone marrow transplantation (BMT). Matching for human leukocyte antigens (HLA) is Correspondence: Dr A van der Meer, Transfusion Service, University Hospital Nijmegen, PO Box 9101, 6500 HB Nijmegen, The Netherlands Received 9 May 1998; accepted 28 July 1998

therefore crucial to reduce the incidence of aGVHD. As the HLA gene cluster is inherited in a Mendelian fashion, HLA identity between patient and donor can be achieved by using sibling donors who have inherited the same HLA haplotypes. However, 20–40% of these patients still develop moderate to severe aGVHD.1 It would obviously be a great achievement if this risk could be identified before transplantation, in order to either search for a more histocompatible (sibling) donor or alternatively, to intensify GVHD prophylaxis. The current procedures used for the identification of HLA-identical siblings are largely based on serological typing of HLA-A, B and DR, whereby class II identity is confirmed either by a mixed lymphocyte culture (MLC) or by molecular typing for HLA-DRB and DQB. This selection procedure does not take into account the following two aspects: firstly, although patient and sibling donor have inherited the same HLA haplotypes, it has become apparent that due to recombination, HLA-DP mismatches can be present in 6–13% of the presumptive HLA-identical siblings.2–6 It has been shown that HLA-DP can function as a target antigen after BMT7,8 and severe aGVHD after BMT with HLA-DP mismatched sibling grafts has been reported.4,6 On the other hand, not all DP mismatches appear to be associated with the occurrence of aGVHD.3,9,10 As HLA-DP mismatches cannot always be detected in a primary MLC2,9 and HLA-DP typing is not routinely performed in a donor selection procedure, HLA-DP mismatches might remain unnoticed. Thus, so-called HLAidentical siblings may in fact be mismatched for at least one HLA locus. Secondly, in fully HLA-identical siblings, distinct sets of polymorphic peptides can be presented in the context of HLA molecules, thus causing aGVHD.11 Only a few of these so-called minor histocompatibility antigens (mH) have been characterised and typing is not yet routinely available. One might circumvent the problems of detecting antigenic mismatches by functional matching of patient and donor by virtue of frequency analysis of alloreactive donor T cells. It has already been shown that a high frequency of IL-2 producing alloreactive donor T cells (HTLp-f) is associated with an increased risk for aGVHD in non-T celldepleted BMT12–14 allthough this was not consistently found.15 In contrast, two limited studies revealed that the frequency of cytotoxic T lymphocyte precursors (CTLp-f)

Donor selection in HLA-identical sibling BMT A van der Meer et al

1050

appeared insufficiently sensitive for detection of alloreactivity in HLA-identical siblings.16,17 However, the number of patients studied thus far is limited, and all studies have been performed with siblings who received an unmanipulated bone marrow graft. In our centre, T cells are partially depleted from the marrow graft by counterflow centrifugation. However, despite T cell depletion, 20% of the patients receiving an HLAidentical sibling graft developed aGVHD ⭓ grade II.18 With the objective of improving the selection of sibling donors we retrospectively analysed our current matching procedures by comparing the outcome of the MLC in relation to DRB/DQB typing and the development of aGVHD. Subsequently, to determine whether HLA-DPB mismatches were responsible for the induction of aGVHD, HLA-DPB was analysed by a sequence-based typing method.19,20 Furthermore, HTLp-f and CTLp-f were determined to investigate whether the presence of alloreactive T cells before transplantation was predictive for development of aGVHD after BMT. Our data show that neither the functional assays (HTLp-f/CTLp-f/MLC) nor detailed HLA typing is predictive for the development of aGVHD in recipients of a partially T cell-depleted HLA-identical sibling graft. Material and methods Patients

Conditioning for transplantation Donor marrow was partially depleted of lymphocytes by counterflow centrifugation as published previously.22,23 0.7 ⫻ 106 to 1.0 ⫻ 106 T lymphocytes per kg body weight were left in the graft. The standard conditioning regimen consisted of cyclophosphamide (2 ⫻ 60 mg/kg body weight) and total body irradiation (2 ⫻ 4.5 Gy or 2 ⫻ 6 Gy). In 42 patients demethoxydaunorubicine (42 mg/m2) was added to the conditioning regimen. Post transplant, all patients received cyclosporin A. HLA-typing PBMC of patients and donors were typed by serology for HLA-A, B and DR by the standard microcytotoxicity assay. Molecular typing for DRB and DQB at intermediate resolution was performed by PCR SSOP as previously described.24–27 HLA-DPB1 sequence-based typing

In the period between January 1990 and May 1995, we performed 153 HLA-identical sibling transplants. Acute GVHD ⭓ grade II, as established by standard criteria,21 occurred in 38 patients. Due to lack of pre-transplant material, only 30 of these 38 recipient/donor combinations could be analysed in this retrospective study. Twenty-two patients had aGVHD grade II, five patients suffered from aGVHD grade III, and three patients had aGVHD grade IV. As a control group, 13 patients who did not develop aGVHD and 20 patients with aGVHD grade I were selected to match the group with aGVHD ⭓ grade II for a number of criteria as depicted in Table 1. Variables were compared by the Wilcoxon/Mann Whitney U test and did not differ significantly between the two groups. Indications for transplantation were: acute lymphoblastic leukaemia (n = 9), acute myeloid leukaemia (n = 21), chronic myeloid leukaemia (n = 12), multiple myeloma (n = 7), refractory anaemia (n = 6), non-Hodgkin’s lymphoma (n = 6) and severe aplastic anaemia (n = 2). All donor–recipient pairs fulfilled Table 1

our standard matching criteria for HLA-identical sibling donors, ie serological identity for HLA-A, B and DR and a negative MLC or serological identity for HLA-A, B and HLA identity at the molecular level for HLA-DRB and DQB. Haplotype segregation patterns could be identified in 24 donor/recipient pairs of the group with aGVHD ⭓ grade II and in 22 in the group with aGVHD group grade 0 or I.

Clinical characteristics of patient–donor combinations

Total No. of patients Median age (range) patients, in years Median age (range) donors, in years Sex-matched Female donor Median No. of T cells in graft

aGVHD grade 0–1

aGVHD grade II–IV

33 40 (17–56) 40 (18–62) 16 18 67 ⫻ 106

30 42 (18–57) 41 (20–50) 14 13 76 ⫻ 106

For sequence-based typing of DPB1 a PCR product is generated by amplification of exon 2 of the DPB1 gene. The primers used for amplification and sequencing are located in introns 1 and 2 (3⬘ primer: TGAATCCCCAACCCAAAGTCCCC, 5⬘ primer: AGGACCACAGAA-CTCGGTACTAGGA). For sequencing a solid phase approach was used as described previously using the Autoread Sequencing kit from Pharmacia Biotech (Uppsala, Sweden).19,20 Typing was performed using Pharmacia Typing Software. Mixed lymphocyte culture (MLC) MLC were performed according to ASHI standards.28 Briefly, 5 ⫻ 104 donor PBMC were incubated with 5 ⫻ 104 irradiated (30 Gy) patient PBMC and vice versa in 96 Ubottom well plates (Greiner, Frickenhausen, Germany) in RPMI-1640 with glutamax (Gibco, Paisley, UK) supplemented with pyruvate containing 100 U/ml penicillin, 100 ␮g/ml streptomycin (all from Gibco) and 10% heat inactivated pooled human serum (PHS). A two-way MLC was performed by mixing 5 ⫻ 104 donor cells with 5 ⫻ 104 patient cells. Cells were incubated for 5 days at 37°C in a humidified atmosphere containing 5% CO2 and then pulsed with 0.5 ␮Ci 3H-thymidine (Amersham, UK) for 18 h. Cells were harvested the following day. Single cell suspensions from two unrelated donors and pooled cells from four unrelated donors served as controls. Stimulation index (SI) was calculated as test c.p.m. divided by auto c.p.m. The MLC was qualified not interpretable in cases of insufficient stimulator or responder activity against the unrelated con-


trols (SI ⬍10) and/or in case of a high autologous response. Combinations were qualified MLC negative if SI in GVH and/or HVG direction did not exceed 2.5 and the proliferation in the two-way MLC did not exceed five times the mean of the autologous response of patient and donor PBMC. Combined helper T lymphocyte precursor frequency assay (HTLp-f) and cytotoxic T lymphocyte precursor frequency assay (CTLp-f) HTLp-f and CTLp-f were determined in the graft-versushost direction from a single limiting dilution assay (LDA) as published previously.29 In brief, a LDA was set up by culturing limiting numbers of responder cells (8 ⫻ 104– 1250 cells/well) with constant numbers (5 ⫻ 104) of irradiated (30 Gy) stimulator cells, 20 wells per dilution were set up. As a reference, irradiated stimulator cells alone were incubated in culture medium. As negative controls effector cells were incubated with irradiated ‘effector’ cells. Completely mismatched third-party cells were used either as effector or stimulator to determine stimulator capacity of the irradiated stimulator cells or the responsive capacity of the effector cells, respectively. After 3 days the plates were centrifuged (1200 g) and half of the supernatants was transferred to new plates for use in the HTLp assay. Fresh medium supplemented with rIL-2 (Eurocetus, Amsterdam, The Netherlands,) was added to the wells of the LDA-culture and this was repeated on day 6. On day 10 CTLp-f were determined by performing a ‘JAM’-test.29,30 For determining HTLp-f the proliferation of the IL-2-dependent cell line CTLL-231 was used as a read out for IL-2 secretion by alloreactive T cells. The day before use in the assay CTLL-2 cells were washed and incubated overnight in medium without rIL-2. On the day of the assay a dose– response curve was set up to determine the cut-off point for the HTLp-f assay. In our opinion the assay is only reliable when measuring in the linear part of the dose– response curve. Therefore, the cut-off value for the assay was set at the beginning of the linear part of the dose– response curve. In practice this was always between 0.01 and 0.04 IU/ml of IL-2. Statistical analysis Frequency estimations were calculated using the computer program developed by Strijbosch et al.32 Frequencies were estimated using the jackknife version of the maximum likelihood analysis. Data were excluded when the goodness of fit was higher than 12.5, corresponding to P values ⬎0.05. The two groups were compared by the Wilcoxon/Mann Whitney U test.

Table 2 Correlation between HLA-DRB and DQB identity and MLC reactivity in HLA-A and B identical siblings

MLC positive MLC negative Total

Total

3 146 149

3 0 3

6 146 152

be done by MLC or by molecular typing of HLA-DRB and DQB. As performing both methods simultaneously is laborious and very likely superfluous, we determined the correlation between DRB/DQB typing and the MLC in HLA-A, B-identical siblings identified during donor searches. The data (Table 2) show that in the 152 pairs for which both techniques were performed, a significant correlation was found between DRB/DQB matching and the outcome of the MLC (P ⬍ 0.0001, two-sided Fisher’s exact test). In only three pairs was there discrepancy, ie DRB/DQB identity associated with a positive MLC. In one case the MLC appeared negative when repeated on fresh blood. Notwithstanding the positive MLC, in two cases the patient was transplanted on the basis of HLA-A, B, DRB and DQB identity. Both patients developed only aGVHD grade I. Next, we analysed the predictive value of the two approaches for the development of aGVHD. The results show that neither the MLC nor HLA-DRB/DQB typing was predictive as still approximately 20% of the patients developed aGVHD (Table 3). Molecular typing of HLA-DPB1 Since we had no data on the capacity of our MLC to detect HLA-DP mismatches, we could not rule out their presence in presumptive HLA-identical siblings. Therefore, we retrospectively typed all our patients with aGVHD for HLADPB1 by sequence-based typing (SBT). A control group that did not develop aGVHD was selected to match the aGVHD group for a number of criteria as depicted in Table 1. Typing for HLA-DPB revealed only a single mismatched donor/recipient pair in the aGVHD group. The patient was typed as HLA-DPB1*02012/0401, while the donor was typed as HLA-DPB1*0401/0402. Since the segregation pattern of the HLA haplotypes within the patient’s family could be identified on the basis of HLA-A, B, DRB and DQB typing, the HLA-DPB mismatch is most likely due to a recombinational event. The patient suffered from aGVHD grade III and died 70 days post BMT from an infection. Correlation of aGVHD with MLC identity

Results

Our current strategy for identifying HLA-identical sibling donors for BMT is based on serological typing for HLAA, B and DR. Further confirmation of class II identity can

DRB/DQB non-identical

The two-sided P value was ⬍0.0001 (Fisher’s exact probability).

Table 3

Comparison of MLC with DRB/DQB typing

DRB/DQB identical

aGVHD 0–1 aGVHD II–IV Total

MLC negative

MLC positive

Total

47 14 61

1 0 1

48 14 62

Not significant (Fisher’s exact probability).

1051


Precursor frequencies of alloreactive donor T cells As the MLC appeared insufficiently sensitive for prediction of aGVHD, we evaluated the predictive value of the more sensitive limiting dilution analysis to determine the frequencies of alloreactive donor T cells. Pre-transplant HTLp-f were determined retrospectively in the graft vs host direction. Due to lack of pre-transplant material HTLp-f and CTLp-f could only be determined in part of the combinations. We observed that the HTLp-f were not significantly higher in patient/donor pairs developing aGVHD ⭓ grade II, as compared to patient/donor pairs who did not develop aGVHD or had aGVHD grade I (Figure 1a). Furthermore, the range of frequencies detected was relatively small (1/8 ⫻ 104 to 1/106) and frequencies were spread equally in the two groups. Therefore, no cut-off value could be established which could distinguish between low-risk and high-risk groups. Also for the CTLp-f no difference was observed between the study group and the control group (Figure 1b). This shows that next to the MLC, also neither the HTLp-f nor the CTLp-f assay in their current set-up are sufficiently sensitive to predict aGVHD ⭓II in genotypically HLA-identical siblings receiving a partially T cell-depleted graft. Discussion In the current study we set out to optimise our selection procedures for HLA-identical sibling donors, as 20% of these patients receiving a partially T cell-depleted bone marrow graft develop aGVHD.18 We found that neither HTLp-f nor CTLp-f nor the MLC were predictive for development of aGVHD ⭓ grade II and that all but one of the donor–recipient pairs were fully HLA identical, including identity for HLA-DP. Our current selection procedure is based on serotyping for HLA-A and B, whereby class II identity is confirmed either by MLC or by molecular typing for DRB and DQB. Regarding the MLC, our data clearly show a very strong correlation with HLA-DRB/DQB identity, such that at present MLC typing for the sake of establishing HLA class II identity is considered obsolete. However, neither the DRB/DQB typing nor the MLC is predictive for the development of aGVHD in HLA-identical siblings. This obviously indicates that the MLC is insufficiently sensitive to detect all functionally relevant mismatches.33 An additional disadvantage of the MLC is that it depends on viable and potent effector and stimulator PBMC, and this is often a problem in patients suffering from haematological disorders. Also, DeGast et al33 showed that MLC reactivity might be due to leukaemia-associated antigens rather than genetic differences between recipient and donor.

a 1000

10 000

HTLP-f (1/y)

Notably, in the in vitro functional histocompatibility assays no or only weak responses were detected (SIGVH = 0.5, HTLp-f ⬍1 ⫻ 106, CTLp-f ⬍1 ⫻ 106). In summary, these results indicate that in all but one of our patients receiving an HLA-identical sibling graft, aGVHD is caused by mismatches for antigens other than HLA.

100 000

1 000 000

Undetectable aGVHD 0-l

aGVHD ll-lV

b 1000

10 000

CTLP-f (1/y)

1052

100 000

1 000 000

Undetectable aGVHD 0-l

aGVHD ll-lV

Figure 1 (a) Host-reactive donor helper T lymphocyte precursor frequencies before transplantation from recipients with no aGVHD or grade I (n = 18) and recipients with aGVHD grade ⭓II (n = 13). (b) Hostreactive donor cytotoxic T lymphocyte precursor frequencies before transplantation from recipients with no aGVHD or grade I (n = 10) and recipients with aGVHD grade ⭓II (n = 9).

Although it has become apparent that HLA-DP, next to the other class II molecules, plays a significant role in the immune response,34 the role of HLA-DPB as a transplantation antigen is as yet much debated.2,35 With regard to presumptive HLA-identical sibling transplantation this controversy is sustained by the limited numbers of patients that have been studied so far.35 Also in our study with 30 cases of aGVHD ⭓ grade II, only a single recipient/donor


pair exhibited an HLA-DPB mismatch. This does not exclude the role of HLA-DPB as a transplantation antigen, but it clearly shows that HLA-DP disparities are not the major cause of aGVHD in our sibling group. The major amino acid difference in this particular combination is located at position 69 (either Lys or Glu) in the fourth hypervariable region of the HLA-DP molecule, a residue that is supposed to have a major influence on the nature of the peptides that can bind to the HLA-DP molecule.36 Disparities for this residue were associated with increased in vitro alloreactivity in unrelated combinations.36 Surprisingly, although our patient suffered from aGVHD grade III no alloreactivity was detectable before transplantation in any of the in vitro functional histocompatibility assays. This confirms that HLA-DPB disparities are not always detectable in a primary MLC.9,33,37 Since a negative MLC was a prerequisite for donor selection, one might argue that the low number of DP mismatches observed in our study is caused by the fact that in a number of cases HLA-DP mismatched sibling pairs might have been excluded during the selection procedure. In fact, during the period of study only three positive MLC were observed in identical sibling pairs. Two donor–recipient pairs were typed for HLA-DPB and no disparities were found, providing further evidence of the low frequency of DP disparities. Consequently, in our opinion the very low frequency of HLA-DP mismatches in HLA-identical siblings does not justify routine HLA-DP typing as part of the selection procedure for HLA-identical sibling donors. In our study, we did not find a correlation between the pre-transplant frequencies of alloreactive helper or cytotoxic donor T cells and development of aGVHD. Since all donors but one were fully HLA-identical siblings, it appears that both assays are insufficiently sensitive to discriminate between clinically relevant and less relevant minor antigen mismatches in genotypically HLA-identical siblings, which is in accordance with previous studies regarding HTLp-f15 and CTLp-f.17 The lack of sensitivity is already reflected in the low frequencies observed in both assays. While in ‘matched’ unrelated donor–recipient pairs the frequencies vary from 1:5 ⫻ 103 up to less than 1:1 ⫻ 106 (A van der Meer, manuscript in preparation), the highest frequency observed in this study was 1:8 ⫻ 104 in the HTLp-f assay and 1:6 ⫻ 104 in the CTLp-f assay while most frequencies were lower than 1:1 ⫻ 105. Our data are in contrast with studies performed by Schwarer et al12 and Weston et al14 who both found a correlation between HTLp-f and the development of aGVHD ⭓ grade II in HLA-identical sibling donor–recipient pairs. The frequencies we observed in our study are lower than those reported in the above-mentioned studies.12,14 As published previously29 this is at least partly a result of measuring IL2 production in the supernatants which were removed from the LDA cultures, instead of adding the CTLL-2 cells directly to the wells of the irradiated LDA cultures. This was necessary to obtain both HTLp-f and CTLp-f from a single LDA but results in a decrease of sensitivity. However, in a preliminary study using the protocols as described by Weston et al14 we were unable to confirm the high frequencies in patients developing aGVHD ⭓ grade III (unpublished results). In fact the HTLp-f did not rise above

1:5 ⫻ 105 in the donor–recipient pairs that were tested. Notably, in contrast to our study, in the previous two studies all patients received grafts that were not manipulated.12,14 Also for the CTLp-f no correlation was found with the occurrence of aGVHD. This confirms a limited study by Irschick et al,17 who were unable to detect CTLp-f before BMT in four patients developing aGVHD. Also a preliminary study by Kaminski et al16 indicated that the assay appeared insufficiently sensitive for detecting alloreactivity, with frequencies varying between 1:4 ⫻ 105 up to undetectable levels. Although the frequencies we detected were somewhat higher (up to 1:6 ⫻ 104), the assay is still insufficiently sensitive for the prediction of aGVHD. In our view, the inability of the functional histocompatibility assays as used in our study to detect cellular reactivity might be overcome by changes in the protocol. First, by using PBMC as stimulator and responder cells the system appears insufficiently sensitive for detecting minor antigen mismatches in a fully HLA-identical background. This problem might be overcome by using antigen presenting cells which exhibit a more profound stimulatory capacity. Second, tissue-specific antigens may play a role in the induction of the alloresponse as aGVHD is mainly restricted to certain organs. Obviously, responses against these minor antigens will not be detected when using PBMC as stimulator cells. The use of tissue-specific cells, ie epidermal cells, as target cells may improve the predictive value of a functional histocompatibility assay.38 Whether an increase in the sensitivity will also lead to a better correlation with aGVHD remains to be determined. An alternative approach for identifying a sibling donor with the highest degree of histocompatibility with the patient is by directly determining the presence or absence of minor antigen mismatches by molecular techniques.39 Obviously, this method is only successful once the minor antigens that play a role in the induction of aGVHD have been identified and characterised. Thus far, only one minor antigen (HA-1) has been characterised that was shown to be associated with the development of aGVHD in HLAidentical siblings.11 Furthermore, because the HA-1 antigen is presented in the context of HLA-A2 this approach is only applicable in a selected group of patients. Also, theoretical models show that only little benefit can be gained by typing and matching for a single minor antigen locus and that this approach in fact can only be successful when at least four minor antigens are typed for.40 In our opinion, this again supports the need for development of more sensitive functional assays to detect functionally relevant minor antigen mismatches. Based on the results of this study, we currently apply the following method for identifying HLA-identical siblings. First, serological typing for HLA-A and B of all available siblings is performed to identify haplotype segregation patterns. Subsequently all HLA-A, B identical siblings are typed at intermediate resolution for DRB/DQB by PCRELISA and PCR-SSP. When at this stage it is not possible to define haplotype segregation patterns, the parents, if still alive, are typed for HLA-A, B by serology and DRB/DQB by molecular typing techniques. In case several HLA-identical siblings are available, criteria such as donor age, gender

1053


1054

and virus serology are taken into account to select the most suitable donor. In addition, we continue to work on the sensitivity and the specificity of the functional histocompatibility assays as outlined above. Acknowledgements This work was financially supported by Bone Marrow Donorbank Nijmegen.

References 1 Bortin MM, Horowitz MM, Mrsic M et al. Progress in bone marrow transplantation for leukemia: a preliminary report from the Advisory Committee of the International Bone Marrow Transplant Registry. Transplant Proc 1991; 23: 61–62. 2 Howell WM, Sage DA, Kohler JA et al. Incidence and clinical significance of HLA-DP patient–donor mismatches in allogeneic bone marrow transplantation (letter). Bone Marrow Transplant 1995; 15: 816–817. 3 Pawelec G, Ehninger G, Schmidt H, Wernet P. HLA-DP matching and graft-versus-host disease in allogeneic bone marrow transplantation. Transplantation 1986; 42: 558–560. 4 Nomura N, Ota M, Kato S et al. Severe acute graft-versushost disease by HLA-DPB1 disparity in recombinant family of bone marrow transplantation between serologically HLAidentical siblings: an application of the polymerase chain reaction-restriction fragment length polymorphism method. Hum Immunol 1991; 32: 261–268. 5 Clay TM, Culpan D, Howell WM et al. UHG crossmatching. A comparison with PCR-SSO typing in the selection of HLADPB1-compatible bone marrow donors. Transplantation 1994; 58: 200–207. 6 Amar A, Nepom GT, Mickelson E et al. HLA-DP and HLADO genes in presumptive HLA-identical siblings: structural and functional identification of allelic variation. J Immunol 1987; 138: 1947–1953. 7 Gaschet J, Mahe B, Milpied N et al. Specificity of T cells invading the skin during acute graft-vs-host disease after semiallogeneic bone marrow transplantation. J Clin Invest 1993; 91: 12–20. 8 Gaschet J, Lim A, Liem L et al. Acute graft versus host disease due to T lymphocytes recognizing a single HLA-DPB1*0501 mismatch. J Clin Invest 1996; 98: 100–107. 9 al Daccak R, Loiseau P, Soulie A et al. HLA-DP genotyping in HLA-A,B, and DR identical intrafamilial bone marrow transplantation. Leukemia 1990; 4: 222–226. 10 Moreau P, Milpied N, Cesbron A et al. Mixed leukocyte culture reactivity, HLA-DP typing and GVHD (letter; comment). Bone Marrow Transplant 1993; 11: 85–86. 11 Goulmy E, Schipper R, Pool J et al. Mismatches of minor histocompatibility antigens between HLA-identical donors and recipients and the development of graft-versus-host disease after bone marrow transplantation. New Engl J Med 1996; 334: 281–285. 12 Schwarer AP, Jiang YZ, Brookes PA et al. Frequency of antirecipient alloreactive helper T-cell precursors in donor blood and graft-versus-host disease after HLA-identical sibling bone-marrow transplantation. Lancet 1993; 341: 203–205. 13 Theobald M, Nierle T, Bunjes D et al. Host-specific interleukin-2-secreting donor T-cell precursors as predictors of acute graft-versus-host disease in bone marrow transplantation between HLA-identical siblings. New Engl J Med 1992; 327: 1613–1617.

14 Weston LE, Geczy AF, Farrell C. Donor helper T-cell frequencies as predictors of acute graft-versus-host disease in bone marrow transplantation between HLA-identical siblings. Transplantation 1997; 64: 836–841. 15 Freidel AC, Michallet M, Gebuhrer L et al. Study of HTLp in adult patients receiving bone marrow transplantation from HLA geno-identical sibs. Hum Immunol 1996; 47: 84 (Abstr.). 16 Kaminski E, Hows J, Man S et al. Prediction of graft versus host disease by frequency analysis of cytotoxic T cells after unrelated donor bone marrow transplantation. Transplantation 1989; 48: 608–613. 17 Irschick EU, Hladik F, Niederwieser D et al. Studies on the mechanism of tolerance or graft-versus-host disease in allogeneic bone marrow recipients at the level of cytotoxic T-cell precursor frequencies. Blood 1992; 79: 1622–1628. 18 Schaap A, Schattenberg A, Bar B et al. Outcome of transplantation for standard-risk leukaemia with grafts depleted of lymphocytes after conditioning with an intensified regimen. Br J Haematol 1997; 98: 750–759. 19 Voorter CE, Rozemuller EH, de Bruyn Geraets D et al. Comparison of DRB sequence-based typing using different strategies. Tissue Antigens 1997; 49: 471–476. 20 Voorter CEM, de Bruym Geraets D, van den Berg-Loonen EM. High-resolution HLA typing for DRB3/4/5 genes by sequence-based typing. Tissue Antigens 1997; 50: 283–290. 21 Glucksberg H, Storb R, Fefer A et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation 1974; 18: 295–304. 22 Schattenberg A, De Witte T, Preijers F et al. Allogeneic bone marrow transplantation for leukemia with marrow grafts depleted of lymphocytes by counterflow centrifugation. Blood 1990; 75: 1356–1363. 23 De Witte T, Hoogenhout J, de Pauw B et al. Depletion of donor lymphocytes by counterflow centrifugation successfully prevents acute graft-versus-host disease in matched allogeneic marrow transplantation. Blood 1986; 67: 1302–1308. 24 Vaughan RW, Lanchbury JS, Marsh SG et al. The application of oligonucleotide probes to HLA class II typing of the DRB sub-region. Tissue Antigens 1990; 36: 149–155. 25 Wood WI, Gitschier J, Lasky LA, Lawn RM. Base composition-independent hybridization in tetramethylammonium chloride: a method for oligonucleotide screening of highly complex gene libraries. Proc Natl Acad Sci USA 1985; 82: 1585–1588. 26 Giphart MJ, Verduijn W. The Eurotransplant HLA-DRB oligonucleotide typing set. Eur J Immunogenet 1991; 18: 57–68. 27 Khalil I, d’Auriol L, Gobet M et al. A combination of HLADQ beta Asp57-negative and HLA DQ alpha Arg52 confers susceptibility to insulin-dependent diabetes mellitus. J Clin Invest 1990; 85: 1315–1319. 28 Mickelson EM, Guthrue LA, Hansen JA. The mixed lymphocyte culture (mlc) test. In: Phelan DL, Mickelson EM, Noreen HS et al (eds). ASHI Laboratory Manual, third edn. ASHI: Lenexa, USA, 1997, pp II.C.1.1–II.C.1.13. 29 Van Der Meer A, Joosten I, Ruiter J, Allebes WA. A single 3 H thymidine-based limiting dilution analysis to determine HTLP and CTLP frequencies for bone marrow donor selection. Bone Marrow Transplant 1997; 20: 149–155. 30 Matzinger P. The JAM test. A simple assay for DNA fragmentation and cell death. J Immunol Meth 1991; 145: 185–192. 31 Gillis S, Ferm MM, Ou W, Smith KA. T cell growth factor: parameters of production and a quantitative microassay for activity. J Immunol 1978; 120: 2027–2032. 32 Strijbosch LW, Buurman WA, Does RJ et al. Limiting dilution assays. Experimental design and statistical analysis. J Immunol Meth 1987; 97: 133–140.


33 DeGast GC, Mickelson EM, Beatty PG et al. Mixed leukocyte culture reactivity and graft-versus-host disease in HLA-identical marrow transplantation for leukemia. Bone Marrow Transplant 1992; 9: 87–90. 34 Meyer CG, May J, Schnittger L. HLA-DP – part of the concert. Immunol Today 1997; 18: 58–61. 35 Moreau P, Cesbron A. HLA-DP and allogeneic bone marrow transplantation. Bone Marrow Transplant 1994; 13: 675–681. 36 Potolicchio I, Brookes PA, Madrigal A et al. HLA-DPB1 mismatch at position 69 is associated with high helper T lymphocyte precursor frequencies in unrelated bone marrow transplant pairs. Transplantation 1996; 62: 1347–1352. 37 Cesbron A, Moreau P, Cheneau ML et al. Crucial role of the third and fourth hypervariable regions of HLA-DPB1 allelic

sequences in primary mixed-lymphocyte reaction: application in allogeneic bone marrow transplantation. Transplant Proc 1993; 25: 1232–1233. 38 Bagot M, Mary JY, Heslan M et al. The mixed epidermal cell lymphocyte-reaction is the most predictive factor of acute graft-versus-host disease in bone marrow graft recipients. Br J Haematol 1988; 70: 403–409. 39 den Haan JM, Meadows LM, Wang W et al. The minor histocompatibility antigen HA-1: a diallelic gene with a single amino acid polymorphism. Science 1998; 279: 1054–1057. 40 Martin PJ. How much benefit can be expected from matching for minor antigens in allogeneic marrow transplantation? Bone Marrow Transplant 1997; 20: 97–100.

1055