* Abbreviations: Ag, antigen; BMT, bone marrow transplantation; CB, cord blood; CBMC, cord blood mononuclear cell; CM, complete medium; CM-ABs, complete medium supplemented with 10% heat-inactivated human AB serum; CM-As, complete medium supplemented with 10% heat-inactivated autologous serum; CTLp, cytotoxic T lymphocyte precursor; GVHD, graft-versus-host disease; GVHR, graft-versus-host reaction; HLA, human leukocyte antigen; HTLp, helper T lymphocyte precursor; IL-2, interleukin-2; LDA, limiting dilution analysis; MLC, mixed lymphocyte culture; PBMC, peripheral blood mononuclear cell.
Human umbilical cord blood (CB*) has been demonstrated to be a rich source of hematopoietic progenitor cells and could substitute for bone marrow in allogeneic stem cell transplantation (1). During recent years, CB has been successfully used for the hematopoietic reconstitution of children with malignant and nonmalignant diseases. In these pediatric recipients, the incidence of acute graft-versus-host disease (GVHD) is extremely low, being 3% in HLA-identical or one antigen (Ag)-mismatched sibling transplants (2) and 10% in partially (up to 3 Ags) HLA-mismatched unrelated transplants (3).
Although these clinical findings suggest that CB transplantation may require less stringent HLA matching and result in less severe GVHD, the immunological mechanisms underlying this phenomenon are not clear. To address this point, efforts have been made to examine the immune function of CB cells. Unfortunately, conflicting results have been observed in many aspects (4), and the immunological advantages of CB transplantation have yet to be proven (5). In an attempt to explore the possible mechanisms of a low GVHD incidence in CB transplants, this study has investigated the alloreactive capacity of CB cells in vitro using both limiting dilution analysis (LDA) and a human skin explant model.
PATIENTS AND METHODS
Informed consent and ethical approval. Written informed consent was obtained from the mother by nurse counselors for CB collection at a prenatal visit. Local ethical committee approval was obtained for all aspects of this study. Written consent for blood and skin biopsy samples was obtained from autotransplant patients.
HLA typing. The HLA typing was carried out at the National Blood Service, Newcastle, United Kingdom. Conventional serological typing was used for identifying HLA-A and -B Ags. HLA-DR typing was performed using polymerase chain reaction amplification with sequence-specific primers.
Culture medium. The complete culture medium (CM) used for mixed lymphocyte culture (MLC) and limiting dilution MLC was RPMI 1640 medium (GIBCO, Uxbridge, UK) containing 100 IU/ml penicillin (GIBCO), 100 µg/ml streptomycin (GIBCO), and 2 mM L-glutamine (GIBCO) supplemented with 10% (v/v) heat-inactivated pooled human AB serum (CM-ABs). The complete culture medium supplemented with 10% heat-inactivated autologous serum to the skin biopsy (CM-As) was used for responder cells/skin biopsy co-culture.
Preparation of cord and peripheral blood mononuclear cells. Peripheral blood was obtained from parental donors or autotransplant patients in remission. Skin biopsies were taken from the same autotransplant patient at the time of transplantation and used in skin explant assays. CB collection was performed in the Department of Obstetrics and Gynaecology at the Royal Victoria Infirmary, New-castle upon Tyne by midwives during the third stage of labor, with the placenta in utero (6). The cord and peripheral blood mononuclear cells (CBMCs and PBMCs) were separated over Lymphoprep (Nycomed Pharma AS, Oslo, Norway) by density gradient centrifugation at 800×g, at 20°C for 15 min (Jouan Centrifuge CR422). Collected mononuclear cells were washed twice in Earle's balanced salt solution (GIBCO) and resuspended in CM-ABs. All CBMCs and PBMCs were greater than 90% viable.
MLC. The MLC was carried out using CBMCs or parental HLA-haploidentical PBMCs as paired responder cells against the same stimulator (autotransplant patient) PBMCs. Responder cells (1×107) were cultured with the same number of irradiated stimulator cells (30 Gy, Cs137 source, Gammacell 1000 Elite, Nordion International Inc., Canada) in 25-cm2 flasks (Nunclon) in 20% CM-ABs at a total volume of 10 ml. After 7 days of incubation at 37°C in a humidified 5% CO2 in 95% air incubator (Flow Laboratories IR 1500), the MLC-primed responder cells were harvested, washed, and then cultured with skin explants taken from an autotransplant patient (the original stimulator).
Skin explant model for the in vitro detection of GVH type alloreactivity. The skin explant model was carried out as described previously (7,8). In brief, standard 4-mm punch biopsy specimens were obtained before transplant from autologous bone marrow transplant (BMT) patients. Under sterile conditions, the skin biopsies were trimmed of excess dermis and divided into 6-10 sections of equal size. Each section was cultured separately with either MLC-primed responder cells (CBMCs or PBMCs) or autologous lymphocytes (1×106 cell/well) or culture medium alone in 96-well round-bottomed microtiter plates (Nunc, Roskilde, Denmark) in CM-As. After 3 days of co-culture, skin explants were fixed in 10% buffered formalin, sectioned, and stained with hematoxylin and eosin. The histopathological evaluation of the skin explants was performed blindly and independently by a histopathologist (L.S.). The grading system (grades I-IV) used was based on that described by Lerner et al. (9). All medium control or autologous cell/autologous MLC control wells gave rise to grade I histopathological damage in skin biopsies, and this was regarded as background for all test biopsies. All biopsies of grade II or above were regarded as positive. In studies where the skin explant model is used to predict GVHD in HLA-identical sibling transplants, positive skin GVH type alloresponses have been shown to be indicative of GVHD grades II or above after transplant (10).
Alloreactive HTLp and CTLp frequency estimation. Alloreactive HTLp and CTLp frequencies were determined using a previously reported combined LDA assay (11). In all experiments, CBMCs and HLA-haploidentical parental PBMCs were used as paired responder cells. In each experiment, paired CBMC/PBMC responder cells were cultured with the same unrelated stimulator PBMCs. Briefly, a constant number (5×104) of irradiated (30 Gy, Cs137 source, Gammacell 1000 Elite) stimulator cells was cultured with graded numbers (5-0.125×104) of responder cells in 96-well round-bottomed microculture plates (Nunc) in a total volume of 200 µl/well. Sixteen to 24 replicates were used for each responder cell concentration. Control cultures for the calculation of background activity consisted of 16-24 wells of irradiated stimulator cells alone or irradiated responder cells plus nonirradiated autologous cells for CTLp and HTLp frequency estimation, respectively. After 3 days of incubation, 100 µl of supernatant was removed from each individual well and transferred to 96-well flat-bottomed microtiter plates (Nunc) for interleukin-2 (IL-2) estimation. The original limiting dilution cultures were fed on day 3 and day 6 with 100 µl of fresh culture medium supplemented with recombinant IL-2 (5 IU/ml). The frequency of HTLp was determined by measuring IL-2 concentration in the supernatant collected from day 3 limiting dilution cultures using a sensitive bioassay. The frequency of CTLp was determined by measuring cytolytic activity of the effector cells in limiting dilution cultures using a standard 4-hr 51Cr release assay.
Estimation of limiting dilution data and statistical analysis. The mean counts per minute plus 3 standard deviations of background control wells was used as a baseline for positivity. The wells that had higher counts per minute than the baseline were scored as positive. The estimation of HTLp and CTLp frequencies from LDA data was based on the Poisson distribution. The values for T cell frequency and 95% confidence intervals and the probability that the data were in conformity with single-hit kinetics were determined by maximum likelihood estimation and chi-square analysis using a computer program based on statistical methods described by Strijbosch et al.(12,13), Only the data that fit a single-hit model have been taken into account for the estimation. The statistical significance of the differences in HTLp and CTLp frequencies between testing groups was determined by one-way analysis of variance. The difference in in vitro skin GVH type alloreactivity induced by CBMCs or PBMCs was tested by Mann-Whitney test for statistical significance. The significance of the association between the levels of alloreactive CTLp frequency and the degree of in vitro skin GVH type alloreactivity (histopathological grades I-IV) was determined by Fisher's exact test.
HLA types and the degree of HLA mismatch between stimulators/responders. In this study, CBMCs and HLA-haploidentical parental PBMCs were used as paired responder cells, which were stimulated with the same unrelated PBMCs. HLA typing was complete for all samples in 16/19 experiments. In these cases, the stimulator/responder pairs were all 3-6 HLA Ag-mismatched (Table 1). Among the pairs, whose HTLp frequencies were measured, the degree of HLA-DR mismatch between stimulator and responder was the same for CBMCs and PBMCs in 8 out of 12 pairs (Table 1, pairs 1, 2, 6, 8, 9, 11, 12, and 15). Among the pairs, whose CTLp frequencies were determined, the degree of HLA-A and B mismatch was the same for CBMCs and PBMCs in seven out of nine pairs (Table 1, pairs 2, 4, 5, 11, 12, 14, and 15).
Alloreactive HTLp and CTLp frequencies in CBMCs and PBMCs. A wide frequency range of alloreactive HTLp and CTLp was observed for both CBMCs and PBMCs (Fig. 1). There was no significant difference in alloreactive HTLp frequencies between CBMCs and PBMCs (1:7,586 and 1:5,976, respectively). However, the mean value of alloreactive CTLp frequency was significantly lower in CBMCs (1:35,694) compared with that of PBMCs (1:5,333) (P=0.018, one-way analysis of variance).
In order to reduce the inaccuracy of statistical analysis induced by large individual variation in alloreactive T cell frequencies, more informative comparisons were made based on each CBMC/PBMC pair. The frequencies were regarded as significantly different if their 95% confidence intervals (approximately 2 SD) did not overlap. The results confirmed that the alloreactive CTLp frequencies in CBMCs were significantly lower than those in PBMCs in 10/12 (83%) pairs and the same as PBMCs in 2/12 (17%) pairs, whereas the alloreactive HTLp frequencies in CBMCs were the same as those in PBMCs in 7/15 (47%) pairs, lower than PBMCs in 5/15 (33%) pairs, and higher than PBMCs in 3/15 (20%) pairs (Table 2).
Alloreactive HTLp and CTLp frequencies were also analyzed in the pairs where the same degree of mismatch was found in HLA-DR (Tables 1 and 2, pairs 1, 2, 6, 8, 9, 11, 12, and 15) or HLA-A and B (Tables 1 and 2, pairs 2, 4, 5, 11, 12, 14, and 15) for CBMCs and PBMCs. The results showed that the mean HTLp frequencies were 1:11,976 and 1:5,302 for CBMCs and PBMCs, respectively. The difference was not statistically significant (P=0.459). In contrast, the mean frequency of CTLp was still significantly lower in CBMCs than in PBMCs (1:95,255 and 1:6,602, respectively, P=0.038).
In vitro skin alloreactivity and its association with CTLp frequency. The grades of skin histopathological damage are shown in Table 3 and Fig. 2 (a-d). In 7 of 15 pairs, CBMCs demonstrated lower grades of in vitro skin alloreactivity than parallel PBMCs (Table 3). The difference in the severity of in vitro skin alloreactivity induced by CBMCs or PBMCs was statistically significant (P=0.022, Mann-Whitney test).
In 8 of 15 pairs, alloreactive CTLp frequencies were determined in parallel with the skin explant assay. The results demonstrated that there was a positive association between the levels of alloreactive CTLp frequency and the severity of in vitro skin alloreactivity for CBMCs (P=0.017). Furthermore, in all samples (CBMCs and PBMCs) with high levels of alloreactive CTLp frequencies (greater than 1:100,000), high grades of in vitro skin alloreactivity (grade III-IV) were observed and those with lower levels of CTLp frequencies (<1:100.000) demonstrated lower grades of in vitro skin alloreactivity, with minimal epidermal cell damage (P<0.001, Fisher's exact test) (Table 3, Fig. 3).
Allogeneic BMT is an accepted curative or life-extending treatment for a number of hematological malignancies, bone marrow failure syndromes, immunodeficiency disorders, and inborn errors of metabolism (14). Successful implementation of allogeneic BMT is limited by the lack of suitable donors and the occurrence of fatal complications, one of the most important being GVHD. To increase the availability of marrow transplant therapy for patients who lack closely matched related donors, human umbilical CB has been used as an alternative source of hematopoietic stem cells for transplantation (1). Beyond the practical benefits associated with CB transplantation, such as bank storage, which offers immediate availability of donor stem cells, absence of donor risk and very low risk of transmissible infectious disease, another important potential advantage of CB transplants is reduced incidence of GVHD (2,3,15,16). This feature may reduce the morbidity and mortality associated with allogeneic BMT and permit a greater degree of HLa disparity between donor and recipient that would consequently expand the available donor pool. Although recent clinical data show that the incidence of severe GVHD after CB transplants is lower than that seen after BMT, the immunological mechanisms underlying this phenomenon are not completely understood. Furthermore, the majority of reported CB transplants are performed in children, in whom there is a naturally reduced incidence of GVHD (17). Few adult CB transplants have been documented, and therefore the capacity of CB cells to mediate severe GVHD in adult recipients is as yet unknown.
Alloreactive HTLp and CTLp frequencies determined by LDA assay are highly sensitive and quantitative parameters indicating the strength of alloreactive immune responses induced by mismatched HLA Ags (18-20). High levels of alloreactive HTLp and CTLp frequencies have been shown to be associated with the incidence of clinical GVHD (21-24). However, only a few studies have been carried out to determine alloreactive HTLp and CTLp frequencies in CB, and the results remain controversial.
The present work demonstrated that there was no significant difference in alloreactive HTLp frequencies between CBMCs and PBMCs. The findings of the comparable levels of alloreactive HTLp frequencies in CBMCs and PBMCs is consistent with an earlier observation demonstrating that the of CB cells to produce IL-2 and respond to IL-2 is intact (25). A number of other studies also confirmed that, after activation with different stimuli, including alloantigens, IL-2 production by CBMCs was comparable to that by PBMCs (26,27). In agreement with our findings, another recent study has similarly reported comparable HTLp frequencies in CBMCs and PBMCs (28). However, these results are in contrast with a previous study showing increased HTLp frequencies in CBMCs as compared with PBMCs (29). There are considerable variations in the methodology used by different investigators, which makes it difficult, sometimes impossible, to compare the results of different research groups. The use of different allogeneic combinations may also lead to discordant results, as some HLA types may be more alloreactive than others in inducing alloresponses (30,31). Moreover, in both our present work and a previous study by others (29), differences between the stimulator and responder at the DQ and DP loci have not been taken into account. Mismatching at the DQ and DP loci could however result a wide range of alloreactive HTLp frequencies (31) and may have influenced these results.
Similarly, alloreactive CTLp frequencies in CBMCs have been reported to be either the same as (28,29) or 10-100 times lower than (32) those in PBMCs. Significantly lower CTLp frequencies were detected in CBMCs as compared with PBMCs in this study. Apart from different methodology, these discrepant results could be a result of differences in the degree or the nature of HLA class I mismatches. The previous study showing no significant difference in CTLp frequencies between CBMCs and PBMCs (29) used the same homozygous B-LCL cell line (A11, B35, DR1) as the stimulator cells for all CBMC and PBMC responders. It has recently been shown that there is at least 10-fold differences in CTLp frequencies in the same responder against different individual HLA Ags (33). The exact mechanisms by which CB cells demonstrate lower levels of alloreactive CTLp frequencies and induce less severe skin epidermal cell damage in vitro are not fully understood. They could be the result of a deficiency in the activation, differentiation and/or functional maturation of CB cytotoxic T cells. As sufficient exogenous IL-2 has been used in the CTLp assay, the low production, by CBMCs, of other cytokines required for the alloresponse may play a part (26,27). A recently published observation demonstrated that CB T cells lack constitutive perforin expression (34). According to recent reports, 80-100% of CB T cells express the CD45RA+/CD45RO- phenotypes (26,27,35). The predominance of naive T cell population in CB may result in more stringent activation requirement of CB cytotoxic T cells. The presence of suppressor cells or other suppressive factors may also prevent the generation or the function of alloreactive cytotoxic T cells in CB (36,37).
It might perhaps have been expected that CB would show less variety in alloreactive T cell frequencies than adult blood, inasmuch as, in adults, the alloreactive T cell repertoire has been influenced by exposure to environmental Ags. The results of this study suggest that, as already observed in adult peripheral blood (19,38), great variations in alloreactive HTLp and CTLp frequencies between different allocombinations also exist in CB. In our present study, among a group of stimulator/CB responder pairs with a similar degree of HLA disparities, alloreactive CTLp frequencies could be either as high as 1:1,667 or as low as <1:500,000. This result reflects the fact that parameters other than the number of HLA mismatches are involved in generating significant alloreactivity by CB cells. These undefined factors may have an influence on CB donor selection and the clinical outcome of CB transplantation. The HLA genes themselves and the maternal immune status during the pregnancy, for example, may influence the development of alloreactive T cell repertoire in CB.
The use of a human skin explant model has shown, for the first time, that CBMCs are less able to induce skin GVH type alloreactivity in vitro, and this was associated with decreased levels of alloreactive CTLp frequencies in CB. However, it is notable that the majority of CBMC responders were capable of inducing skin GVH type alloreactivity in vitro, although the grade of histological damage was lower than that induced by PBMC responders. The potential of GVHD incidence after CB transplants may therefore not be absent, but rather the severity of GVHD may be reduced.
In conclusion, this study presents the first clear in vitro demonstration that a proportion of CBMC samples can tolerate high levels of HLA mismatches (3-6 Ags) without generating high levels of alloreactive CTLs or severe skin GVH type alloreactivity. The results provide a cellular basis for the possibility of using HLA-mismatched unrelated CB cells as transplant material. Further investigations regarding the expression of cell surface adhesion molecules, the function of cytolytic pathways, and the regulation of cytokines would provide important insights into the biological and immunological nature of CB cells.
Acknowlegments. The authors thank the staff of the tissue typing laboratory at the National Blood Service, Newcastle, for the HLA typing of adult blood and CB used in this study.
1. Broxmeyer HE, Douglas GW, Hangoc G, et al. Human umbilical cord blood as a potential source of transplantable hematopoietic stem/progenitor cells. Proc Natl Acad Sci USA 1989; 86: 3828.
2. Wagner JE, Kernan NA, Steinbuch M, et al. Allogeneic sibling umbilical cord blood transplantation in children with malignant and non-malignant disease. Lancet 1995; 346: 214.
3. Kurtzberg J, Laughlin M, Graham M, et al. Placental blood as a source of hematopoietic stem cells for transplantation into unrelated recipients. N Engl J Med 1996; 335: 157.
4. Apperley JF. Umbilical cord blood progenitor cell transplantation. Bone Marrow Transplantation 1994; 14: 187.
5. Broxmeyer HE. Questions to be answered regarding umbilical cord blood hematopoietic stem and progenitor cells and their uses in transplantation. Transfusion 1995; 35: 694.
6. Ademokun JA, Chapman C, Dunn J, et al. Umbilical cord blood collection and separation for haematopoietic progenitor cell banking. Bone Marrow Transplant 1997; 19: 1023.
7. Dickinson AM, Sviland L, Carey PJ, et al. Skin explant culture as a model for cutaneous graft versus host disease in humans. Bone Marrow Transplant 1988; 3: 323.
8. Sviland L, Dickinson AM, Carey PJ, et al. An in vitro predictive test for clinical graft versus host disease in allogeneic bone marrow transplant recipients. Bone Marrow Transplant 1990; 5: 105.
9. Lerner KG, Kao GF, Storb R, et al. Histopathology of graft versus host reaction (GVHR) in human recipients of marrow from HLA-matched sibling donors. Transplantation 1974; 18: 295.
10. Dickinson AM, Sviland L, Hamilton PJ, et al. Cytokine involvement in predicting clinical graft-versus-host disease in allogeneic bone marrow transplant recipients. Bone Marrow Transplant 1994; 13: 65.
11. Wang XN, Proctor SJ, Dickinson AM. Frequency analysis of recipient-reactive helper and cytotoxic T lymphocyte precursors using a combined single limiting dilution assay. Transplant Immunol 1996; 4: 247.
12. Strijbosch LWG, Buurman WA, Does RJMM, et al. Limiting dilution assays: experimental design and statistical analysis. J Immunol Methods 1987; 97: 133.
13. Strijbosch LWG, Does RJMM, Buurman WA. Computer aided design and evaluation of limiting and serial dilution experiments. Int J Biomed Comput 1988; 23: 279.
14. Gratwohl A, Hermans J, Baldomero H. Blood and marrow transplantation activity in Europe 1995. Bone Marrow Transplant 1997; 19: 407.
15. Wagner JE, Rosenthal J, Sweetman R, et al. Successful transplantation of HLA-matched and HLA-mismatched umbilical cord blood from unrelated donors: analysis of engraftment and acute graft-versus-host disease. Blood 1996; 88: 795.
16. Kurtzberg J, Graham M, Casey J, et al. The use of umbilical cord blood in mismatched related and unrelated hemopoietic stem cell transplantation. Blood Cells 1994; 20: 275.
17. Eisner MD, August CS. Impact of donor and recipient characteristics on the development of acute and chronic graft-versus-host disease following pediatric bone marrow transplantation. Bone Marrow Transplant 1995; 15: 663.
18. Speiser DE, Loliger CC, Siren MK, et al. Pretransplant cytotoxic donor T cell activity specific to patient HLA class I antigens correlating with mortality after unrelated BMT. Br J Haematol 1996; 93: 935.
19. Schwarer A, Jiang YZ, Deacock S, et al. Comparison of helper and cytotoxic antirecipient T cell frequencies in unrelated bone marrow transplantation. Transplantation 1994; 58: 1198.
20. Kaminski E, Sharrock C, Hows J, et al. Frequency analysis of cytotoxic T lymphocyte precursors: possible relevance to HLA-matched unrelated donor bone marrow transplantation. Bone Marrow Transplant 1988; 3: 149.
21. Spencer A, Brookes P, Kaminski E, et al. Cytotoxic T lymphocyte precursor frequency analysis in bone marrow transplantation with volunteer unrelated donors: value in donor selection. Transplantation 1995; 59: 1302.
22. 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. N Engl J Med 1992; 327: 1613.
23. Schwarer AP, Jiang YZ, Brookes PA, et al. Frequency of anti-recipient alloreactive helper T-cell precursors in donor blood and graft-versus-host disease after HLA-identical sibling bone-marrow transplantation. Lancet 1993; 341: 203.
24. 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.
25. Fairfax CA, Borzy MS. Interleukin 2 production, proliferative response, and receptor expression by cord blood mononuclear cells. J Clin Lab Immunol 1988; 27: 63.
26. Harris DT, Schumacher MJ, Locascio J, et al. Phenotypic and functional immaturity of human umbilical cord blood T lymphocytes. Immunology 1992; 89: 10006.
27. Roncarolo MG, Bigler M, Ciuti E, et al. Immune responses by cord blood cells. Blood Cells 1994; 20: 573.
28. Keever CA, Abu-Hajir M, Graf W, et al. Characterization of the alloreactivity and anti-leukemia reactivity of cord blood mononuclear cells. Bone Marrow Transplant 1995; 15: 407.
29. Deacock SJ, Schwarer AP, Bridge J, et al. Evidence that umbilical cord blood contains a higher frequency of HLA class II specific alloreactive T cells than adult peripheral blood. Transplantation 1992; 53: 1128.
30. Roelen DL, van Bree SPMJ, van Beelen E, et al. Cytotoxic T lymphocytes against HLA-B antigens are less naive than cytotoxic T lymphocytes against HLA-A antigens. Transplantation 1994; 57: 446.
31. 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.
32. Harris DT, Locascio J, Besencon FJ. Analysis of the alloreactive capacity of human umbilical cord blood: implications for graft-versus-host disease. Bone Marrow Transplant 1994; 14: 545.
33. Breur-Vriesendorp B, Vingerhoed J, van Twuyver E, et al. Frequency analysis of HLA-specific cytotoxic T lymphocyte precursors in humans. Transplantation 1991; 51: 1096.
34. Berthou C, Legros-Maida S, Soulie A, et al. Cord blood T lymphocytes lack constitutive perforin expression in contrast to adult peripheral blood T lymphocytes. Blood 1995; 85: 1540.
35. Beck R, Lam-Po-Tang, PR. Comparison of cord blood and adult blood lymphocyte normal ranges: a possible explanation for decreased severity of graft versus host disease after cord blood transplantation. Immunol Cell Biol 1994; 72: 440.
36. Han P, Hodge G, Story C, et al. Phenotypic analysis of functional T-lymphocyte subtypes and natural killer cells in human cord blood: relevance to umbilical cord blood transplantation. Br J Haematol 1995; 89: 733.
37. Roncarolo MG. The role of interleukin-10 in transplantation and GVHD. In: Ferrara JLM, Deeg HJ, Burakoff SJ, eds. Graft-vs-host disease. New York: Marcel Dekker Inc., 1996; 694.
38. Sharrock CEM, Man S, Wanachiwanawin W. et al. Analysis of the alloreactive T cell repertoire in man. I. Differences in precursor frequency for cytotoxic T cell responses against allogeneic MHC molecules in unrelated individuals. Transplantation 1987; 43: 699.