Specific depletion of alloreactive T cells in HLA-identical siblings: a method for separating graft-versus-host and graft-versus-leukaemia reactions

Accumulating evidence indicates that alloreactive donor T cells confer both graft-versus-host (GVH) and graft-versus-leukaemia (GVL) reactivity following allogeneic bone marrow transplantation. We have developed a method to deplete alloreactive donor T cells with an immunotoxin targeting the alpha chain of the IL-2 receptor. In patients with chronic myeloid leukaemia and their HLA-identical sibling donors, we measured donor helper T-lymphocyte precursor frequencies (HTLPf) against recipient peripheral blood mononuclear cells (PBMNC; donor versus host), recipient leukaemia cells (donor versus leukaemia) and third-party PBMNC, before and after the depletion. In seven pairs there was a 4.3-fold reduction of donor-versus-host HTLPf (P=0.017), without a significant change in the donor frequencies against third party (P=0.96). In eight further donor-recipient pairs, immunotoxin-depleted donor versus patient PBMNC HTLPf 4.5-fold (mean 1/155,000 before and 1/839,000 after depletion, P=0.007). There was a smaller non-significant 1.8-fold reduction in donor-versus-leukaemia HTLPf from 1/192,000 to 1/334,000 (P=0.19). These results suggest that selective T-cell depletion can significantly deplete donor anti-host reactivity while conserving anti-leukaemia reactivity in HLA-matched donor-recipient pairs.

T cell-depleted granulocyte colony-stimulating factor (G-CSF) modified allogenic bone marrow transplantation for hematological malignancy improves graft CD34+ cell content but is associated with delayed pancytopenia

To increase the stem cell content of T cell-depleted bone marrow transplants (BMT), we treated 12 patients with hematological malignancies with BMT from HLA-identical sibling donors given G-CSF 10 microg/kg/day for 5 days before marrow harvest. After CD34+ cell selection, patients received a median of 1.7 (range, 0.82-3.1) x 10(6) CD34+ cells/kg and 2.3 (range, 0.25-4.0) x 10(5) CD3+ cells/kg. All patients had initial engraftment but four developed pancytopenia between days 55-130 post-BMT. In two patients, this required a second infusion of G-CSF-mobilized donor peripheral blood progenitor cells. We observed no delayed pancytopenia in a matched historical group of 24 patients receiving T cell-depleted BMT without prior G-CSF stimulation. Compared to this control group, G-CSF-stimulated marrow recipients showed a significant decline in neutrophil and monocyte counts after 8 weeks. However, outcome after BMT was otherwise comparable, with a similar incidence of acute graft-versus-host disease and transplant-related mortality. Disease-free survival was 63 vs 67% for controls matched for CD34+ cell dose (P = NS). These results indicate that G-CSF stimulation can increase the CD34+ cell content of T cell-depleted marrow but carries a risk of late graft failure.

Preferential usage of T cell receptor (TCR) V beta by allogeneic T cells recognizing myeloid leukemia cells: implications for separating graft-versus-leukemia effect from graft-versus-host disease

New understanding of the alloresponse following bone marrow transplantation supports the possibility that the graft-versus-host disease (GVHD) response can be separated from a favorable graft-versus-leukemia (GVL) effect. We used chronic myeloid leukemia (CML) cells to generate 122 recipient-reactive T cell clones from a closely HLA-matched sibling responder. Clones were tested for their proliferative response to stimulator CML cells or PHA-transformed (non-leukemic) lymphoblasts. Of 78 clones tested, 32 recognized both leukemia cells and PHA blasts, 19 only CML and four only PHA blasts. The remainder were non-specific responders. This functional specificity corresponded to distinct patterns of T cell receptor (TCR) V beta usage: clones recognizing CML cells preferentially used V beta 5, V beta 6/7 while clones recognizing both CML and PHA blasts or only PHA blasts preferentially used V beta 3 and V beta 8. It may therefore be possible to identify in vitro-generated myeloid leukemia-restricted donor T cells by their pattern of V beta usage. TCR V beta antibodies could thus be used to select and expand leukemia-restricted donor T cells for transfusion after BMT to specifically enhance the GVL response.

Association of natural killer cell immune recovery with a graft-versus-leukemia effect independent of graft-versus-host disease following allogeneic bone marrow transplantation

There is good evidence that T lymphocytes play an important role in the graft-versus-leukemia (GVL) effect following allogeneic bone marrow transplantation (BMT) for hematologic malignancies. However, the role of natural killer (NK) cells in GVL is less clear. To further investigate a possible association of NK cells with GVL we studied 15 patients undergoing BMT for chronic myeloid leukemia (CML), correlating T-cell (CD4+ and CD8+) and NK-cell (CD16+ 56+) recovery with relapse and graft-versus-host disease (GVHD). Patients were studied on three occasions up to 9 months after BMT, for lymphocyte surface phenotype and for spontaneous and IL-2-stimulated (LAK cell) cytotoxic function. Circulating CD8+ and NK but not CD4+ cell numbers were significantly lower in five patients who relapsed compared with those remaining in remission after BMT (mean 0.03 vs 0.32 x 10(9)/l, p = 0.002 for CD8+ cells: mean 0.03 vs 0.11 x 10(9)/l, p = 0.002 for NK cells). There was no correlation of CD4+. CD8+, or NK cell numbers and development of grade-II or more acute GVHD. Spontaneous NK cytotoxic function rose to within the normal range in the first month after BMT. LAK function remained low during the study period. These results link NK cell recovery more closely with a GVL than with a GVH effect.

Specific depletion of alloreactivity against haplotype mismatched related individuals by a recombinant immunotoxin: a new approach to graft-versus-host disease prophylaxis in haploidentical bone marrow transplantation

Haploidentical bone marrow transplantation (BMT) is associated with a high risk of severe graft-versus-host disease (GVHD). While pan-T cell depletion of the graft is the most effective means of preventing severe GVHD, it is associated with delayed recovery of T cell function leading to fatal infections. We used two related Pseudomonas exotoxin-based immunotoxins, anti-Tac(Fv)-PE38 and anti-Tac(Fv)-PE38KDEL, that both target the IL-2 receptor on activated T cells, to specifically deplete alloreactive lymphocytes against haploidentical stimulators. The functional capacity of the remaining lymphocytes was tested in proliferative assays against the original haploidentical stimulator and pooled cells from other mismatched donors (third party). We varied the recombinant toxin concentration and schedule to determine the optimum conditions for selective depletion. In 10 experiments, the mean residual reactivity after depletion was 7.6 +/- 1.4% against the haploidentical stimulator and 64.2 +/- 5% against the third party, expressed as a percentage of the undepleted response to the same stimulators. Depletion was shown to be specific for mixed lymphocyte culture (MLC)-activated lymphocytes. The immunotoxin did not affect CFU-GM growth of normal BM cells. This selective depletion of haploidentical alloreactivity could be used to prevent GVHD while conserving immune recovery following haploidentical BMT.

Interaction of natural killer cells with MHC class II: reversal of HLA-DR1-mediated protection of K562 transfectant from natural killer cell-mediated cytolysis by brefeldin-A

Major histocompatibility complex (MHC) class I antigens on tumour cell surfaces have been shown to modulate target susceptibility to natural killer (NK) cell-mediated lysis in some, although not all, systems investigated. MHC class II expression may also affect NK cell function, but the mechanism by which MHC class II antigen regulates NK cell activity has not been fully examined. In this study we induced HLA-DR1 expression by gene transfection into the classic NK-sensitive K562 cell line to study the interaction of NK cells with MHC class II molecules and the effect of brefeldin-A (BFA), an endogenous antigen-processing pathway blocker, on NK-target cell interaction. We demonstrated that the expression of HLA-DR1 on the cell surface reduced K562 cell susceptibility to NK lysis by peripheral blood monuclear cells and a NK cell line. The effect was demonstrable in prolonged (8 hr) cytotoxicity assays and was blocked by pretreatment of target cells with anti-HLA-DR antibody. Treatment of K562 DR transfectant with BFA abrogated the resistance of K562 transfectant to NK-mediated cytolysis. These findings indicate that HLA class II molecules regulate NK cell function and target recognition, and suggest that endogenous peptides presented through MHC molecules are responsible for regulating NK cytolysis.

Distinct T cell populations distinguish chronic myeloid leukaemia cells from lymphocytes in the same individual: a model for separating GVHD from GVL reactions

Donor lymphocyte responses to minor histocompatibility antigen (mHA) differences are involved in allo-responses between HLA matched pairs causing GVHD and graft-versus-leukaemia (GVL). Since some mHA are tissue-restricted, GVHD and GVL responses may be separable. We studied donor lymphocyte responses to patients with CML in a series of 10 HLA-matched sibling and 10 unrelated donor-recipient pairs comparing proliferation to recipient PHA blasts and CML cells and attempting to selectively deplete responses to PHA blasts in vitro. Responses in counts per min (c.p.m) to CML cells and PHA blasts were, respectively, 2809 +/- 2205 (SD) and 7376 +/- 1877 in related and 12,107 +/- 7191 and 26,136 +/- 22,479 in unrelated pairs. Autologous responses to PHA blasts were significantly lower (mean 779 +/- 735) (p < 0.001). Results correlated with clinical outcome: higher responses to recipient cells correlated with transplant-related death (p = 0.02 for CML and p = 0.06 for PHA blasts). Higher responses to CML correlated with GVHD grade > or = II (p = 0.025). Donor lymphocytes exposed to recipient PHA blasts for 5 days and treated with a ricin-conjugated anti-CD25 antibody retained over 75% of their response to CML but < 10% to PHA blasts. Similarly, depletion of response to CML but not to PHA blasts occurred when CML was the primary challenge. These results indicate that distinct populations of donor T cells respond to recipient leukaemic and non-leukaemic cells, and provide the basis for a clinically applicable technique to selectively deplete donor GVHD reacting cells while conserving GVL.