Donor lymphocyte infusion (DLI) induces complete remissions in 70–80% of patients with a relapse of CML after allogeneic stem cell transplantation (SCT). However, in some patients molecular persistence of bcr-abl positive stem cells can be found. To determine possible mechanisms of this persistence we studied in detail a patient who received DLI for a hematological relapse after HLA identical sibling SCT. After obtaining a hematological and cytogenetic remission in the absence of any graft versus host disease persistence of molecular disease was observed despite multiple additional escalating doses of DLI. To determine the nature of this T cell response, donor T cells isolated from bone marrow of the patient at the time of clinical response were stimulated with CML cells from the patient. T cells responding to the leukemia by secreting IFN-γ were isolated, directly cloned by cell sorting and expanded.

Four different types of CD8+ HLA class I restricted CTL clones were obtained. All clones were cytolytic against EBV-LCL of the patient, but did not lyse EBV-LCL of the donor, indicating recognition of a minor histocompatibility antigen (mHag). PHA blasts of patient and donor were not lysed, indicating that the clones recognized an antigen present on B-cells but not on T-cells. To further study the lineage restriction of the clones, bone marrow and peripheral blood CML cells were stained with CFSE and incubated with the clones. The survival of different cell populations was measured by counterstaining with different monoclonal antibodies (

). Two clones only lysed CD14 positive monocytic cells, two other clones lysed CD14 positive monocytic cells as well as CD13–33 positive myeloid cells. Interestingly, immature myeloid cells were not lysed, whereas recognition by the T cells did occur after maturation of these cells in vitro by growth factors. By morphological examination it was demonstrated that the cell clones are capable of lysing myelocytes but not myeloblasts. These results indicate that the clones recognize a mHag expressed during myeloid maturation.

Subsequently, we investigated whether the CTL clones could lyse CML progenitor cells using a hematopoietic progenitor cell inhibition assay (

van der Hoorn,
). In the presence of the CTL clones persistent proliferation of 20–40% was found, in contrast to the complete proliferation inhibition of CML progenitors observed when mHag clones HY-A1 and HA-1 were used as control. Secondly, survival studies showed that the CTL clones were not able to lyse CFSE labelled CD34 positive CML cells, in contrast to the control clones HY-A1 and HA-1.

In conclusion, different types of CD8+ mHag-specific CTL clones recognizing lineage specific antigens have been isolated from a relapsed CML patient with molecular persistence of bcr-abl after repeated administrations of DLI. These clones recognize mHag’s present on B-cells, monocytic cells and mature myeloid CML cells, but not on CML progenitor cells. The isolation of these CTL’s in the absence of T cells against CML progenitor cells explains the clinical hematological remission but molecular disease persistence of this patient after DLI. These findings suggest that for adequate eradication of CML a cytotoxic T cell response against leukemic progenitor cells is needed.

Author notes

Corresponding author