Abstract

The advent of imatinib, and subsequently, other tyrosine kinase inhibitors (TKIs) has relegated allogeneic stem cell transplantation (SCT) to second-or third-line therapy for chronic phase-chronic myeloid leukemia (CML). A significant shortcoming of TKIs in the large majority of good-responder patients is the persistent detection of BCR-ABL transcripts despite achievement of complete cytogenetic response (CCyR) or major molecular response (greater than 3-log reduction of BCR-ABL levels below a standardized baseline in minimal residual disease (MRD) measurements from total mononuclear cells). The presence of this MRD, and the demonstration that primitive CD34+ cells from patients in CCyR still harbor BCR-ABL (Bhatia, et al, Blood 2003) strengthens the opinion that TKIs do not eradicate all CML cells. Conversely, SCT has apparently cured some CML patients with more than 20 years follow-up. Of note, MRD is also quite frequently found in CML patients more than 5 years post-SCT, and donor lymphocyte infusions (DLI) are usually given to patients who satisfy criteria for molecular relapse. We compared the level of BCR-ABL transcripts in primitive hematopoietic cells from patients who were treated with T-depleted SCT with scheduled T-cell addback from D+45 to D+100 post-SCT (n=34) with levels in those taking TKIs (n=4). CD34+ progenitor cells were isolated from leukapheresis collections at D+120 post-SCT (n=13), peripheral blood from D+60 – 66 months post-SCT (n=19), bone marrow aspirates in patients on long-term follow-up (4–10 years) post-SCT (n=8) or 15 – 18 months post-TKI treatment alone (n=4); serial post-SCT samples were available in 6 patients. Using fluorescence activated cell sorting, hematopoietic stem cells (HSC, CD34+CD38-Lin-CD90+), common myeloid progenitors (CMP, CD34+CD38+Lin-IL3Rα+CD45RA-) and granulocyte-macrophage progenitors (GMP, CD34+CD38+Lin-IL3Rα+CD45RA+) were collected. BCR-ABL expression in primitive CD34+ subpopulations and total leukocytes from peripheral blood (PB) was measured using real-time quantitative polymerase chain reaction, with the sensitivity to detect one BCR-ABL-positive cell in 106 normal cells. A median of 112 x 106 mononuclear cells (range 16 – 582 x 106) were available per patient sample, with a median of 3 x 106 CD34+ cells (range 1 – 90 x 106) sorted. There was no difference between the number of HSC, CMP or GMP CD34+ cells collected between MRD negative or MRD positive-post-SCT patients, or TKI-treated patients. All patients with negative MRD in PB (n=12 post-SCT, n=1 post-TKI) did not have detectable BCR-ABL transcripts in primitive CD34+ subpopulations. Furthermore, in PB MRD-positive patients, 3/21 (14%) post-SCT and 2/2 (100%) post-TKI did not have detectable BCR-ABL transcripts in HSC, CMP or GMP populations. 18/21 (86%) PB MRD-positive patients who were post-SCT had detectable BCR-ABL transcripts in at least one primitive CD34+ subpopulation. A rise in BCR-ABL levels in GMP populations tended to herald impending relapse. In post-SCT patients on long-term follow-up with persistent PB MRD positivity not fulfilling criteria for molecular relapse, the highest BCR-ABL levels were in HSC populations. In comparable patients with a major molecular response, post-SCT patients appeared to harbor a greater amount of residual leukemia cells in CD34+ subpopulations than TKI-treated patients. Our observations suggest that although SCT is a curative treatment for CML, the graft-versus-leukemia effect may eliminate only more mature leukemic CD34+ subpopulations in some patients who have enduring positive MRD post-SCT, with persistence of the most primitive leukemic HSCs, which are presumably constrained in patients who do not fulfill criteria for relapse. Conversely, TKI appears to reduce the number of BCR-ABL-positive primitive CD34+ subpopulations, especially GMPs and CMPs, more efficiently. Our data support TKI-treatment as an adjunct to DLI to treat CML relapse post-SCT, and concurrent vaccination strategies which are able to target surface proteins on HSC to eradicate disease.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author