Persistence of leukemia stem cells in chronic myelogenous leukemia patients in prolonged remission with imatinib treatment

Abstract

Imatinib mesylate treatment markedly reduces the burden of leukemia cells in chronic myelogenous leukemia (CML) patients. However, patients remain at risk for relapse on discontinuing treatment. We have previously shown that residual BCR-ABL(+) progenitors can be detected in CML patients within the first 2 years of imatinib treatment. However, reduced rates of relapse and continued decline of BCR-ABL levels with prolonged treatment, together with the ability of selected patients to maintain remission after discontinuing treatment, led us to investigate whether prolonged imatinib exposure resulted in reduction or elimination of BCR-ABL(+) stem cells. We evaluated BCR-ABL expression in CD34(+)CD38(+) (38(+)) committed progenitors and CD34(+)CD38(-) (38(-)) stem/primitive progenitor cells in samples from CML patients on imatinib treatment for at least 4 years with cytogenetic and molecular response. High levels of BCR-ABL expression were maintained over time in the 38(-) stem cell fraction. The absolute frequency of BCR-ABL(+) cells as determined by limiting dilution analysis was consistently higher in 38(-) compared with 38(+) cells. Transplantation into NOD/SCID-IL2Rγ-chain knockout mice demonstrated that BCR-ABL(+) cells had long-term in vivo repopulating capacity. These results directly demonstrate that BCR-ABL(+) stem cells persist in CML patients despite prolonged treatment with imatinib, and support ongoing efforts to target this population.

Figure 1

Reliability of BCR-ABL measurement with a small number of cells. RNA extracted from K562 cells was diluted 1:100 into RNA from HL60 cells. Ten-fold serial dilutions were performed to obtain 200 ng, 20 ng, 2 ng, and 0.2 ng, equivalent to 20 000, 2000, 200, and 20 cells, respectively, which were then used for Q-PCR analysis of BCR-ABL and BCR. (A) BCR-ABL and BCR measurements following Q-PCR of different dilutions of RNA are shown. (B) The ratio of BCR-ABL to BCR at the different dilutions is shown.

Figure 2

Persistence of BCR-ABL⁺ cells in primitive progenitor/stem cells in CML patients in CCR with imatinib mesylate treatment for 5 years. (A) Experimental procedures for isolating CD34⁺CD38⁺ and CD34⁺CD38⁻ cells by flow cytometry sorting followed by Q-PCR for BCR-ABL and BCR are shown. (B) The cumulative results for Q-PCR analysis for BCR-ABL levels in MNC, CD34⁺CD38⁺, and CD34⁺CD38⁻ cells in CML patients after 1 to 5 years of imatinib mesylate treatment are shown using box and whisker (Tukey) plots. Significance values comparing MNC and CD34⁺CD38⁻ cells are *.05 > P < .1, **P < .01, ***P < .001; comparing CD34⁺CD38⁺ and CD34⁺CD38⁻ cells are ^.05 > P < .1, ^P < .05.

Figure 3

Frequency of BCR-ABL⁺ cells in CD34⁺38⁺ or CD34⁺38⁻ populations. (A) A schematic of experimental and analytical procedures for measurement of the frequency of BCR-ABL–positive cells within flow cytometry selected CD34⁺CD38⁻ and CD34⁺CD38⁺ cells using limiting dilution analysis is shown. (B) The frequencies of BCR-ABL⁺ cells within the CD34⁺CD38⁺ (black bars) and CD34⁺CD38⁻ (grey bars) subpopulations in 14 samples obtained at different time points from 6 patients is shown (2-3 serial samples per patient).

Figure 4

Engraftment of NOD/SCID-IL2Ry-chain knockout (NSG) mice with CD34⁺ BM cells from CML patients treated with imatinib mesylate. (A) The schematic for these experiments is shown. BM CD34⁺ cells from CML patients in remission (n = 6) were transplanted via tail vein injection into sublethally irradiated NSG mice (3-4 mice per sample, total of 20 mice). Mice were evaluated for human cell engraftment and BCR-ABL status of engrafted cells after 11-12 weeks. (B) The percentage of human CD45⁺ cell engraftment in BM, spleen and PB of mice transplanted with CD34⁺ cells are shown. Results represent the mean ± SEM of human CD45⁺ cells engrafted in BM, spleen, and PB for each BM sample transplanted. (C) The percentage of human CD45⁺ cells expressing different surface markers in mouse BM, spleen, and PB are shown (n = 20). (D) The BCR-ABL status of engrafted cells was evaluated using Q-PCR. BCR-ABL expression in human cells engrafted in BM of individual mice is shown (n = 15). Horizontal bars represent median values. Samples from the 5 mice that were excluded also expressed BCR-ABL, but BCR values were too low to allow quantitation.