Higher Percentage of CD34⁺CD38⁻ Cells Detected by Multiparameter Flow Cytometry From Leukapheresis Products Predict Unsustained Complete Remission in AML

Myeloablative chemotherapy followed by autologous PBSCT remains one treatment strategy in adult AML patients. Relapse has been shown higher for those who received the highest CD34⁺ PB doses. Although highly active against the leukemic bulk, intensive chemotherapy often spare the hardier leukemia stem cells (LSCs) responsible for relapse. Detection of MRD in harvest products may reflect inadequate in vivo purging at least in part responsible for relapses. Although recent data have challenged the CD34⁺CD38⁻ phenotype of LSCs, this cell population remains generally considered enriched for LSCs. In this setting, MRD remaining during CR should be relatively enriched in CD34⁺CD38⁻ leukemic cells and their persistence should correlate with disease recurrence.

CD34⁺ cells were harvested after CR achievement in 123 AML patients [median age: 53 y (25–72)] treated by induction chemotherapy in our Institution between 10/1994 and 04/2003. Patients were included in different clinical trials planning autologous SC harvest in CR and autologous SCT in absence of donor or allogeneic SCT indication. Seventy-one of them received effectively autologous PBSCT. Harvests performed in 15 normal donors were used as controls. CD34/CD38 cell profile was analyzed in harvests in one single tube by multicolor flow cytometry using multiple MoAbs. The gating strategy was based on CD45^low/SSC and CD34⁺CD45^low cell populations from total FSC/SSC viable cells. Three populations of CD34⁺ were distinguished: CD34⁺CD38^–; CD34⁺CD38^low; and CD34⁺CD38⁺.

Patients from the entire cohort with higher percentage of CD34⁺ cells (cut-off level: 1%) in PBSC products were associated with shorter EFS [median: 5.6 months (3-y EFS: 13%) vs 13.6 (37%); p=0.0005] and OS [median: 10 months (3-y OS: 19%) vs 23.4 (47%); p=0.004]. This was also the case when analyzing only patients who received autologous SCT: [median EFS: 5 months (3-y EFS: 13%) vs 22.2 (48%); p=0.0006, and median OS: 9.1 months (3-y OS: 21%) vs 43.3 (57%); p=0.001]. Among CD34⁺ populations, only CD34⁺CD38^– had a prognostic impact on EFS and OS. At a cut-off level of 0.9%, median EFS was 8.2 months (3-y EFS: 29%) for those with higher percentage vs 91.9 (62%) for those with lower percentage and median OS was 14.2 months (3-y OS: 36%) vs 95.4 (69%) respectively for the entire cohort. These results were confirmed in patients undergoing autologous SCT: median EFS was 7.3 months (3-y EFS: 31%) vs 91.1 (70%) (p= 0.05), and median OS was 14.4 months (3-y OS: 39%) vs 94.6 (80%). CD34⁺CD38^low and CD34⁺CD38⁺ populations did not show any prognostic impact. Harvests from AML patients were divided into 3 groups: Group A: 51 patients with CR duration <1 y; Group B: 22 patients with CR duration >1 y; and Group C: 50 patients without relapse. Harvests from 15 normal donors (Group D) were used as controls. Significant differences were only observed when comparing Group A and Group D for total CD34⁺ cells (mean ± SEM 2.5 ± 0.5 vs 1.2 ± 0.3; p < 0.05) and CD34⁺CD38^– (4.5 ± 0.7 vs 2.3 ± 0.5; p < 0.05). To confirm the prognostic value of CD34⁺CD38^–, 19 patients (Group 1) with evidence of leukemic contamination in harvests (abnormal cytogenetics at presentation found in aphereses) were compared with 22 patients (Group 2) without evidence of contamination (abnormal cytogenetics at presentation not found in aphereses). Median EFS was 10.1 months (3-year EFS: 45%) in Group 2 vs 6.3 (13%) in Group 1 (p=0.01), and median OS was 36.6 months (3-year OS: 55%) vs 10.8 (23%), respectively (p=0.03). Harvests from 15 normal donors (Group 3) served as controls. Significant differences were noted in harvest products regarding CD34⁺CD38^– between Group 1 and Group 3 (mean ± SEM 6.0 ± 1.5 vs 2.3 ± 0.5; p = 0.04) and Group 1 and Group 2 (6.0 ± 1.5 vs 2.4 ± 0.5; p = 0.03), while there were no differences between Group 2 and controls. There were no significant differences between groups regarding CD34⁺CD38^low and CD34⁺CD38⁺. We also measured MFI of CD13, CD33, CD123, CD117 in CD34⁺ subpopulations. Phenotypes were compared among the different groups.

Higher proportions of CD34⁺CD38⁻ in apheresis products appear to reflect inadequate in vivo purging and distinguish samples as enriched in ‘leukemic cells’ from those with lower CD34⁺CD38⁻ as largely constituted of ‘normal cells’. This could serve as detection of MRD and help to identify samples associated with high-risk of relapse after autologous SCT.

No relevant conflicts of interest to declare.