Persistence of Residual Circulating CD26+ Leukemia Stem Cells in Chronic Myeloid Leukemia Patients in Treatment Free Remission

Chronic myeloid leukemia (CML) patients achieving a deep and stable molecular response with tyrosine kinase inhibitors may discontinue TKI therapy and count on 50% about probability to maintain a "treatment free remission" (TFR) which is today the best surrogate for "cure" in CML. Albeit indispensable for attempting TFR, deepness and duration of molecular response are not the only criteria for a successful TFR as half of CML patients (pts) meeting those criteria still loose the response within 6-12 months after TKI discontinuation. Available data suggest that relapse after TKIs withdrawal is due to the persistence of leukemic stem cells (LSCs) intrinsically resistant to TKIs, however the presence and the impact of residual LSCs in CML patients in TFR has not been explored yet.

CD34+/CD38-/Lin- CML LSCs specifically co-express dipeptidylpeptidase IV (CD26). We have recently demonstrated that CD26+ LSCs are easily detectable by flow-cytometry also in peripheral blood of CML pts both at diagnosis and during tyrosine kinase inhibitors (TKIs) treatment (Bocchia M, Blood 2016 128:942). The number of circulating residual CD26+LSCs appeared not correlating to the degree of molecular response, suggesting that residual TKI resistant quiescent LSCs may be transcriptionally low/silent and thus potentially not quantifiable by standard qRT-PCR analysis.

We here evaluated the presence of circulating CD26+ LSCs in CML patients in TFR referring to several Italian hematology centers. During a follow-up visit, in which patients were checked for standard PB BCR-ABL1 molecular response, 3-6 mls of PB were collected in EDTA and sent to Siena Hematology Lab to detect CD34+/CD38-/CD26+ LSCs by multicolor flow cytometry. Cells were incubated with anti CD45, CD34 (581), CD38 (HIT2), CD26 (M-A261). Acquisition and analysis were performed by FACSCantoII using DIVA8 software.

A total of 84 CML pts in TFR for a median of 31 months (range 2-192) were evaluated. All 84 patients were in first line treatment at the time of TKI withdrawal. Fifty-one out of 84 (60%) pts discontinued imatinib treatment and were in TFR for a median of 37 months (range 2-192); 14/84 (17%) pts discontinued dasatinib and were in TFR for a median 32 months (range 5-90) and 19/84 (23%) pts were in TFR for a median of 18 months (range 3-39) after nilotinib treatment.

PB CD26+ CML LSCs were measurable in 54/84 (64%) pts with a median number of 0,0132 cells/µL (range 0,0053-1,77) while only 24/84 (28%) pts showed detectable BCR-ABL copies (median BCR-ABL/ABL^IS ratio 0,0062 range 0,0032-0,493). Indeed, 18/84 (21%) pts showed both circulating CD26+ LSCs and detectable BCR-ABL, 26/84 (31%) pts scored negative for both determinations, while 36/84 (43%) pts showed detectable CD26+ LSCs and undetectable BCR-ABL. Only 4/84 (5%) pts showed measurable BCR-ABL copies with undetectable CD26+ LSCs.

Kendall rank correlation coefficient, Mood test and bi-linear relation model of the whole cohort showed no correlation between BCR-ABL/ABL^IS ratio and number of residual LSCs, while a significant linear inverse correlation was found between number of circulating CD26+ LSCs and TFR duration (r=-0.055 p <0.01 p=0.0063). Similar results were found also when considering only 51 pts in TFR after imatinib. A statistical analysis focusing on dasatinib and nilotinib subgroup of pts was not possible for due to the reduced sample size.

This is the first time that residual circulating CML LSCs are detected in TFR CML pts even after long lasting TKI discontinuation and deep stable molecular response. Intriguingly, the evaluation of CD26+ LSCs appeared a more sensitive method than qRT-PCR to detect residual "quiescent" disease. The reason why CML LSCs may persist in pts who discontinued TKI, without resulting in an overt disease, it's still unknown but likely a CML LSCs specific immune control plays a crucial role. An ongoing prospective study, recruiting a consistent number of pts, monitoring the dynamics of circulating CD26+LSCs from TKI discontinuation may answer some still open questions. LSCs' clinical role, correlation with type and duration of previous TKI treatment as well interaction with active specific immune system will be investigated.