Despite well-known Natural Killer (NK) cells antitumor activity, its potential in regulating Chronic Myeloid Leukemia (CML) is not fully stablished. Tyrosine kinase inhibitors (TKI) treatment, besides reducing BCR-ABL1 transcripts, seems to result in immunomodulation that partly restore the immune dysregulation of the disease. Numerical, functional and maturation defects might be associated with CML risk stratification, molecular response profile or even with resistance to treatment with TKI. Therefore, we aimed to immunophenotypically characterize frequency, subtypes and maturation profile of NK cells in CML patients, by Multiparametric Flow Cytometry (MFC), during monitoring of TKI treatment.

11 Peripheral Blood samples from patients with CML (Clinical Hospital of Ribeirão Preto - HCRP-USP) and 10 healthy donors (controls) were analyzed. Clinical, laboratory and ELN2017 risk stratification were assessed. NK cells were defined as CD19-CD3-CD56+ and: CD56brightCD16- (secretory), CD56dimCD16+ (cytotoxic), CD11b-CD27- (tolerant); CD27+CD11b- (immature secretor), CD27+CD11b+ (secretory) and CD11b+CD27- (cytotoxic). Based on molecular response levels assessed by BCR-ABL1 RT-qPCR, patients were classified in G1 (without loss of molecular response to the 1st line of TKI; n = 3), G2 (loss of molecular response to 1st line of TKI; n = 3) and G3 (loss of molecular response to 2nd generation TKI; n = 5). Mann Whitney or Kruskal-Wallis tests (p < 0.05) were applied using SPSS (v.20).

A higher frequency of lymphocytes in CML patients was observed when compared to controls (p = 0.0242), but no difference was found in the frequency of NK cells. G1 and G2 showed a higher frequency of NK cells than G3 (p = 0,7774). Even though no statistical significance was found, CD27-CD11b- and CD27+CD11b- NK cells showed a higher frequency, while CD27+CD11b+ NK cells showed a lower frequency in CML when compared to controls. Frequency of CD27+CD11b- NK cells was higher in G2 and G3 when compared to G1. G3 showed an immature profile with higher percentages of CD27-CD11b- and CD27+CD11b- NK cells when compared to controls, while the more mature CD27+CD11b+ NK cells showed a lower frequency in CML. Furthermore, a lower frequency of CD56brightCD16- (p = 0.0002) and CD56dimCD16+ (p = 0.0448) NK cells was observed in CML, as wells as the frequency of CD27-CD11b+ cytotoxic NK cells, consistent to what we found in G3 and supporting our hypothesis.

We observed that patients with loss of molecular response in treatment with 1st and 2nd TKI generation presented an immature non-effective NK cell profile, with deficient cytotoxicity. This scenario might result in impaired tumor surveillance of CML, meaning that phenotypic changes in NK cells may directly affect risk stratification and treatment response in CML. However due to the small number of patients and the lack of a functional assay, it is not possible to confirm that the cytotoxic cell may actually be deficient. Our findings and further functional studies can contribute to a better understanding of the physiopathology, prognosis and monitoring of CML, response to TKI therapy, as well as help to define therapeutic strategies based on immunotherapy mediated by NK cells.