Chronic Myeloid Leukemia Patients with Deep Molecular Responses to Tyrosine Kinase Inhibitors Have Increased Effector Natural Killer and Cytotoxic T Cell Immune Responses to Leukaemia-Associated Antigens and Concomitant Reduced Immune Suppressors

We hypothesized that immune responses contribute to deep BCR-ABL molecular responses in chronic phase chronic myeloid leukaemia (CML) patients on tyrosine kinase inhibitors (TKI). We studied 32 CML patients; 16 at diagnosis, patients treated with imatinib (n=20), nilotinib (n=9) or dasatinib (n=3).

Methodology: The effector immune responses of Natural Killer (NK) cells were characterized by flowcytometry and functional analysis by CD107a degranulation assay. Cytotoxic T lymphocyte (CTL) responses to leukaemia-associated antigens (LAAs) WT1, BMI-1, PR3 and PRAME were quantified by interferon-gamma ELISPOT using peptide libraries of 15-mer peptides overlapping by 11 amino acids spanning the entire protein, or HLA-A0201 specific peptides in HLA-A0201+ patients. Immune suppressor regulatory T cells (Treg), Myeloid Derived Suppressor Cells (MDSC), Programmed cell death-1 (PD-1) expression on T cells, NK cells, B cells and monocytes, and major B cell subsets were extensively characterized by flowcytometry.    

Results: Patients in deep molecular response (MR^4.5; BCR-ABL <0.0032%) displayed increased antigen-specific CTL responses to LAAs, both in the number of positive LAAs and frequency of responses, compared to patients at diagnosis and major molecular response (MMR; BCR-ABL <0.1%). The most abundant LAA response was to PRAME [51% of patients in MR^4.5 compared to 31% in MMR and 0% at diagnosis] and WT1 [31% of patients in MR^4.5 compared to 28% in MMR and 0% at diagnosis]. PR3-specific immune responses were the least abundant, with no difference in response between MR^4.5and MMR (both 3%) compared to 0% at diagnosis.

Immunophenotypic analysis revealed a shift toward a more mature, cytolytic NK cell phenotype (CD57⁺, CD161⁺CD62L^-) in MMR and MR^4.5, consistent with up-regulation of the CD94/NKG2 family of inhibitory/activating receptors (NKG2A, NKG2C and NKG2D), the cytotoxicity triggering receptor NKp46 and a functional increase in NK cell cytotoxicity capacity against K562 target cells. The percentage of CD3^-CD56^dimCD16^bright cytolytic NK cells as a proportion of total lymphocytes was significantly increased in MMR and MR^4.5 [33.6% ± 6.6 p=0.0008 and 33.1% ± 4.1 p=0.01, respectively] compared to 7.8% ± 2.8 at diagnosis.

The absolute Treg number/µl was significantly lower in patients in MMR and MR^4.5 [13.9 ± 1.7 and 10 ± 1.1, respectively] compared to 32.7 ± 4.4 at diagnosis. Similarly, MDSC were significantly reduced in patients in MMR and MR^4.5 [3.9 ± 0.9 and 1.9 ± 0.5 MDSC/µl] compared to diagnosis [18.3 ± 3.9]. A predominantly granulocytic (CD66b⁺CD15⁺) MDSC phenotype was seen in CML patients at diagnosis. PD-1 expression as a proportion of total lymphocytes was significantly decreased in cytotoxic CD8⁺ T cells in MR^4.5 [5.7% ± 1.2] compared to MMR [12.3% ± 2.0, p=0.008] and patients at diagnosis [21.7% ± 5.2, p=0.0003]. PD-1 expression was decreased in CD4⁺ helper T cells in MR^4.5 [7.5% ± 1.7] compared to MMR [11.4% ± 1.5, p=0.07] and diagnosis [17% ± 2.9, p=0.008]. Overall, PD-1 expression was lower in NK cells in CML patients, albeit significant in MMR and MR^4.5 [0.24% ± 0.09, p=0.006 and 0.36 ± 0.07, p=0.02, respectively] compared to [1.42% ± 0.4] at diagnosis. No difference in PD-1 expression was seen in B cells or monocytes.

No significant difference was observed in CD3^-CD19⁺ B cells in MMR and MR^4.5 or at diagnosis.  Analysis of major B cell subsets revealed no difference in the proportion of transitional, naïve or memory B cells, plasma blasts or plasma cells.           

Conclusion: Enhanced effector immune responses of NK and LAA-specific CTLs are associated with concomitant reduction in immune suppressor activity, and may increase the rate of deep molecular responses to TKIs in CML. Methods to augment these responses may result in greater rate of success in TKI cessation studies.