Mature, Adaptive-like CD56DIM NK Cells in Chronic Myeloid Leukemia Patients in Treatment Free Remission

Background: Tyrosine kinase inhibitors (TKIs) have significantly improved the treatment of CML.  Even though TKI treatment is generally not considered curative, recent studies have shown that nearly half of CML patients who have achieve good and durable responses are able to stop the TKI treatment. However, patients who have successfully discontinued TKI treatment still have residual disease. We hypothesized that the immune system plays a role in treatment free remission (TFR), and our preliminary results in the EURO-SKI trial showed that patients who relapse early after imatinib discontinuation have decreased numbers and frequencies of NK cells. In EURO-SKI trial relapse was defined as the loss of major molecular response (MMR). We now aimed to analyze in more detail the phenotype and function of the NK cells in order to understand their role in TFR.

Methods: Lymphocyte subclass analysis (the number of NK-, T- and B-cells) was performed at the time of therapy discontinuation and 1 month after the imatinib discontinuation in patients participating in the EURO-SKI stopping trial in the Nordic countries (n=105, results are presented from patients who have reached 6 months follow-up). More detailed immune phenotype and functional assays (NK-cell degranulation and secretion of Th1 type of cytokines IFN-γ/TNF-α) were analyzed from a proportion of patients (n=31).  

Results: Imatinib treated patients remaining in remission for 6 months (non-relapsing, n=48, median age 60,5 years) displayed an increased amount of NK cells at the time of drug discontinuation (18.6% vs. 11.0%, p=0.02, NK-cell count 0.25 x10⁹ cells/L vs. 0.184 x10⁹ cells/L m, p=0.059) compared to patients who relapsed early (before 5 months, n=29, median age 60,5 years). Furthermore, the NK cell frequency in non-relapsing patients was even higher than in healthy controls (11.5%, n=48, p=0.001).  T and B cell counts and frequencies showed no differences between the groups.

Detailed analysis of the NK cell compartment displayed a more mature phenotype for the NK cells in non-relapsing patients. Larger frequencies of NK cells from early relapsing patients was CD56^bright compared to non-relapsing patients  (4.8% vs. 2.7% of CD56 NK cells, p=0.04). Furthermore, patients who had higher frequencies of CD56^bright NK cells than median had decreased TFR at 6 months (42%) compared to patients with lower frequency (70%, p=0.01). In addition, there was a trend towards more CD57^pos (78% (n=21) vs. 66% (n=10), p=0.09) CD56^dim NK cells in non-relapsing patients.

To further study the mature NK cells in non-relapsing patients, recently identified markers (FceRg^neg, PLZF^neg , SYK^neg , EAT-2^neg ) for adaptive NK cells were analyzed. Interestingly, there was a trend that non-relapsing patients had higher frequencies of adaptive-like NK cells. For example, non-relapsing patients had more CD56^dim NK cells that had down regulated EAT-2 (2.8% (n=6) vs. 1.3% (n=5) of lymphocytes, p=0.03) and more CD56^dim NK cells expressing NKG2D (11.2% vs. 2.6% of lymphocytes, p=0.02) and NKp46 (13.6% vs. 3.9% of lymphocytes, p=0.05). Moreover, after imatinib discontinuation the expression of transcription factor Eomes increased in the CD56^dim NK cells of the early relapsing group (baseline MFI 2045 vs. 1 month 3480, p=0.06), while in non-relapsing group it seemed to even decrease (baseline MFI 2273 vs. 1 month 1980, p=0.13) pointing towards an adaptive phenotype.

No significant differences between the groups were observed when degranulation against K562 cell line was studied. However, CD16^neg NK cells from non-relapsing patients responded to K562 stimulation by secreting more TNFα/IFNγ compared to the early relapsing patients (21% vs. 13% of CD56^posCD16^neg NK cells, p=0.01). Furthermore, patients whose CD16^neg NK cells had higher than median TNFα/IFNγ secretion when stimulated with K562 cells showed an increased TFR at 6 months (78%) compared to patients who had lower TNFα/IFNγ secretion than median (37%, p=0.005).

Conclusions: CML patients who successfully discontinued imatinib therapy displayed a higher number and frequency of peripheral blood mature, adaptive-like NK cells capable of secreting cytokines TNFα/IFNγ relative to relapsing patients. How such NK cells may contribute to maintenance of treatment free remission is still unknown. Nonetheless, our results warrant further clinical studies with NK-cell modulating agents.