This multicenter prospective study aimed to evaluate the immunophenotypic and cytokine profiles of newly diagnosed chronic myeloid leukemia (CML) patients receiving dasatinib as frontline therapy.

A total of 28 treatment-naïve adult patients with chronic-phase CML were enrolled across five medical centers in Taiwan. The peripheral blood was collected serially at predefined six timepoints (V1 to V6), including baseline (before dasatinib treatment, V1) and five subsequent visits (V2 to V6). The duration between each visit was 28±3 days. The peripheral blood from healthy individuals (n=16) was also collected for baseline comparison. Flow cytometry profiled NK (CD57, CD62L, CD244, NKp30/44/46, CD226, NKG2A/C/D) and T (CD4, CD8) cell markers, and eleven cytokines (IFN-γ, IL-10, IL-12p40, IL-12p70, IL-15, IL-18, IL-27, MCP-1, MIP-1α, MIP-1β, TNF-α) were measured by MILLIPLEX® multiplex assay. The patients who achieved major molecular response or better by six months post-treatment were classified as Group A (n=12), with all remaining patients classified as Group B (n=16).

The baseline immunophenotypes and cytokines were analyzed in CML patients and healthy individuals. The results revealed CML patients exhibited impaired NK cytotoxicity vs healthy individuals, with reduced activating receptors (NKp30, NKp44, CD244, CD226) and increased inhibitory NKG2A. Group A showed a higher percentage of NKp44+ and CD226+ NK cells than Group B. Cytokine profiling revealed that CML patients had lower IFN-γ but elevated IL-10, IL-18, IL-27, MIP-1α, and MIP-1β. Compared to Group B, Group A had lower baseline IL-10 and IL-18 and higher MCP-1, with their IL-10 levels approximating those of healthy individuals.

The serial phenotypic investigation after dasatinib treatment demonstrated that, among all CML patients, the percentages of CD57+CD62L+, NKp30+, CD226+, and NKG2C+ NK cells were increased after dasatinib administration. The phenotypic change between Groups A and B was further examined, revealing opposite NKG2A+ expression trends: decreasing in Group A and increasing in Group B after dasatinib treatment. Phenotype ratio changes (LS-Means), calculated as differences between baseline and post-treatment values, were compared between Groups A and B, revealing significant differences in several phenotypes, including CD244 and NKp44. Regarding cytokine analysis, several cytokines, including IFN-γ, IL-10, IL-12p40, IL-15, IL-18, IL-27, MIP-1β, and TNF-α, showed a significant decreasing trend after dasatinib administration. Additionally, Group B exhibited a greater magnitude of decrease in IL-10 and IL-18 levels than Group A. Moreover, the serum level of MCP-1 decreased after dasatinib treatment in Group A, whereas an increase was observed in Group B. There was no significant difference observed in T cell phenotype analysis.

To develop a predictive model, the five markers (CD244, IL-10, IL-18, MCP-1, and NKp44) that showed distinct changes following dasatinib treatment were selected. These markers were converted to binary variables by assigning each a score of 0 or 1, and the combined composite scores (0–5) were calculated for each patient. The result revealed that Group A patients were more likely to exhibit higher composite scores. Specifically, 75.0% of Group A had composite scores ≥ 3, compared to 18.8% in Group B (P = 0.0061). The stratified logistic regression of pooled marker categories demonstrated strong predictive ability for identifying good responders (Group A), with an odds ratio of 13.00 (95% CI: 2.12–79.59; P = 0.0055) and a C-index of 0.781. Further analysis based on age and gender showed better predictive ability among patients aged < 65 years or males.

In summary, CML patients exhibited compromised anti-cancer cytotoxic activity in NK cells. The patients who achieved faster and deeper molecular responses may partly preserve NK cell function or be less immunosuppressive. Further analyses showed significant changes in cytokine and immunophenotype after dasatinib treatment, indicating robust modulation in inflammation or immune signaling pathways. Finally, a model was established with stronger predictive power to identify the subgroup with faster and deeper molecular response.

This content is only available as a PDF.
2025
Sign in via your Institution