Introduction and aims: With the rapid development of immunotherapeutic strategies in acute myeloid leukemia (AML), it has become increasingly important to understand the complex interactions between leukemic blasts and the surrounding immune environment. Increasing evidence indicates that immune cell composition plays a critical role in AML pathogenesis, influencing both disease progression and patient response to therapy. This study aims to comprehensively profile the peripheral blood immune landscape in AML patients and explore how the relative abundance of specific immune cell populations correlates with clinical outcome. Furthermore, by integrating these immune profiles with transcriptomic signatures of leukemic blasts, we aimed to elucidate how the leukemic clone adapts to and interacts with the surrounding immune environment to promote its survival and evade immune surveillance.

Methods: Peripheral blood samples were collected from 32 patients with non-acute promyelocytic AML at diagnosis. For cell profiling by digital cytometry, a targeted sequencing library was prepared from whole leukocyte RNA using the SureSelect CD CiberMed Heme + HiRes kit (Agilent). For whole-transcriptome library preparation (NEBNext Ultra II, NEB), ribodepleted RNA (RiboCop, Lexogen) from pure leukemic blasts was used. Both libraries were sequenced on NovaSeq (Illumina). Relative fractions of 12 lymphoid-derived cell types were calculated using the iSORT Fractions CiberMed software, with proportions normalized to the total lymphoid fraction. Gene expression was quantified using StringTie2 (v1.3.6), and analyses were performed in R (v4.0.0), GraphPad Prism 10, and MetaScape.

Results: Using PERMANOVA analysis based on lymphoid cell fractions, patients who achieved complete remission (CR; n=15) after the first cycle of 3+7 chemotherapy and those who did not (n=17) were significantly separated (p=0.014). In responders, the relative fractions of CD8⁺ T cells (p=0.005) and resting NK cells (p=0.019) were significantly higher compared to non-responders. The largest difference was observed for the total lymphoid cytotoxic cell fraction (LCCF; CD8⁺ T cells, resting NK cells, activated NK cells, and γδ T cells), which was 45% on average in responders compared to 31% in non-responders (p<0.001).

Gene set enrichment analysis of 112 differentially expressed genes in leukemic blasts between high and low LCCF (H/L-LCCF) patients (divided by median) highlighted pathways in interferon signaling, cytoskeletal remodeling, and cell–microenvironment interactions, alongside tissue development processes. Several genes associated with proliferation, immune evasion, immune recognition and regulation, microenvironmental remodeling, and survival signaling, including SPARC, LAG3, PRAME, SETBP1, and NFKBIZ – were significantly downregulated in blasts from H-LCCF patients compared to L-LCCF. Conversely, the tumor suppressor gene DLC1, which is involved in cell migration and proliferation, was significantly upregulated in blasts from H-LCCF patients.

Conclusions: H-LCCF was identified as a marker of better therapy response and was associated with a distinct blast transcriptomic profile. Under strong cytotoxic surveillance, leukemic blasts exhibited reduced expression of genes involved in survival and proliferation (e.g., SETBP1), immune evasion and T cell inhibition (e.g., LAG3), T cell recognition (e.g., PRAME), and microenvironmental remodeling (e.g., SPARC), while upregulating tumor suppressors (e.g., DLC1). This suggests that strong immune pressure may drive transcriptional reprogramming and immune-mediated selection, leading to a less aggressive and less immunogenic leukemic phenotype. Conversely, in patients with low immune pressure, the overexpression of immune-inhibitory genes such as LAG3 may contribute to the suppression of cytotoxic T cells and the establishment of an immunosuppressive microenvironment, resulting in reduced LCCF.

Our results highlight the interplay between leukemic blasts and the peripheral immune landscape, revealing candidate mechanisms of immune modulation and evasion in AML. Furthermore, we identified potential biomarkers of leukemic immunogenicity that may be relevant to the optimization of immunotherapeutic approaches.

Funding: This project was supported by MH CZ-DRO (UHKT 00023736).

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