Background

Acute myeloid leukemia (AML) has been recognized as a heterogeneous hematological malignancy and the most common acute leukemia in adults, which is mainly caused by multifaceted somatic mutations in myeloid differentiation and monoclonal proliferation of immature progenitors [1, 2]. Despite the approval and application of novel drugs for patients, AML still remains as a major field of unmet medical need among diverse hematologic malignancies attributes to the deficiency of details information upon the cytopathologic and pathogenic features [3]. For decades, considerable progresses have been achieved in elucidating the pathogenesis of AML both at the cellular and molecular levels [4]. For instance, diverse inherited genetic loci and novel biomarkers (e.g., epigenetic, genetic, and protein) involved in myelodysplastic syndrome (MDS) and AML development in patients have been identified for the risk stratification and treatment assessment of AML patients on the basis of epigenetic and omics profiles [3, 5]. According to the World Health Organization (WHO) Classification of leukemia and myeloid neoplasms, AML can be divided into eight categories according to the recurrent genetic abnormalities and into three categories according to the indicated gene mutations. However, the genes with sufficient accuracy for unraveling the risk-stratification schemes and clinically targetable treatment decision-making of AML are still largely obscure [6].

Natural killer (NK) cells are critical innate lymphoid cells (ILCs) for their effect in tumor immunosurveillance and antiviral immunity dispense with presensitization, which have been involved in both innate and adoptive immune responses via the cytotoxic and cytokine-secreting approaches [7,8,9]. NK cells have been considered as heterogeneous populations generated from hematopoietic stem cells (HSCs) in the bone marrow environment (BME) with limited functional and phenotypic diversity, which are composed of two distinct subsets in human, including the cytotoxic CD3−CD56dimCD16high and the IFN-γ-producing CD3−CD56brightCD16+ counterparts [10, 11]. To date, adoptive NK cells and the concomitant chimeric antigen receptor-transduced NK cells (CAR-NKs) have been explored in various cancer settings including AML, and aiming to attain better therapeutic outcomes [12, 13]. For example, Terrén et al. and Albinger et al. reported the application of cytokine-induced memory-like (CIML) NK cells and primary CD33-targeting CAR-NK cells for the treatment of patients with AML, respectively [14, 15]. Furthermore, Soldierer et al. and Xu et al. highlighted the genetic engineering of human cells for the CAR-enhanced immunotherapy of hematological malignancies including AML [16, 17]. Very recently, Gauthier and the colleagues reported the successful control of AML by a trifunctional NKp46−CD16a−NK cell engager (NKCE) with prolonged pharmacodynamic effects and very low inflammatory cytokine induction against CD123 antibody-dependent cell cytotoxicity (ADCC) [18]. Of note, Crinier et al. showed the trajectories of NK cell differentiation from the resident CD56bright NK cells to the CD56dim NK1-like NK cells and the CD56bright NK2-like NK cells, yet the major cytophenotypic and omics features of resident and expanded NK cells in the bone marrow of AML patients (AML-NKs) are largely obscure.

For the purpose, in this study, we isolated resident NK cells (rNKs) from both AML patients (rAML-AMLs) and healthy donors (rHD-AMLs), and conducted ex vivo NK cell expansion and activation from the corresponding rNKs after a 14-day’s induction (eAML-AMLs, eHD-AMLs). With the aid of multifaceted cellular and molecular analyses, we verified the cytophernotic (e.g., cytomorphology, immunophenoty**, cellular vitality, and cytotoxicity against diverse tumor cell lines) and transcriptomic characteristics (e.g., gene expression profiling, genetic variations) between AML-NKs and HD-NKs in the bone marrow microenvironment, which collectively indicated the pathogenic effect of NK cells in patients with AML and would benefit the further development of novel biomarkers for clinical diagnosis and NK cell-based cytotherapy in future.

Methods

Preparation of mononuclear cells (MNCs) from bone marrow

Human bone marrow was obtained with the consent of AML patients and healthy donors (HD) and the approval of the Ethics Committee of Gansu Provincial Hospital and the guideline of Helsinki Declaration (2023-120). For MNCs isolation from bone marrow of AML patients (AML-MNCs) and HD (HD-MNCs), the Ficoll-based density gradient centrifugation were conducted as we described recently with several modifications [40]. The further exploration of the alterations in AML-NK cells would benefit our understanding upon pathogenesis and therapeutic regimens of AML from the aspect of numerical, receptor expression, and maturation defects of NK cells, along with checkpoint overexpression inhibitors and epigenetic modifications [40]. Of note, it’s of great interesting to further explore the potential variations of NK cells in bone marrow environment of AML patients before treatment and remission, which will help illuminate the biofunction and omics features of AML and benefit the development of targeted therapy in clinical practice.

State-of-the-art literatures have highlighted the involvement of dysimmunity in the pathogenesis and drug resistance of AML patients. For instance, Zhang et al. verified the multiple chemoresistant properties in LSCs and the oxidative phosphorylation (OXPHOS) signatures in pediatric AML [41]. Very recently, Naldini et al. reported the longitudinal single-cell profiling and functional validation of chemotherapy response in AML patients, and identified the variations in the stemness and quiescence signatures of the OxPhoslow miR-126high LSCs [42]. Of note, Abbas et al. outlined the T cell receptor repertoire profiling and the T cell landscape in AML patients with PD-1 blockade therapy [43]. Instead, we conducted systematic and detailed characterization of the multifaceted signatures of NK cells and the variations in subsets and gene expression pattern between resident and expanded AML-NKs and HD-NKs. To our knowledge, there were minimal reports upon the specific subsets (e.g., CD3−CD56+CD25+) of AML-NKs in AML patients. Similarly, Stringaris and the colleagues verified the upregulation of NKG2A (a inhibitory receptor) and the downregulation of NKp46 in NK cells of AML patients, which indicated the long-lasting changes and effector function impairment in AML-NK cells induced by AML blasts [44]. Interestingly, Crinier and the colleagues revealed the trajectories and the stress signature of NK cell differentiation in bone marrow, which further indicated the involvement of AML-NKs in the pathogenesis of AML [45]. Therefore, it is of great interesting to further dissect the single-cell mutation and gene expression profiling of clonal evolution as well as cellular hierarchy in AML patients, which will collectively benefit the development of accurate diagnosis and targeted therapy [46, 47].

Conclusion

Overall, in this study, we systematically and meticulously dissected the biological phenotypes, cellular vitality and cytotoxicity of both resident and expanded NK cells between AML-NKs and HD-NKs from the landscape of gene expression pattern and somatic variations. Our findings put forward the potential pathogenic role of AML-NK cells in bone marrow microenvironment (BME) in AML patients, which could be conducive to the further exploration of the pathogenesis of AML and the development of NK cell-based targeted therapy in future.