Low-Density Granulocytes Are Decreased in Acute Myeloid Leukemia and in Myelodysplastic Syndromes with Negative Prognostic Factors

Introduction. Myelodysplastic syndromes (MDS), a group of clonal hematological diseases, are characterized by ineffective hematopoiesis, progressive peripheral blood (PB) cytopenia(s), and increased risk of developing acute myeloid leukemia (AML). Classification and risk stratification are constantly under revision for a better estimation of prognosis in those patients. Investigation of immune biomarkers is needed, because immune dysregulation also plays an important role in dysplastic hemopoiesis and immunological escape of neoplastic clones. Here, we studied frequency of low-density granulocytes (LDGs), a neutrophil subset with immunoregulatory functions, in MDS and AML at diagnosis and during treatments.

Methods. A total of 17 patients (M/F, 14/12; median age, 69 years old; range, 21-84 years) and seven healthy subjects were enrolled at the Hematology and Transplant Center, University Hospital "San Giovanni di Dio e Ruggi d'Aragona", Salerno, Italy, between October 2020 and July 2021. Patients were diagnosed with AML (N = 7), or MDS (N = 10) according to the 2016 World Health Organization criteria. For immunophenotyping, fresh EDTA whole PB was stained with the ollowing antibodies: CD45; HLA-DR; CD15; CD3; CD56; CD19; CD11b; CD33; CD34; CD14; and CD16 (all from Beckman Coulter, Brea, CA). Acquisition was carried out using a Navios EX flow cytometer, and Navios software v1.3 (Beckman Coulter). Post-acquisition compensation and analysis were performed using FlowJo software (v.10.7.1, Becton Dickinson). LDGs were identified as CD3-CD56-CD19-CD11b+CD33+CD14-CD15+ cells, following previously published gating strategies (Rahman S, et al. Ann Rheum Dis. 2019). Data were analyzed using Prism (GraphPad software, La Jolla, CA). A P < 0.05 was considered statistically significant.

Results. Frequencies of circulating LDGs were significantly reduced in AML patients at diagnosis compared to controls (P = 0.0018) and MDS (P = 0.0077) and were slightly decreased compared to AML in complete remission (P = 0.1605). MDS patients were then divided based on Revised International Prognostic Scoring System (IPSS-R), and very-low and low-risk MDS patients displayed significantly higher circulating LDG frequencies compared to AML at diagnosis (P = 0.0083), while no differences were described between AML at baseline and intermediate-risk MDS (P = 0.1103). Subsequently, LDGs were correlated with clinical and phenotypic features by correlation analysis showing significant negative correlations between LDGs and blasts identified by flow cytometry (r = -0.5463; P = 0.0057) but not by cytology (P = 0.1346), between LDGs and lymphocytes (r = -0.4407; P = 0.0311) or flow cytometric normalized blast count (NBC; r = -0.5283; P = 0.0096) as previously defined (Giudice V, et al. Biomedicines. 2021). A slight negative correlation was described between LDGs and WT1 expression levels (r = -0.5369; P = 0.0719), particularly evident in MDS patients (r = -0.9980; P = 0.0402), supporting our previous findings of negative prognostic impact of WT1 expression in MDS and AML. Finally, we investigated CD16 expression on LDGs, because CD16 is essential for neutrophil degranulation. Despite no differences were described between percentage of LDG subsets among patients' groups, various correlations were identified by Pearson analysis. In particular, CD16+ LDGs negatively correlated with blasts (P = 0.0229), while positively correlated with lymphocytes (P = 0.0404) detected by flow cytometry. Conversely, CD16int and CD16- LDGs negatively correlated with lymphocytes (P = 0.0109 and P = 0.0021, respectively) and positively correlated with granulocytes identified by flow cytometry (P = 0.0024 and P = 0.0008, respectively). In addition, CD16int LDGs negatively correlated with blasts detected by flow cytometry (r = -0.65; P = 0.0414).

Conclusions. Our preliminary results suggested a possible role of LDGs in prognostic definition of AML and MDS patients especially when combined with other biomarkers, such as WT1 expression levels or NBC. Moreover, our data supported the hypothesis of biological heterogeneity of granulocytes, as LDG subsets variously correlated with lymphocytes and leukemic cells suggesting different roles in suppression or activation of immune responses. However, our findings need further validation in larger cohorts and in in vitro studies.