A total of 60 patients with AML (38 men and 22 women; age, 62.2+/-5.7 years) and 60 healthy controls (38 men and 22 women; age, 62.1+/-5.8 years) who were admitted to the Xing'an League People's Hospital (Ulanhot, China) between May 2017 and May 2020, were enrolled in the present study. All healthy controls showed normal physiological parameters in systemic physiological examination, including bone marrow blasts <4%, and had no history of AML. All patients were diagnosed using a bone marrow test. All enrolled patients had a percentage of bone marrow blasts >20% (with a mean of 21.5%) and no previous history of AML or other malignancies. Patients with other clinical disorders (such as metabolic disorders, chronic diseases and severe infection), and who had initiated therapy for such disorders within 3 months prior to admission, were excluded. The present study was approved by the Ethics Committee of Xing'an League People's Hospital (approval no. 323SE), and all patients and control subjects provided written informed consent.

Bone marrow was collected from all patients and healthy controls by biopsy, and used to isolate BMMNCs using lymphocyte separation medium (TBD Science; Tian Jin Hao Yang Biological Manufacture Co., Ltd.). The isolation procedure was conducted following the manufacturer's instructions. Briefly, 2 ml bone marrow mixed with medium was centrifuged at 400 × g for 15 min. The second layer of the supernatant (containing the lymphocytes) was then used for cell culture. The Beckman MoFlo Astrios high-performance, live-cell sorting system (Beckman Coulter, Inc.) was used to isolate BMMNC subpopulations, as outlined in a previous study (17). PE-conjugated mouse anti-human CXCR4 (cat. no. 60089PE.1; Stemcell Technologies, Inc.) and APC-conjugated rat anti-human CD45 (cat. no. 28145-1; Signalway Antibody LLC) were used to isolate the BMMNCs, per the sorting identification strategy of a previous study (17). The cells were resuspended in minimal essential medium with Earle's salts (Gibco; Thermo Fisher Scientific, Inc.) containing 10% fetal calf serum and 1% antibiotics (penicillin/streptomycin; Gibco; Thermo Fisher Scientific, Inc.); 1×10⁷ resuspended bone marrow cells were seeded into 100-mm cell culture dishes and incubated at 37°C in a humidified incubator with 5% CO₂. Kasumi-3 and Kasumi-6 AML cell lines were purchased from the ATCC, and cultured in RPMI-1640 medium (10% FBS) at 37°C in an incubator with 5% CO₂ and 95% humidity. Cells used in the subsequent assays were collected at ~85% confluence.

Overexpression of circ-ATAD1 and miR-34b was achieved by transfecting 1×10⁶ Kasumi-3 or Kasumi-6 cells with a circ-ATAD1 expression vector (1 µg; pcDNA3.1 backbone vector; Invitrogen; Thermo Fisher Scientific, Inc.) or miR-34b mimics (10 nM; Sigma-Aldrich; Merck KGaA) using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) (1:2). Untransfected and empty vector- or NC miRNA-(Sigma-Aldrich; Merck KGaA) transfected cells were used as the normal control and negative control (NC) cells, respectively. The sequences of the miR-34b mimics and miR-NC are displayed in the Table I. After incubation with the transfection mixture for 6 h at 37°C, cells were immediately washed with fresh medium, followed by culture in fresh medium for another 48 h prior to use.

Total RNA was isolated from BMMNCs and in vitro cultured cells using RNAzol reagent (Sigma-Aldrich; Merck KGaA), and treated with DNA eraser (Takara Bio, Inc.) at 37°C until the OD260 nm/280 nm ratio had reached ~2.0 (pure RNA). The integrity of all RNA samples was determined by electrophoresis using a 5% urea-PAGE gel. Total RNA was used for first-strand cDNA synthesis with the Prime Script RT reagent kit (Takara Bio, Inc.) per the manufacturer's instructions. Circ-ATAD1 expression was determined using the SYBR^® Green Quantitative RT-qPCR Kit (Sigma-Aldrich; Merck KGaA) with GAPDH as the internal control. Mature miR-34b expression was analyzed using the All-in-One™ miRNA qRT-PCR Detection Kit (GeneCopoeia, Inc.) with U6 as the internal control. All operations were completed following the manufacturers' instructions. The qPCR thermocycling conditions for all genes, cirRNA and miRNAs were as follows: 95°C for 1 min, followed by 40 cycles of 95°C for 10 sec and 60°C for 45 sec. PCR reactions were performed using the CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.). Ct values of the targeted genes were normalized to the corresponding internal controls based on the 2^−ΔΔCq method, which was used to quantify gene expression (18). The qPCR primers are listed in Table II.

Both the nuclear and cytoplasmic fractions of Kasumi-3 and Kasumi-6 cells were prepared using the Nuclei Isolation Kit: Nuclei EZ Prep (Sigma-Aldrich; Merck KGaA) according to the manufacturer's instructions, and used for RNA isolation and RT-qPCR to detect circ-ATAD1, with GAPDH as the internal control.

IF staining with anti-histone H3 (1:300; cat. no. ab6002; Abcam) was in reference to a previous study (19). The Blocking buffer was PBS with 1% BSA (Thermo Fisher Scientific, Inc.). The Alexa fluor 555 anti-rabbit antibody (cat. no. ab150078; Abcam) was used as the secondary antibody (1:500). FISH was performed using Kasumi-3 and Kasumi-6 cells as described previously (19). The probe of cir-ATAD1 (TACCACAGCCTGGAGGCCCATAG) was synthesized and labeled with digoxigenin (DIG-dUTP) by Sangon Biotech (Shanghai) Co., Ltd. Specifically, the slices covered by cells were fixed in 4% paraformaldehyde (MilliporeSigma) for 10 min at room temperature. Before pre-hybridization, cells were permeabilized with cold 0.1% Triton X-100 and pre-hybridized with a hybridization buffer at 37°C for 1 h. The slides were incubated with a hybridization buffer containing the FISH probe at 95°C for 5 min, and then at 37°C overnight in the dark in a humid chamber. The samples were washed with 2 × saline sodium citrate buffer (SSC) for 10 min at 37°C, 1 × SSC for 2 × 5 min at 37°C, and 0.5 × SSC for 10 min at room temperature. The slides were then incubated with anti-DIG-488 (1:300; cat. no. ab150077; Abcam) at 37°C for 50 min, and the nuclei were counterstained with DAPI at room temperature for 30 min. Finally, the slices were sealed in fluorescence decay-resistant medium and images were obtained under a fluorescence microscope (Nikon Corporation).

Genomic DNA isolation from transfected Kasumi-3 and Kasumi-6 cells was performed using a routine method (20). All genomic DNA samples were processed using the EZ DNA Methylation-Gold™ Kit (Zymo Research Corp.) per the manufacturer's protocol. Then, qPCR and routine PCR were performed to detect methylation of the miR-34b gene promoter using PCR Master Mix ×2 (Invitrogen; Thermo Fisher Scientific, Inc.). The MSP primers can be used to amplify methylated template, while the primers for unmethylation-specific PCR (USP) do not amplify these products. Both the MSP and USP conditions were as following: 95°C for 5 min, followed by 40 cycles of 95°C for 30 sec, 55°C for 30 sec and 72°C for 50 sec, and then 72°C for 10 min. All primers used for MSP and USP are listed in Table III.

A total of 3×10³ transfected Kasumi-3 and Kasumi-6 cells were transferred to each well of a 96-well plate in 0.1 ml medium, and cultured at 37°C for 48 h before the addition of BrdU. The experiment was conducted using the BrdU Cell Proliferation Assay (cat. no. QIA58; Sigma-Aldrich; Merck KGaA) according to the manufacturer's protocol. Then, cells were cultured with 20 µl/well diluted BrdU reagent (10 mM) for 6 h, fixed with Fixing solution and incubated for 30 min. After fixation, the cells were incubated for another 48 h with peroxidase-coupled anti-BrdU-antibody (supplied by the kit), followed by washing twice with ice-cold PBS. After incubation with peroxidase substrate for 3 h, OD values were measured at 450 nm. For the represent images, the anti-Brdu (1:500; cat. no. ab6362; Abcam) was used as primary antibody, and Goat anti-rabbit Alexa Fluor^® 546 (1:2,000; cat. no. A11010; Invitrogen; Therno Fisher4 Scientific, Inc.) was used as the secondary antibody. The detailed method has been previously published (21).

AML and control groups were compared by unpaired t-test. Comparisons among multiple independent cell transfection groups were analyzed by one-way ANOVA followed by Tukey's test. Correlations were analyzed by Pearson's correlation coefficient. Data are presented as the mean ± standard deviation of three independent replicates. P<0.05 was considered to indicate a statistically significant difference.

Samples of BMMNCs from both patients with AML (n=60) and healthy controls (n=60) were subjected to RNA isolation and RT-qPCR to determine the differential expression of circ-ATAD1 and miR-34b. Compared with the controls, circ-ATAD1 was highly expressed in AML (Fig. 1A, P<0.01), while miR-34b expression was lower in AML (Fig. 1B, P<0.01), suggesting that circ-ATAD1 upregulation and miR-34a downregulation may be involved in AML. To study the crosstalk between circ-ATAD1 and miR-34b, their correlations across both AML and control samples were determined using Pearson's correlation analysis. The data showed that circ-ATAD1 and miR-34b were closely and inversely correlated across AML samples (Fig. 1C; P<0.0001), but not across the control samples (Fig. 1D). Therefore, circ-ATAD1 and miR-34b may interact with each other in AML.

To further study the crosstalk between circ-ATAD1 and miR-34b, Kasumi-3 and Kasumi-6 cells were transfected with either a circ-ATAD1 overexpression vector or miR-34b mimics, followed by expression confirmation every 24 h until 96 h. It was observed that both circ-ATAD1 and miR-34b were overexpressed between 24 and 96 h post transfection (Fig. 2A and B, P<0.05). In addition, circ-ATAD1 overexpression decreased that of miR-34b (Fig. 2C, P<0.05). By contrast, miR-34b overexpression failed to significantly alter circ-ATAD1 expression (Fig. 2D). Therefore, circ-ATAD1 may downregulate miR-34b expression in AML cells.

A subcellular fractionation assay was used to determine the subcellular location of circ-ATAD1 in both Kasumi-3 and Kasumi-6 cells, and circ-ATAD1 was only detected in the nuclear fractions (Fig. 3A). The relative expression of circ-ATAD1 and miR-34b in Kasumi-3 cells (nuclei), in Kasumi-6 cells (nuclei), in the control and that in the AML patient samples is also provided. The expression levels of circ-ATAD1 were highest in AML samples (Fig. 3B; P<0.01), and those of miR-34b were highest in the healthy control samples (Fig. 3C; P<0.01). FISH (Figs. S1 and S2) also showed similar results. In addition, the effects of circ-ATAD1 overexpression on miR-34b gene methylation were analyzed by MSP. Compared with cells transfected with empty pcDNA3.1 vector, cells transfected with the circ-ATAD1 expression vector showed increased methylation of miR34b (Fig. 3D). These results indicate the significantly different expression of these two molecules in healthy and tumor cells. In addition, the expression of miR-34b was regulated by circ-ATAD1 through methylation-related mechanisms in Kasumi-3 and Kasumi-6 cells.

The role of circ-ATAD1 and miR-34b in regulating Kasumi-3 (Fig. 4A and C) and Kasumi-6 (Fig. 4B and D) cell proliferation was analyzed using a BrdU assay. Circ-ATAD1 overexpression increased cellular proliferation, while miR-34b overexpression decreased cellular proliferation (both (P<0.05). Moreover, co-transfection analysis showed that the effect of circ-ATAD1 overexpression on AML cell proliferation was reduced by miR-34b overexpression (P<0.05). The results indicated Circ-ATAD1 overexpression promotes the proliferation of AML cells through miR-34b. The Circ-ATAD1-miR-34b cell signaling axis, which contributed the tumor cell proliferation, has been revealed. However, methylation-related mechanisms were the major regulation mechanism between Circ-ATAD1 and miR-34b.