Bone marrow samples were extracted from 152 adult patients with MDS and 29 patients with AML. All patients were diagnosed according to the 2008 World Health Organization (WHO) criteria (12). Of the patients with MDS, 92 were men and 60 were women, with a median age of 56 years (range, 18-86 years). A total of 84 patients had refractory cytopenia with multilineage dysplasia (RCMD), 8 had RCMD with ring sideroblasts (RCMD-RS), 17 had refractory anemia with an excess of blasts type 1 (RAEB-1), 29 had RAEB-2, five had refractory anemia (RA), two had refractory anemia with ring sideroblasts (RARS), six had an unclassifiable form of MDS (MDS-U) and one had 5q-syndrome. Of the patients with AML, 16 were diagnosed with acute myelomonocytic leukemia, six with acute monocytic leukemia (AML-M5) and seven with AML with mixed-lineage leukemia rearrangements. The prognostic score for each patient with MDS was calculated using the International Prognostic Scoring System (IPSS) (13). All subjects with MDS were recruited on January 1, 2008 by the Sino-U.S. Shanghai Leukemia Cooperative Group (School of Public Health, Fudan University, Shanghai, China), and were followed up until they succumbed to the disease or until the last follow-up date (December 30, 2015). Two subjects did not undergo chromosomal examination due to examination failure. Finally, 69 subjects succumbed and 14 subjects progressed to AML. In addition, bone marrow samples from 40 subjects without hematological malignancies were obtained between December 30, 2015 and January 1, 2018, and were analyzed as controls. The control subjects were diagnosed as having immune thrombocytopenia by bone marrow smears and chromosomal examinations. Subsequently, a 2-year follow-up period ensured that these controls did not develop MDS or other clonal disorders. No statistical difference was found with regards to age and sex among the subjects with MDS and AML and the controls. The present study was approved by the Ethics Committee of Huashan Hospital, Fudan University (approval no. KY2015-269; Shanghai, China), and informed consent was obtained from each participant.

The mononuclear cells were harvested by bone marrow aspiration and were isolated by centrifugation (600 × g for 20 min at room temperature) in lymphocyte separation medium (Ficoll^® Paque Plus; GE Healthcare, Chicago, IL, USA). Total RNA was isolated from bone marrow mononuclear cells using the QIAamp RNA Blood Mini kit (Qiagen GmbH, Hilden, Germany). TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used to isolate total RNA from the cultured leukemia cells after culture medium was removed and the cells were washed with pre-cooled PBS (4°C) three times. The purity of RNA was assessed by measuring the optical density (OD) 260/280 ratio using a NanoDrop spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc., Wilmington, DE, USA). The RNA samples were then reverse transcribed into cDNA using Takara PrimeScript RT Master Mix (Takara Bio, Inc., Otsu, Japan). The 10-µl reaction mixture comprised 2 µl 5X PrimeScript RT Master Mix, 500 ng RNA and RNase-Free distilled H₂O, which was used to ensure the total reached 10 µl. Subsequently, the mixture was amplified at 37°C for 15 min, 85°C for 5 sec and was stored at 4°C for further use.

For RT-qPCR, cDNA samples were amplified using an ABI Prism 7500 Sequence Detection system (Applied Biosystems; Thermo Fisher Scientific, Inc.) with Takara SYBR Premix Ex Taq PCR reagents (Takara Bio, Inc.). A 20-µl reaction mixture containing 10 µl SYBR-Green premix, 0.4 µl each primer (10 µM), 0.4 µl ROX Reference Dye II, 2 µl cDNA and 6.8 µl DEPC-treated water was amplified at 95°C for 30 sec, followed by 40 cycles at 95°C for 5 sec and 60°C for 34 sec. The PCR primer sequences for ABAT were as follows: Forward, 5′-TTTTCTGTCTCCTCCACCTGTC-3′ and reverse, 5′-CTGGCT GGGTTCATCCTAAG-3′, which resulted in amplification of a 113-bp PCR band. The GAPDH pr i mer sequences were as follows: For wa rd, 5′-TGGATGAAGTTGGTGGTGAG-3′ and reverse, 5′-CAGCATCAGGAGTGGACAGA-3′, which resulted in amplification of an 87-bp PCR band. The relative abundance of ABAT mRNA was normalized to the levels of GAPDH mRNA using the 2^−ΔΔCq (fold change) method (14).

Total DNA was isolated from bone marrow mononuclear cells using QIAamp DNA Blood Mini kit (Qiagen GmbH), according to the manufacturer’s protocol. DNA quality and purity were assessed using the aforementioned methods, and DNA samples were stored at −20°C for further use.

For MS-HRM, DNA samples from all patients, and human methylated and nonmethylated DNA standards (Zymo Research Corp., Irvine, CA, USA) were modified using the EZ DNA Methylation-Gold kit (Zymo Research Corp.), according to the manufacturer’s protocol. In addition, the human methylated and nonmethylated DNA standards were mixed at a ratio of 1:9, 1:3, 1:1 and 3:1 to perform DNA methylation modification. All modified DNA samples were amplified using the C1000 Touch Thermal Cycler (Bio-Rad Laboratories, Inc.) with Takara Premix Taq Hot Start Version (Takara Bio, Inc.). A 25-µl reaction mixture containing 12.5 µl Premix Taq Hot Start Version, 1 µl each primer (10 µM), 1 µl modified DNA and 9.5 µl DEPC-treated water was amplified at 95°C for 2 min, followed by 40 cycles at 95°C for 30 sec, 59°C for 30 sec and 72°C for 30 sec, and an extension step at 72°C for 5 min. Subsequently, the amplified product was added to 0.9 µl SYTO 9 Green Fluorescent Nucleic Acid Stain (Invitrogen; Thermo Fisher Scientific, Inc.) and reacted at 94°C for 60 sec and 40°C for 60 sec. Data were collected between 75°C and 85°C using Rotor Gene 6000 (Qiagen GmbH). The methylation primers for ABAT were as follows: Forward, 5′-GTAAGAAGGGTTGGTAGGGTTTT-3′ and reverse, 5′-ACCATTTACACCCTCAAAACTACA-3′, which resulted in amplification of a 183-bp PCR band.

In the present study, SKM-1 and THP-1 leukemia cell lines were used. The SKM-1 cell line is derived from a patient with leukemia transformed from MDS; therefore, this cell line has similar characteristics to MDS. The THP-1 cell line is derived from a patient with AML-M5. Mutations in the ABAT gene were detected in both cell lines used in this study. The ABAT gene mutations detected in the SKM-1 cell line were as follows: 309C>T, V103V; 984C>A, V328V and 167G>A, Q56R. The ABAT gene mutations in the THP-1 cell line were: 984C>A, V328V and 167G>A, Q56R. The mutation sites have no influence on enzyme activity; therefore, these two cell lines were used in the present functional study. The human myeloid leukemia cell line THP-1 was obtained from the Chinese Academy of Sciences (Beijing, China), and the SKM-1 cell line was purchased from Health Science Research Resources Bank (Sennan, Japan). All cells were cultured in RPMI-1640 (HyClone; GE Healthcare Life Sciences, Logan, UT, USA) supplemented with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) at 37°C in an atmosphere containing 5% CO₂.

shRNA sequences that targeted the 3′ untranslated region of ABAT were designed and cloned into pGMLV-SC5 lentiviral vectors (Genomeditech, Shanghai, China). The oligonucleotide sequences were: Sh1-ABAT, 5′-GCGGGAGGACCTGCTAAATAA-3′; Sh2-ABAT, 5′-GCTGGAGACGTGCATGATTAA-3′; and Sh3-ABAT, 5′-GGTGACAAATCCATTCGTTTC-3′. SKM-1 and THP-1 cells (1×10⁵) were infected with three independent shRNA lentiviruses and a scramble lentivirus (negative control, NC; Genomeditech) at a multiplicity of infection of 80 at 37°C; cells without lentiviral infection were considered normal control cells. Infection efficiency was evaluated by flow cytometry, according to the percentage of green fluorescent protein (GFP)-positive cells, and RT-qPCR and western blotting after 72 h. Puromycin (cat. no. P8230; Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) was used to improve the percentage of GFP-positive cells at 72 h.

For drug treatment, the two cell lines were cultured in complete medium supplemented with 0, 100, 200 and 400 µmol/l (±)-vigabatrin (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), which is a specific inhibitor of the GABAT enzyme; the medium was replaced by new complete medium containing (±)-vigabatrin every 24 h. After 10 days, the cells were harvested for further use (15).

GABAT enzymatic activity was detected using the GABAT assay kit (cat. no. E-134; Biomedical Research Service Center, University at Buffalo, Buffalo, NY, USA), according to a previous report (15). Briefly, this assay is based on sequential GABA transamination reaction and glutamate dehydrogenase reaction, which couples the reduction of iodonitrotetrazolium to iodonitrotetrazolium-formazan (ε = 18 mM⁻¹cm⁻¹ at 492 nm). Reactions were terminated by adding 3% acetic acid (cat. no. A6283; Sigma-Aldrich; Merck KGaA) and OD was measured at 492 nm using a plate reader (Thermo Fisher Scientific, Inc.). A mean of the readings was obtained, and control well readings were subtracted from sample well readings (ΔOD). GABAT activity was calculated using the following formula: GABAT activity [µmol/(l•min)] = (ΔOD × 1,000 × 150 µl)/(60 min × 0.6 cm × 18 × 5 µl) = ΔOD × 46.3, where 150 µl is the total reaction volume, 0.6 cm is the light path in the 96-well plate, and 5 µl is the volume of the sample in each well.

Cell viability was measured using the Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Inc., Rockville, MD, USA) according to the manufacturer’s protocol. Briefly, cells treated with the GABAT inhibitor, cells infected with shRNA lentiviruses, NC cells and normal control cells were seeded in a 96-well plate at 5×10³ cells/well. The plates were then incubated for 24, 48 and 72 h at 37°C in a humidified incubator containing 5% CO₂. Subsequently, 10 µl CCK-8 reagent was added to each well. After 4 h at 37°C, absorbance was measured at 450 nm. Three independent experiments were performed.

The two cell lines (1×10⁵ cells) infected with lentiviral vectors or treated with a GABAT inhibitor were harvested and washed with PBS. Cell apoptosis was assessed using Phycoerythrin (PE) Annexin V Apoptosis Detection kit I (cat. no. 559763) and Fluorescein Isothiocyanate (FITC) Annexin V Apoptosis Detection kit I (cat. no. 556547) (both from BD Biosciences, Franklin Lakes, NJ, USA), respectively. Briefly, the cells were harvested and the supernatant was discarded; subsequently, cells were suspended in PBS containing 3% bovine serum albumin (BSA; cat. no. A8020; Beijing Solarbio Science & Technology Co., Ltd.) and were adjusted to 1×10⁶ cells/ml. A 100-µl aliquot was labeled with 5 µl PE and 5 µl 7-aminoactinomycin D (7-AAD) for cells transfected with lentiviruses, or with 5 µl FITC and 5 µl propidium iodide (PI) for cells treated with drugs, and cells were incubated at room temperature for 25 min. Flow cytometry was performed immediately after incubation. Data acquisition and analysis were performed using the BD FACSCanto flow cytometer (BD Biosciences) with FCS Express 3.0 software (De Novo Software, Glendale, CA, USA) and FlowJo 7.6 software (FlowJo LLC, Ashland, OR, USA). Normal controls were used to set voltage for flow cytometry. Cells that were Annexin V-positive and 7-amino-actinomycin D (7-AAD)/propidium iodide (PI)-negative were the early apoptotic fraction, whereas cells that were double positive were the late apoptotic fraction.

For cell cycle analysis, the aforementioned leukemia cells (1×10⁵ cells) were harvested, washed twice with ice-cold PBS, and fixed with 75% ethanol at −20°C overnight. Prior to analysis, the fixed cells were washed twice with ice-cold PBS and suspended in 300 µl PI for 15 min in the dark at 4°C. The data were acquired using the BD FACSCanto flow cytometer with FCS Express 3.0 software. Data were analyzed using ModFit LT software (version 3.2; Verity Software House, Inc., Topsham, ME, USA).

For western blot analysis, total protein was extracted from the cultured cells using protein lysis buffer (cat. no. 89900; Thermo Fisher Scientific, Inc.) and was quantified using a bicinchoninic acid assay kit (Beyotime Biotechnology, Haimen, China), with BSA (Beijing Solarbio Science & Technology Co., Ltd.) as a standard. Proteins (20 µg) were fractionated by 10% SDS-PAGE and were transferred onto polyvinylidene fluoride membranes (Bio-Rad Laboratories, Inc.). The membranes were blocked in 5% non-fat dry milk at room temperature for 2 h and were immunostained overnight at 4°C using antibodies against ABAT (cat. no. ab216465, 1:1,000; Abcam, Cambridge, MA, USA), cyclin D1 (cat. no. 2978S, 1:1,000), cyclin D3 (cat. no. 2936S, 1:1,000), cyclin-dependent kinase (CDK)4 (cat. no. 12790S, 1:1,000), CDK6 (cat. no. 13331S, 1:1,000), p16^INK4a (cat. no. 80772S, 1:1,000), p21^Waf1/Cip1 (cat. no. 2947S, 1:1,000), caspase-3 (cat. no. 9665T, 1:1,000), cleaved caspase-3 (cat. no. 9664T, 1:1,000), B-cell lymphoma-2 (Bcl-2)-associated X protein (Bax; cat. no. 2772T, 1:1,000), Bcl-2 (cat. no. 4223T, 1:1,000), GAPDH (cat. no. 2118S, 1:1,000) and β-actin (cat. no. 8457S, 1:1,000) (Cell Signaling Technology, Danvers, MA, USA). GAPDH and β-actin antibodies were used as loading controls. After washing with Tris-buffered saline containing 0.1% Tween, the membranes were incubated with the following secondary antibodies: Anti-rabbit immunoglobulin G (IgG) (cat. no. 7074P2, 1:5,000; Cell Signaling Technology, Inc.) and anti-mouse IgG (cat. no. ab6728, 1:5,000; Abcam) at room temperature for 1 h. The protein bands were visualized using the enhanced chemiluminescence system (Beyotime Institute of Biotechnology). Densitometric analysis of the protein bands was conducted using ImageJ software (ImageJ 1.48v; National Institutes of Health, Bethesda, MD, USA). The ratio was calculated by comparing the gray value of proteins to GAPDH or β-actin.

All statistical analyses were performed using Stata version 14.0 software (StataCorp LP, College Station, TX, USA); comparisons between two groups were analyzed by independent t-test when data conformed to normal distribution, if not, a non-parametric Kruskal-Wallis test was performed. The 95% confidence interval (CI) of the expression and methylation levels of ABAT in the 40 control samples was defined as the normal level. Values that were below or above the confidence limit were considered low or high expression/methylation. The independent t-test was used to compare the levels between two groups. One-way analysis of variance followed by Bonferroni post hoc test was used for multiple comparisons. χ² tests or Fisher’s exact tests were used for prognostic analysis.

OS was assessed from the day of diagnosis until death due to any cause, or until the last follow-up date (December 30, 2015). Subjects who succumbed prior to leukemia evolution were censored at the time of death. Kaplan-Meier curves were generated to assess the association of gene expression and methylation with survival, and the log-rank test was performed to analyze the association between different clinical indicators and survival. The Cox regression model was used for multivariate analysis, in order to identify independent prognostic factors affecting OS. P<0.05 was considered to indicate a statistically significant difference.

The mRNA expression and methylation levels of ABAT were detected in samples obtained from 152 patients with MDS, 29 patients with AML and 40 controls individuals (Fig. 1A-D). The average expression level in the 152 patients with MDS was 0.31 (0.21-0.40), which was significantly lower compared with in the controls (P<0.0001; Fig. 1A and Table I). When the number of bone marrow blast cells was increased, the mRNA expression levels of ABAT were also upregulated (Table I). According to the WHO subtypes of MDS, patients with high-risk subtypes (i.e. RAEB-1 and RAEB-2) exhibited high expression levels compared with low-risk subtypes (i.e., RA, RARS, RCMD, RCMD-RS, 5q-syndrome and MDS-U) (P=0.09; Fig. 1C and Table I). The low-risk and high-risk subgroups of patients with MDS exhibited lower expression levels than patients with AML (P=0.0001 and P=0.0039, respectively).

The fluorescence intensities of 0, 10, 25, 50, 75 and 100% methylated DNA standards were detected (data not shown) and the fluorescence intensity of nonmethylated DNA standards was subtracted to determine differential fluorescence intensity. A standard curve was established, with y = 0.622x + 0.326, R²=0.998 (data not shown). Subsequently, the fluorescence intensities of DNA methylation of 152 patients with MDS, 29 patients with AML and 40 controls were detected, in order to calculate the percentage of ABAT methylation according to the standard curve. The percentage of ABAT gene methylation was higher in subjects with MDS than in controls (P=0.0001) (Table I and Fig. 1B). The high-risk subgroup of MDS patients exhibited a higher degree of methylation compared with the low-risk subgroup (P<0.0001). No difference was determined between patients with AML and low-risk subjects (P=0.417; Fig. 1D).

Univariate analysis revealed that age, hemoglobin levels, WHO classification, marrow blast percentage, IPSS karyotype, IPSS cytopenias, IPSS risk group, and the expression and methylation levels of ABAT were significant associated with OS in patients with MDS (Table II). The normal expression levels of ABAT (95% CI, 0.46-1.25) and the normal methylation percentage of ABAT (95% CI, 7.98-12.39) were defined in the 40 control samples (Table I). Values below the lower confidence limit were considered the low-expression and low-methylation group; values above the high confidence limit were considered the high-expression and high-methylation group; the other values were considered normal. Therefore, according to this classification, subjects with MDS were categorized into 14 normal-, 131 low- and 7 high-expression patients. In terms of ABAT gene methylation, subjects were categorized into 16 low-, 19 normal- and 117 high-methylation subjects (Table II).

Subjects with normal and high ABAT mRNA expression had a shorter OS compared with those with low expression (P=0.015; Fig. 1E). The median survival time of subjects with high expression was 17.88 months. Furthermore, when the degree of methylation increased, OS decreased (P=0.005; Fig. 1F).

The multivariate Cox regression analysis demonstrated that old age, high blast count and high methylation had a strong impact on the OS of patients with MDS (hazard ratios=2.51, 2.26 and 1.93, respectively); therefore, these were considered independent risk factors for the survival of patients with MDS (Table III). Although ABAT mRNA expression had prognostic value in the univariate analysis, the multivariate analysis revealed that the expression of ABAT was not an independent factor affecting OS (P=0.856).

The mRNA and protein expression levels of ABAT were detected in the various cell groups (Fig. 2). There was no difference in ABAT expression between SKM-1 and THP-1 cell lines (data not shown). HL-60 is also a human leukemia cell line; however, due to its low expression of ABAT (data not shown), SKM-1 and THP-1 cells were used in this study. Three lentiviral vectors were constructed containing three different shRNA sequences (i.e. Sh1-ABAT, Sh2-ABAT and Sh3-ABAT), in order to explore the biological function of ABAT in leukemia cell lines. Following infection with the three vectors, the mRNA expression levels of ABAT were decreased by 90, 70 and 50% in SKM-1 cells compared with the NC group, respectively (Fig. 2A). In THP-1 cells, following infection with the three vectors, the mRNA expression levels of ABAT were decreased by 80, 60 and 40%, respectively (Fig. 2B). There was no difference between the normal control and negative control groups with regards to ABAT mRNA expression. The protein expression levels of ABAT were decreased by ~90, 80 and 50% (Fig. 2C and E) in the Sh1-ABAT, Sh2-ABAT and Sh3-ABAT SKM-1 cells, and by 60, 55 and 30% in the THP-1 cells (Fig. 2D and F) compared with the NC group, respectively. Since the Sh3-ABAT lentivirus exerted minor effects on the mRNA and protein expression levels of ABAT, Sh1-ABAT and Sh2-ABAT lentiviruses were used for further functional analysis.

Relative GABAT activity was detected following infection of SKM-1 and THP-1 cells with shRNA lentiviruses. Relative GABAT activity was decreased by 65 and 40% in the Sh1-ABAT and Sh2-ABAT groups compared with in the NC group in SKM-1 cells (Fig. 2G), and by 35 and 40% in THP-1 cells (Fig. 2H), respectively. (±)-vigabatrin is a specific inhibitor of GABAT; therefore, relative activity of GABAT was also detected in cells following vigabatrin treatment. When the dose of vigabatrin was increased, GABAT activity was decreased to 70% in the 100 µM group, 60% in the 200 µM group and 55% in the 400 µM group compared with in the control group in SKM-1 cells (Fig. 2I). In THP-1 cells, relative GABAT activity was decreased to 70, 40 and 45% in the 100, 200 and 400 µM groups, respectively, compared with in the control group (Fig. 2J).

SKM-1 and THP-1 cells were infected with shRNA lentiviruses and the effects of ABAT knockdown on viability were evaluated. Cells that were not infected with lentivirus or treated with the GABAT inhibitor were considered normal controls; all results were compared to the normal control group. The results revealed that when the expression levels of ABAT were decreased, cell viability was significantly inhibited (Fig. 3A and B). Furthermore, treatment with the GABAT inhibitor also inhibited the viability of SKM-1 and THP-1 cells (Fig. 3C and D). Taken together, these data indicated important roles for ABAT and GABAT in the pathogenesis of MDS.

Cell cycle distribution was detected in the two cell lines, in order to identify whether the inhibitory effects of ABAT gene silencing and GABAT inhibition on growth could be explained by alterations in the cell cycle. In SKM-1 cells, following infection with Sh1-ABAT and Sh2-ABAT, the percentage of cells in G₁ phase was increased to 53.09±4.69 (P<0.01) and 52.85±2.54% (P<0.01) compared with in the negative control group, and the percentage of cells in S phase was decreased to 34.28±0.29 and 34.89±0.22%, respectively (Fig. 5A). The same trend was observed in the THP-1 cells (Fig. 5B). There was no difference between the negative control and normal control groups. In these two cell lines, cell cycle-associated proteins were also detected. The protein expression levels of CDK6, CDK4, cyclin D1 and cyclin D3 were significantly downregulated, whereas the expression levels of the periodic inhibitory proteins, p16^INK4a and p21^Waf1/Cip1, were upregulated in SKM-1 cells in response to ABAT knockdown (Fig. 5C and D). Conversely, in THP-1 cells, the expression of p21^Waf1/Cip1 was decreased (Fig. 5E and F).

Cell cycle distribution was also measured following drug treatment. When SKM-1 and THP-1 cells were treated with 100, 200 and 400 µM vigabatrin for 10 days, the percentage of cells in G₀/G₁ phase was slightly increased (Fig. 6A and B). Furthermore, in SKM-1 cells, the percentage of cells in S phase was decreased from 50.28±5.82% to 43.51±1.12, 43.83±2.49 and 44.32±0.87% in response to 100, 200 and 400 µM vigabatrin (Fig. 6A). In THP-1 cells, the percentage of cells in S phase was decreased from 25.44±1.65% to 21.90±0.61, 22.09±1.44 and 21.61±0.30% in response to 100, 200 and 400 µM vigabatrin (Fig. 6B). The expression levels of cell cycle-associated proteins were also detected by western blotting. A similar trend was observed in cells treated with the GABAT inhibitor as in cells with ABAT gene knockdown (Fig. 6C-F).