Correlation of long non-coding RNA taurine-upregulated gene 1 with disease conditions and prognosis,as well as its effect on cell activities in acute myeloid leukemia

Abstract

OBJECTIVE:

This study aimed to investigate the correlation of long non-coding RNA taurine-upregulated gene 1 (lncRNA TUG1) with clinicopathological characteristics and prognosis in acute myeloid leukemia (AML) patients, as well as its function in cell proliferation and apoptosis.

METHODS:

Two hundred and thirty six de novo AML patients were consecutively enrolled and then underwent conventional induction chemotherapy. Bone marrow samples were obtained from all AML patients and controls. Quantitative polymerase chain reaction assay was performed to detect lncRNA TUG1 expression. KG-1 cells were transfected by TUG1 inhibitor (TUG1 ( $-$ )) and blank inhibitor (NC ( $-$ )) plasmids. Cell proliferation and apoptosis were evaluated by CCK8 and AV/PI assays, and apoptotic markers expressions were detected by Western blot assay.

RESULTS:

LncRNA TUG1 expression was higher in AML patients compared to controls, and it was positively correlated with white blood cell counts as well as poor risk stratification. Additionally, elevated lncRNA TUG1 expression was observed in non-complete remission (non-CR) patients compared to CR patients, and it was correlated with shorter event-free survival and overall survival in AML patients. In the in vitro experiments, lncRNA TUG1 expression was upregulated in AML cell lines compared to control cells, and cell proliferation ability was reduced, but cell apoptosis rate was promoted in TUG1 ( $-$ ) group compared to NC ( $-$ ) group at 72 hours after transfection in KG-1 cells.

CONCLUSIONS:

LncRNA TUG1 predicts advanced disease conditions and poor prognosis in AML patients, and its knockout decreases proliferation and increases apoptosis of AML cells.

1. Introduction

Acute myeloid leukemia (AML), a heterogeneous hematopoietic malignancy, is characterized by accumulated clonal myeloid progenitor cells with limited ability to differentiate into mature blood cells including red blood cells, platelets and granulates, which is clinically manifested by anemia, hematostaxis and serious infection [1]. The incidence rate of AML in infants aged below 1 year is 1.5 per 100,000, and it is less in those younger than 25 years, while for the population elder than 25 years, its rate increases with the growth of age and eventually reaches 25 per 100,000 in octogenarians [2]. Although AML has become curable in a limited number of patients during the past 3 decades with the adoption of remissive induction and post-remission consolidation treatments, the prognosis in the vast majority of AML patients remains unsatisfactory, bringing forward a clinical challenge in AML treatment [3]. Therefore, it is essential to explore novel biomarkers to monitor disease progression as well as to improve the prognosis in AML patients.

Long non-coding RNAs (lncRNAs), broadly defined as non-protein coding RNA molecules with lengths over 200 nucleotides, are enormously abundant in mammalian cells with approximately 28000 distinct transcripts [4]. Recent studies have exhibited that lncRNAs are pivotal regulators of genes and are involved in the initiation and development of various cancers including AML [4, 5, 6, 7]. LncRNA taurine upregulated gene 1 (TUG1), a member of lncRNAs firstly identified in mouse retinal cells, has been found to be highly expressed in human malignant tissues and correlates with adverse clinicopathological characteristics as well as poor prognosis in patients with several carcinomas (e.g., bladder cancer and lung adenocarcinoma) [8, 9, 10]. However, only a recently published study reveals that lncRNA TUG1 is upregulated in AML and correlates with worse prognosis in AML patients in a small population [11], while a microarray data published on GEO database with entry No. GSE 96535 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE96535) discloses that lncRNA TUG1 is not dysregulated in AML patients, indicating that the role of lncRNA TUG1 in AML development and progression is still obscure.

Thus, the purpose of this study was to investigate the correlation of lncRNA TUG1 with clinicopathological characteristics and prognosis in AML patients with large sample size and to further explore its function in cell proliferation and apoptosis in AML cells.

2. Materials and methods

2.1 Participants

This study consecutively enrolled 236 de novo AML patients at the Department of Hematology, Renmin Hospital Affiliated to Hubei University of Medicine from 2015/1/1 to 2017/12/31. Patients with the following conditions were included: (1) Diagnosed as primary AML according to World Health Organization (WHO) Classification of Tumors of Hematopoietic and Lymphoid Tissues (2008). (2) Age between 18 and 65 years. Patients with the following conditions were excluded from this study: (1) Diagnosed as acute promyelocytic leukemia (APL), central nervous system leukemia (CNSL) or secondary AML; (2) History of other malignant tumors (3) Previous exposure to radiotherapy, chemotherapy, targeted therapy, immunotherapy or allogeneic stem cell transplantation; (4) Severe or uncontrolled heart, renal or liver disease; (5) Pregnancy or breastfeeding women. Meanwhile, 118 age and gender-matched controls who underwent bone marrow biopsy (non-hematologic malignancy patients) were also recruited. The Ethics Committee of Renmin Hospital Affiliated to Hubei University of Medicine had approved this study before its initiation, and all patients provided written informed consents.

2.2 Sample acquisition and lncRNA TUG1 detection

Bone marrow sample was obtained by biopsy from AML patients and controls before initiation of any treatments, and lncRNA TUG1 expression in bone marrow was subsequently determined by quantitative polymerase chain reaction (qPCR) assay.

2.3 Baseline data collection and assessments

Baseline information of AML Patient was collected including age, gender, white blood cells (WBC), French-American-British (FAB) classification, cytogenetics, molecular genetics and risk stratification. The FAB classification was categorized in line with FAB classification system; the cytogenetics was evaluated according to An International System for Human Cytogenetic Nomenclature (ISCN 2009), and the risk status was classified according to National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology of AML (Version 2.2013). In addition, data of controls including age and gender were also documented.

2.4 Treatment and outcome evaluation

After the enrollment, AML patients underwent conventional induction chemotherapy regimens including: (1) IA, Idarubicin 8–12 mg/m ${}^2$ /d (days 1–3); Ara-C 100–150 mg/m ${}^2$ /d (days 1–7); (2) MA, Mitoxantrone 8–12 mg/m ${}^2$ /d (days 1–3), Ara-C 100–150 mg/m ${}^2$ /d (days 1–7); (3) DA, Daunorubicin 45 $\sim$ 90 mg/m ${}^2$ /d (days 1–3), Ara-C 100 $\sim$ 200mg/m ${}^2$ /d (days 1–7); (4) HAD, Homoharringtonine 2.0–2.5 mg/m ${}^2$ /d (days 1–7) (or 4 mg/m ${}^2$ /d (days 1–3)); Ara-C 100–150 mg/ m ${}^2$ /d (days 1–7); Daunorubicin 45–90 mg/m ${}^2$ /d (days 1–3). If complete remission (CR) was realized at first course of induction chemotherapy, consolidation treatment or transplantation was applied based on clinical conditions (such as risk stratification, donor source or patient’s health conditions); while if CR was not achieved, an intermediate or intensive dose Ara-C was added to second course referring to the original regimens and patient’s disease conditions. If CR was still not realized after 2 courses of chemotherapy, other regimens was considered.

Treatment response was evaluated after chemotherapy, and CR was defined as bone marrow (BM) with at least 20% cellularity and BM blasts below 5% at steady state after treatment, without cytological evidence of leukemia, no transfusion requirement, leucocyte count above 1 $\times$ 10 ${}^9$ /L and platelet count above 100 $\times$ 10 ${}^9$ /L. All AML patients were regularly followed up to 2017/12/31 with the median follow-up duration 17 months (range from 1–36 months), and event-free survival (EFS) was calculated from the date of treatment to the date of recurrence, progression or death for any cause while overall survival (OS) was calculated from the date of treatment to the date of death from any cause.

2.5 Cell culture

AML cell lines including KG-1, MOLM-14, Kasumi-6, NB-4, THP-1 and HL-60 were purchased from Cell Biology of the Chinese Academy of Sciences (Shanghai, China) or American Type Culture Collection (ATCC) (Manassas, USA), while normal control cells were obtained by isolating CD34 ${}^{+}$ cells from bone marrow aspirating from healthy donors. KG-1, HL-60 and NB-4 cells were cultured in 80% IMDM medium (Gibco, USA) supplemented with 20% fetal bovine serum (FBS) (Gibco, USA); MOLM-14, THP-1 cells and control cells were cultured in 90% RPMI-1640 medium (Sigma-Aldrich, USA) supplemented with 10% FBS (Gibco, USA); Kasumi-6 cells were cultured in 80% RPMI-1640 medium (Sigma-Aldrich, USA) supplemented with 20% FBS (Gibco, USA). All cells were cultured at 37 ${}^{\circ }$ C and 5% CO ${}_2$ . After culturing, lncRNA TUG1 expression was determined by qPCR assay in all AML cells and control cells.

2.6 Transfection and assays

LncRNA TUG1 inhibitor and blank inhibitor plasmids were constructed and transfected into KG-1 cells as TUG1 ( $-$ ) and NC ( $-$ ) group. At 24 hours post-transfection, qPCR assay was used to detect lncRNA TUG1 expression in each group; at 0, 24, 48 and 72 hours post-transfection, CCK8 assay was used to determine cell proliferation ability in each group; AV/PI assay was used to measure cell apoptosis rate in each group at 72 hours post transfection; Western blot assay was adopted to assess the expressions of apoptotic markers (C-Caspase3 and Bcl-2) at 24 hours post-transfection.

Table 1

AML patients’ characteristics

Parameters	AML patients
	( $N=$ 236)
Age (years)	41.6 $\pm$ 11.8
Gender (male/female)	131/105
WBC ( $\times$ 10 ${}^9$ cell/L)	15.9 (7.8–28.5)
FAB Classification (n/%)
M1	2 (0.8)
M2	83 (35.2)
M4	61 (25.8)
M5	79 (33.5)
M6	11 (4.7)
Cytogenetics (n/%)
t(6;9)	1 (0.4)
t(9;22)	2 (0.8)
11q23	3 (1.3)
inv(3) or t(3;3)	3 (1.3)
$-$ 5 or 5q $-$	4 (1.7)
t(9;11)	4 (1.7)
$-$ 7 or 7q $-$	6 (2.5)
$+$ 8	7 (3.0)
t(8;21)	16 (6.8)
inv(16) or t(16;16)	18 (7.6)
Others (not included in better or poor risk)	25 (10.6)
CK	28 (11.9)
NK	119 (50.4)
Monosomal karyotype (n/%)	20 (8.5)
FLT3-ITD (n/%)	55 (23.3)
Isolated biallelic CEBPA mutation (n/%)	23 (9.7)
NPM1 (n/%)	85 (36.0)
Risk stratification (n/%)
Better-risk	66 (28.0)
Intermediate-risk	79 (33.5)
Poor-risk	91 (38.5)

Data were presented as mean value $\pm$ standard deviation, median (1/4–3/4 quartiles) or count (percentage). AML, acute myeloid leukemia; WBC, white blood cell; FBA Classification, French-American-British classification systems; CK, complex karyotype; NK, normal karyotype; FLT3-ITD, internal tandem duplications in the FMS-like tyrosine kinase 3; CEBPA, CCAAT/enhancer-binding protein $\alpha$ ; NPM1, nucleophosmin (controls $N=$ 118: age: 41.9 $\pm$ 15.1, $P=$ 0.836, Gender: 60man/58 female, $P=$ 0.407).

Figure 1.

LncRNA TUG1 relative expression in AML patients and controls. (A) LncRNA TUG1 expression was higher in AML patients compared to controls. (B) ROC curve showed that lncRNA TUG1 had a good potential to distinguish AML patients from controls. Wilcoxon rank sum test was used to compare lncRNA TUG1 expression in AML patients and controls, and ROC curve was used to evaluate the predictive value of lncRNA TUG1 for AML. $P$ value $<$ 0.05 was considered significant. LncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; AML, acute myeloid leukemia; ROC, receiver operating characteristics.

2.7 qPCR assay

Total RNA was extracted using TRIzol reagent (Invitrogen, USA) and qualified by concentration and purity, as well as agarose gel electrophoresis. Subsequently, RNA was reversely transcribed into cDNA using PrimeScript™ RT reagent kit (TAKARA, Japan): (1) mix 1 ug of RNA with distilled water to constant volume 14 ul, then place at 65 ${}^{\circ }$ C for 5 min; (2) add 1 ul of reverse transcriptase mixture, 1 ul of primer and 4 ul of buffer and place at 42 ${}^{\circ }$ C for 18 min then at 95 ${}^{\circ }$ C for 5 min. The expression of lncRNA TUG1 was then quantitated using SYBR Premix Ex TaqTM II (TaKaRa, Japan). The calculation of qPCR result was carried out using the 2 ${}^{-\mathrm{\Delta }\mathrm{\Delta }\text{Ct}}$ method, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the internal reference. The primers of lncRNA TUG1 and GAPDH were as follows: lncRNA TUG1 forward primer: TAGGAGTGG ATGTGTTCTGTAGCA; lncRNA TUG1 reverse primer: TGGTCGTGGAATATGGTCAATGAG; GAPDH forward primer: GAGTCCACTGGCGTCTTCAC; GAPDH reverse primer: ATCTTGAGGCTGTTGTC ATACTTCT.

2.8 CCK-8 assay

Ten ul of CCK-8 (Dojindo, Japan) and 90 ul of medium were added to each group of KG-1 cells, and following that, the cells were incubated at 37 ${}^{\circ }$ C under 95% air and 5% CO ${}_2$ . Cell proliferation ability was then determined.

2.9 AV/PI assay

Before being suspended in 100 ul blinding buffer, KG-1 cells were initially digested using pancreatin and washed with phosphate buffer solution (PBS). Subsequently, 2 ul AV (Invitrogen, USA) was added to the cells; then the cells were placed on the ice for 15 mins in darkness. Following that, 1 ul PI (Invitrogen, USA) was added to the cells, and cell apoptosis rate was analyzed using flow cytometry (FCM) (Becton Dickinson, USA).

2.10 Western blot assay

KG-1 cells were centrifuged after lysing with RIPA lysis buffer (Thermo Scientific, USA). The total protein concentration was detected by the BCA™ protein assay kit (Pierce, USA). Proteins were separated by electrophoresis on sodium dodecyl sulfate-polyacrylamide (SDS) gel. Following that, proteins were transferred to PVDF membrane (Millipore, USA) and the membrane was then blocked for 1 hour at room temperature with 5% non-fat dried milk in PBST. After incubation with C-Caspase3, Bcl-2 and GAPDH antibodies (Cleaved Caspase 3 Rabbit mAb (CST, USA), Anti-Bcl-2 Rabbit Antibody (Abcam, USA) and GAPDH Rabbit mAb (CST, USA)), the PVDF membrane was finally incubated with the mouse anti-rabbit IgG HRP-conjugated secondary antibody (CST, USA). The specific bands indicating protein abundance were visualized by ECL advanced Western blot analysis detection kit (BD, USA).

Table 2

Correlation of LncRNA TUG1 relative expression with clinical characteristics of AML patients

Clinical characteristics	LncRNA TUG1	$P$ value
	relative expression
Age		0.504
$⩾$ 40 years	1.530 (0.738–3.158)
$<$ 40 years	1.526 (0.729–2.577)
Gender		0.063
Male	2.348 (1.174–3.424)
Female	1.728 (0.822–3.363)
WBC		0.010
$⩾$ 10 ( $\times$ 10 ${}^9$ cell/L)	2.404 (1.166–3.448)
$<$ 10 ( $\times$ 10 ${}^9$ cell/L)	1.396 (0.840–3.285)
FAB Classification		0.034
M1	4.283
M2	1.891 (0.873–3.376)
M4	2.348 (1.032–3.587)
M5	1.948 (0.898–2.929)
M6	2.785 (2.092–4.240)
Cytogenetics		0.058
NK	1.891 (0.859–3.190)
Others	2.412 (1.120–3.563)
Monosomal karyotype		0.065
Yes	3.049 (2.372–3.555)
No	1.952 (0.991–3.359)
FLT3-ITD mutation		0.943
Yes	2.086 (0.980–3.560)
No	2.054 (0.979–3.348)
Isolated biallelic CEBPA mutation		0.474
Yes	2.582 (1.053–4.070)
No	2.011 (0.947–3.370)
NPM1		0.845
Yes	1.901 (0.940–3.416)
No	2.179 (0.980–3.363)
Risk stratification		0.022
Better-risk	1.665 (0.835–3.220)
Intermediate-risk	1.985 (0.895–3.114)
Poor-risk	2.776 (1.354–3.603)

Data were presented as median (1/4–3/4 quartiles). Comparison was determined by Wilcoxon rank sum test or Kruskal-Wallis rank sum test. $P<$ 0.05 was considered significant. AML, acute myeloid leukemia; WBC, white blood cell; FAB Classification, French-American-British classification systems; CK, complex karyotype; NK, normal karyotype; FLT3-ITD, internal tandem duplications in the FMS-like tyrosine kinase 3; CEBPA, CCAAT/enhancer-binding protein $\alpha$ ; NPM1, nucleophosmin.

2.11 Statistics

Analysis of statistics was performed using GraphPad Prism 5.01 software (GraphPad Software Inc, USA) and SPSS 21.0 software (IBM, USA). Data were presented as mean $\pm$ standard deviation, median (25 ${}^{\text{th}}$ –75 ${}^{\text{th}}$ quantile) or count (percentage). Comparison between two groups was determined by Wilcoxon rank sum test or t-test, and comparison among groups was carried out using Kruskal-Wallis rank sum test. Receiver operating characteristic (ROC) curve was used to assess the ability of lncRNA TUG1 in distinguishing AML patients from controls. Kaplan-Meier (K-M) curves were performed to evaluate EFS and OS in AML patients, and Log-rank test was used to compare the difference in EFS and OS between lncRNA TUG1 high and low expression groups. $P$ value $<$ 0.05 was considered significant.

Figure 2.

LncRNA TUG1 expression in non-CR patients and CR patients. Compared to CR patients, lncRNA TUG1 expression was observed to be higher in non-CR patients. Wilcoxon rank sum test was used for the comparison of lncRNA TUG1 expression between non-CR patients and CR patients, and $P$ value $<$ 0.05 was considered significant. LncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; CR, complete remission.

3. Results

3.1 Baseline characteristics

The mean ages for AML patients ( $N=$ 236) and controls ( $N=$ 118) were 41.6 $\pm$ 11.8 and 41.9 $\pm$ 15.1 respectively ( $P=$ 0.836), and there were 131 males and 105 females in AML group, while 60 males and 58 females in control group ( $P=$ 0.407). For AML patients, the median WBC count was 15.9 (7.8–28.5) $\times$ 10 ${}^9$ cell/L. As to FAB classification, the numbers of patients with M1, M2, M4, M5 and M6 were 2 (0.8%), 83 (35.2%), 61 (25.8%), 79 (33.5%) and 11 (4.7%) respectively. There were 66 (28.0%) patients with better-risk stratification, 79 (33.5%) with intermediate-risk stratification and 91 (38.5%) with poor-risk stratification. Other baseline characteristics of AML patients were listed in Table 1.

Figure 3.

Correlation of lncRNA TUG1 expression with EFS and OS in AML patients. (A) Worse EFS was presented in AML patients with lncRNA TUG1 high expression compared to patients with lncRNA TUG1 low expression. (B) Shorter OS was observed in AML patients with lncRNA TUG1 high expression compared to patients with lncRNA TUG1 low expression. K-M curve was used to assess EFS and OS in AML patients, and comparison of median EFS and OS between patients with lncRNA TUG1 high expression and low expression was carried out by log-rank test. $P$ value $<$ 0.05 was considered significant. LncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; EFS, event-free survival; OS, overall survival; AML acute myeloid leukemia; K-M curve, Kaplan-Meier curve.

Figure 4.

LncRNA TUG1 expression in AML cell lines and control. LncRNA TUG1 expression was increased in AML cell lines including KG-1, MOLM-14, Kasumi-6, THP-1 and HL-60 compared to control cells, and the most significant increase was presented in KG-1 cells. Comparison between two groups was performed using t-test. NS non-significant, * $P<$ 0.05, ** $P<$ 0.01, *** $P<$ 0.001. LncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; AML, acute myeloid leukemia.

3.2 LncRNA TUG1 relative expression in AML patients and controls

LncRNA TUG1 expression in AML patients (2.070 (0.981–3.373)) was increased compared to controls (0.746 (0.343–1.686)) ( $P<$ 0.001) (Fig. 1A). ROC curve illustrated that lncRNA TUG1 had good predictive value in distinguishing AML patients from controls, with area under curve (AUC) of 0.754 (95% CI: 0.704–0.805) (Fig. 1B).

3.3 Correlation of lncRNA TUG1 expression with clinicopathological characteristics in AML patients

LncRNA TUG1 expression was positively correlated with WBC counts ( $P=$ 0.010) and poor risk stratification ( $P=$ 0.022) in AML patients. And its expression was different in patients with M1, M2, M4, M5 and M6 FAB classification ( $P=$ 0.034). No correlation was observed between lncRNA TUG1 expression and other clinicopathological characteristics in AML patients (Table 2).

3.4 LncRNA TUG1 expression in non-CR patients and CR patients

There were 56 non-CR patients and 180 CR patients after induction chemotherapy, and lncRNA TUG1 expression in non-CR patients (2.611 (1.393–3.725)) was higher than that in CR patients (1.946 (0.857–3.315)) ( $P=$ 0.009) (Fig. 2).

3.5 EFS and OS in AML patients with lncRNA TUG1 high expression and low expression

The median EFS in AML patients with lncRNA TUG1 high expression (9.0 (95% CI: 7.1–10.9) months) was shorter than ( $P<$ 0.001) that in patients with lncRNA TUG1 low expression (21.0 (95% CI: 16.4–25.6) months) (Fig. 3A). In addition, worse median OS ( $P=$ 0.002) was observed in patients with lncRNA TUG1 high expression (21.0 (95% CI: 15.7–26.3) months) compared to patients with lncRNA TUG1 low expression (32.0 (95% CI: 28.1–35.9) months) (Fig. 3B).

Figure 5.

Correlation of lncRNA TUG1 expression with AML cell proliferation and apoptosis. (A) At 24 hours post-transfection in KG-1 cells, lncRNA TUG1 expression was reduced in TUG1 ( $-$ ) group compared to NC ( $-$ ) group. (B) Cell proliferation ability was reduced in TUG1 ( $-$ ) group compared to NC ( $-$ ) group at 72 hours post-transfection. (C, E) Cell apoptosis rate was promoted in TUG1 ( $-$ ) group compared to NC ( $-$ ) group at 72 hours post-transfection. (D) C-Caspase3 expression was increased, while Bcl-2 expression was decreased in TUG1 ( $-$ ) group compared to NC ( $-$ ) group at 24 hours post-transfection. Comparison between two groups was performed using t-test, and protein expression was determined by Western blot assay. NS non-significant, ** $P<$ 0.01, *** $P<$ 0.001. LncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; AML, acute myeloid leukemia; NC, negative control.

3.6 LncRNA TUG1 expression in AML cell lines and control cells

In order to further understand the underlying mechanism of lncRNA TUG1 in AML, cells experiments were performed. LncRNA TUG1 expression was increased in AML cell lines including KG-1 ( $P<$ 0.001), MOLM-14 ( $P<$ 0.05), Kasumi-6 ( $P<$ 0.01), THP-1 ( $P<$ 0.05) and HL-60 ( $P<$ 0.01) compared to control cells, and the most significant increase was in KG-1 cells (Fig. 4). Thus, KG-1 cells were used for transfection in the following experiments.

3.7 Function of lncRNA TUG1 in cell proliferation and apoptosis in AML

At 24 hours post-transfection in KG-1 cells, qPCR assay was performed and disclosed that lncRNA TUG1 expression was decreased in TUG1 ( $-$ ) group compared to NC ( $-$ ) group ( $P<$ 0.001) (Fig. 5A). Compared to NC ( $-$ ) group, cells proliferation ability was reduced in TUG1 ( $-$ ) group ( $P<$ 0.01) at 72 hours post-transfection in KG-1 cells (Fig. 5B). Regarding cell apoptosis, AV/PI assay displayed that cell apoptosis rate was promoted ( $P<$ 0.01) in TUG1 ( $-$ ) group compared to NC ( $-$ ) group at 72 hours post-transfection (Fig. 5C and E). Also, cell apoptotic markers expressions were visualized by Western blot assay at 24 hours post-transfection, which showed that C-Caspase3 expression was upregulated, and Bcl-2 expression was downregulated in TUG1 ( $-$ ) group compared to NC ( $-$ ) group (Fig. 5D). These results implied that the knockout of lncRNA TUG1 inhibited proliferation and promoted apoptosis in KG-1 cells.

4. Discussion

In the present study, we observed that: (1) LncRNA TUG1 high expression was correlated with advanced clinicopathological characteristics as well as shorter EFS and OS in AML patients. (2) Knockout of lncRNA TUG1 repressed proliferation and facilitated apoptosis of KG-1 cells in vitro.

LncRNA TUG1 is considered as a novel oncogene, which is highly expressed in various cancers and correlates to advanced disease conditions in cancer patients. In a recent study, lncRNA TUG1 expression is detected to be elevated in endometrial carcinoma tissues compared to adjacent normal tissues [12]. Another study demonstrated that lncRNA TUG1 is overexpressed in intrahepatic cholangiocarcinoma, and it associates with poor clinicopathological characteristics including advanced tumor stage, intrahepatic metastasis, lymph node metastasis and perineural invasion [13]. Also, lncRNA TUG1 is highly expressed in gastric cancer and positively correlates with invasion depth and TNM stage in gastric cancer patients [14]. These previous studies illuminate that lncRNA TUG1 high expression is associated with aggravated disease conditions in patients with different cancers. In addition, several studies have illustrated the cancerogenic function of lncRNAs such as lncRNA H19 and lncRNA zinc finger antisense 1 (ZFAS1) in AML [15, 16]. However, considering that the tumorigenic function of lncRNA TUG1 in hematological malignancies, especially in AML, remains obscure, we aimed to investigate the correlation of lncRNA TUG1 with clinicopathological characteristics and prognosis in AML patients. From our result, lncRNA TUG1 expression was elevated in AML patients compared to controls and was positively correlated with WBC count and poor risk stratification in AML patients. The possible explanation could be that upregulation of lncRNA TUG1 altered corresponding signaling pathways to promote proliferation and suppress apoptosis of AML cells, hence facilitated disease progression and worsened disease conditions in AML patients, which was supported by the following in vitro experiment.

In the meantime, the prognostic value of lncRNA TUG1 has been investigated in researches on cancer, which indicate that lncRNA TUG1 is negatively correlated with prognosis in cancers including muscle-invasive bladder cancer, colorectal cancer and esophageal squamous cell carcinoma [17, 18, 19]. Although one recently published study by Wang et al. illustrates the negative correlation of lncRNA TUG1 with EFS and OS in AML patients, the sample size in that previous study is relatively small, and their patients mainly come from East China [11]. Therefore, in the present study, we enrolled 236 patients, which was relatively larger sample size and from a different region (Middle China) and discovered that lncRNA TUG1 high expression was correlated with worse EFS and shorter OS in AML patients. Our results were in accordance with that of Wang et al. and the possible explanations were: (1) LncRNA TUG1 altered genes and corresponding cellular signaling pathways to promote AML cell growth, facilitating disease progression and increasing disease severity, thereby contributed to poor prognosis in AML patients. (2) LncRNA TUG1 might attenuate drug-sensitivity of cells to alter the chemotherapy outcomes, hence increased the recurrence of disease and led to poor prognosis in AML patients. However, the follow-up duration in this study was relatively short with the median follow-up time being 17 months (range from 1-36 months). Thus, the long-term effect of lncRNA TUG1 on prognosis in AML patients was not investigated, and further study with longer follow-up was needed. In addition, due to lack of non-hematologic malignancy patients underwent bone marrow biopsy, the number of controls ( $N=$ 118) was not matched with the number of AML patients ( $N=$ 236), which might have negative impact on statistical power.

In order to explore the underlying mechanisms of lncRNA TUG1 in various cancers, in vitro experiments have been extensively performed. For example, lncRNA TUG1 promotes cell proliferation, migration and invasion by inhibiting miR-29c in bladder cancer and enhances cancer cell growth through sponging miR-382 in pancreatic cancer [10, 20]. Also, lncRNA TUG1 increases proliferation and decreases apoptosis in osteosarcoma cells via downregulating miR-132-3p and upregulating expression of SOX4 [9]. These studies imply that lncRNA TUG1 promotes tumorigenesis by initiating cell proliferation and inhibiting cell apoptosis in different cancers, however, its mechanism in leukemias remains unclear. To further understand the function of lncRNA TUG1 in AML, we transfected TUG1 inhibitor mimic and blank mimic plasmids into KG-1 cells and assessed cell proliferation and cell apoptosis rate. Our result revealed that knockout of lncRNA TUG1 suppressed proliferation and accelerated apoptosis of KG-1 cells, suggesting that it was involved in the pathogenesis of AML by regulating cellular proliferation and apoptosis, which was in line with the results from previously stated in vitro experiments and revealed the potential therapeutic application of lncRNA TUG1 in AML.

In summary, lncRNA TUG1 predicts increased disease severity and poor prognosis in AML patients, and its knockout decreases proliferation and increases apoptosis of AML cells.