Long non-coding RNA taurine-upregulated gene 1 correlates with unfavorable prognosis in patients with refractory or relapsed acute myeloid leukemia treated by purine analogue based chemotherapy regimens

Abstract

OBJECTIVE:

This study aimed to explore the correlation of long non-coding RNA taurine-upregulated gene 1 (lncRNA TUG1) expression with clinicopathological features and its predictive value for treatment response and survival profiles in refractory or relapsed acute myeloid leukemia (R/R AML) patients.

METHODS:

Seventy three R/R AML patients who received cladribine combined with cytarabine and granulocyte colony-stimulating factor (G-CSF) (CLAG) or fludarabine combined with cytarabine and G-CSF (FLAG) based chemotherapy and 37 non-malignant controls were recruited. LncRNA TUG1 expression was detected in bone marrow sample obtained before treatment. Complete response (CR), partial response (PR), overall response rate (ORR) and overall survival (OS) were evaluated.

RESULTS:

LncRNA TUG1 expression was upregulated in R/R AML patients compared to controls. It was also elevated in R/R AML patients with age $\geqslant$ 60 years (vs. age $<$ 60 years, $P=$ 0.030) and in patients with secondary AML (vs. primary AML, $P=$ 0.035). R/R AML patients with lncRNA TUG1 high expression achieved numerically lower CR ( $P=$ 0.053), decreased ORR ( $P=$ 0.028) and shorter OS ( $P<$ 0.001) than patients with lncRNA TUG1 low expression. Univariate logistic regression and COX’s regression disclosed that lncRNA TUG1 high expression correlated with declined ORR, numerically decreased CR, and reduced OS. Furthermore, multivariate analyses verified that lncRNA TUG1 high expression was an independent predictive factor for decreased ORR and worse OS.

CONCLUSIONS:

In conclusion, lncRNA TUG1 expression was elevated in R/R AML patients, and it might serve as a potential biomarker for poor prognosis in R/R AML patients treated with CLAG or FLAG based chemotherapy.

Keywords

LncRNA TUG1 refractory or relapsed acute myeloid leukemia (R/R AML)CLAG FLAG prognosis

1. Introduction

Acute myeloid leukemia (AML), a common myeloid disease, is characterized by the proliferative, clonal, abnormally differentiated, or occasionally poorly differentiated cells penetrating into the bone marrow, blood and other tissues [1, 2]. The hemopoietic progenitor cells in AML lose the normal differentiation function and the regular respond to proliferation, and such a disorder results in fatal infection, bleeding or organ infiltration [2]. Although the complete response (CR) rate reaches 80% in patients receiving initial induction chemotherapy, there are still some patients classified as refractory or relapsed AML (R/R AML) patients: approximately 20% to 50% AML patients ultimately diagnosed as refractory disease, and 20% to 70% patients relapse after achieving CR with initial induction chemotherapy [3, 4, 5]. For R/R AML patients, the 1-year survival is only 29% and 5-year survival is as low as 11% [6]. Hence, effective chemotherapy for R/R AML patients is needed.

One of the common therapies for R/R AML patients is the chemotherapy regimen consisting of cytarabine, purine analogs and granulocyte colony-stimulating factor (G-CSF) [3, 7]. For purine analogs, fludarabine and cladribine are widely applied in treating R/R AML [8, 9, 10]. The regimen of fludarabine combined with cytarabine and G-CSF (FLAG) achieves CR of 46% to 63% in R/R AML patients [4, 11, 12]. Meanwhile, another frequently-used regimen, cladribine with cytarabine and G-CSF (CLAG), has obtained CR in approximately 78.8% [13, 14]. For Chinese R/R AML patients, CLAG has been proposed as the first-line treatment [13, 14]. These chemotherapies elevate the remission rates in R/R AML patients, whereas only around 20% of patients attain long-term relapse-free survival [15]. Increasing evidence has demonstrated that knowledge of prognostic biomarkers for R/R AML would contribute to extending survival of these patients. Thus, exploration of novel biomarkers for prognosis is important to improve the treatment outcomes of R/R AML patients [16].

Long non-coding RNA (lncRNA) is a class of endogenous RNA molecules with transcript length more than 200 nucleotides [16]. As reported, lncRNA lacks protein-coding potential and plays a crucial role in hematological malignancies [16, 17]. LncRNA taurine-upregulated gene 1 (TUG1), located at chromosome 22q12, is initially identified as a key regulator in the formation of photoreceptors and retinal development [18, 19]. Of late, the fact that lncRNA TUG1 takes part in carcinogenesis in a number of cancers such as breast cancer, bladder cancer and glioma is reported [18, 20]. Whereas, the expression and biologic effect of lncRNA TUG1 in R/R AML patients are rarely investigated. Hence, this study was conducted to evaluate the correlation of lncRNA TUG1 expression with clinicopathological features. Furthermore, the predictive value of lncRNA TUG1 for treatment response and survival profile in R/R AML patients treated by CLAG or FLAG chemotherapy was investigated.

2. Methods

2.1 Participants

This study was a part of the cohort study investigating the efficacy of CLAG compared to FLAG treatment in R/R AML patients. Seventy three R/R AML patients who received purine analogue based chemotherapy regimens (CLAG or FLAG) from Jul 2013 to Jun 2016 at Department of Hematology in Affiliated Hospital of North Sichuan Medical College, were consecutively enrolled in this prospective cohort study. The inclusion criteria were as follows: (1) Diagnosed as AML according to classification of morphology, immunology, cytogenetics, molecular biology (MICM); (2) Refractory or relapsed disease; (3) Bone marrow (BM) blasts above 10%; (4) About to receive purine analogue based chemotherapy regimens (CLAG or FLAG); (5) Able to be followed up regularly. The exclusion criteria were: (1) Previous purine analogue based chemotherapy (which indicated if multiple-lines of purine analogue based chemotherapy were performed in one patient, only the first time would be included); (2) Severe heart dysfunction, severe arrhythmia, severe lung dysfunction, severe hepatic or renal dysfunction; (3) Other solid tumor history; (4) Known allergic or resistant to drugs included in CLAG or FLAG regimens; (5) Patients in pregnancy or lactation, or planned for pregnancy.

Besides, 37 age and gender-matched non-malignant patients who received bone marrow biopsy (including 24 thrombocytopenic purpuras, 8 infectious diseases and 5 health donors) were consecutively enrolled as controls to elevate the expression of lncRNA TUG1 in R/R AML and controls.

The Institutional Review Board of Affiliated Hospital of North Sichuan Medical College approved this study, and written informed consents were obtained from all patients. This study was conducted according to Declaration of Helsinki.

2.2 Sample collection

Bone marrow samples were collected before treatment in R/R AML patients, and were also collected in non-malignant controls. Total RNA was extracted with TRIzol Reagent (Invitrogen, USA), according to the manufacturer’s instructions.

2.3 Quantitative polymerase chain reaction (qPCR) assay for lncRNA TUG1 expression

The RNA was quantified by OD 260 (Takara, Japan). Then 1 $\mu$ g of total RNA from each sample was used for the synthesis of cDNA with reverse transcription kit (TaKaRa, Japan). Subsequently, cDNA product was subjected to qPCR with SYBR Green kit (TaKaRa, Japan). The amplification in qPCR was conducted in the following conditions: 95 ${}^{\circ}$ C for 5 min, followed by 40 cycles of 95 ${}^{\circ}$ C for 5 s, and finally 61 ${}^{\circ}$ C for 30 s. Results of qPCR were calculated by the method of 2 ${}^{-\Delta\Delta\text{Ct}}$ and U6 was served as the internal reference. The primers sequences were as follows: lncRNA TUG1: Forward 5’ TAGGAGTGGATGTGTTCTGTA GCA 3’ Reverse 5’ TGGTCGTGGAATATGGTCAAT GAG 3’. U6: Forward 5’ CTCGCTTCGGCAGCACA TA3’ Reverse 5’CTCGCTTCGGCAGCACATA3’.

2.4 Treatments

The treatments of R/R AML patients were not intervened in this cohort study, and patients received purine analogue based chemotherapy regimens (CLAG or FLAG) “on demand” according to disease condition and patients’ wills. Patients were divided into two groups according to the actual treatment regimens: (1) CLAG group ( $N=$ 27), 5 mg/m ${}^{2}$ /d cladribine (days 1–5); 2 g/m ${}^{2}$ /d cytarabine (days 1–5) and 300 $\mu$ g/d filgrastim (days 0–5), and 4 patients combined with mitoxantrone treatment at a dose of 10 mg/m ${}^{2}$ /d on days 1–3. (2) FLAG group ( $N=$ 46), 30 mg/m ${}^{2}$ /d fludarabine (days 1–5), 2 g/m ${}^{2}$ /d cytarabine (days 1–5), and 300 $\mu$ g/d filgrastim (days 0–5), and 5 patients combined with idarubicin treatment at a dose of 10 mg/m ${}^{2}$ /d on days 1–3.

2.5 Assessments

CR, partial response (PR) and overall response rate (ORR) were evaluated in 62 patients (available for assessment). Percentage of patients achieving allo-HSCT was also assessed in all patients. Overall survival (OS) was calculated from the date of treatment to the time of death from any cause. The median follow-up duration was 10.0 months (1/4–3/4 quartiles: 4.0–15.5 months), and the last follow up month was Feb 2017.

2.6 Definitions

Refractory AML was defined as (1) patients did not achieve CR after two courses of induction chemotherapy by standard protocol; (2) patients relapsed within 6 months after the first CR; (3) patients relapsed at 6 months or above after first CR and fail by the subsequent induction chemotherapy; (4) patients relapsed more than 2 times; (5) extra-medullary infiltration of leukemia. Relapsed AML was defined as leukemic cells reappeared in peripheral blood or the percentage of bone marrow blasts over 10%. Risk stratification was assessed by cytogenetics and molecular abnormalities according to NCCN guidelines. CR was defined as 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. PR was defined as either BM blasts 5–25%, or a 50% or better decrease in BM blasts, or BM blasts $<$ 5% but with Auer rods’ presence. ORR was deemed as patients achieving CR plus PR.

Table 1
Baseline characteristics of R/R AML patients

Items	R/R AML patients
	( $N=$ 73)
Age (years)	48.73 $\pm$ 18.84
Gender (male/female)	36/37
Disease status
Relapsed (n/%)	45 (61.6)
Refractory (n/%)	28 (38.4)
Primary or secondary disease
Primary (n/%)	57 (78.1)
Secondary (n/%)	16 (21.9)
Risk stratification
Favorable (n/%)	9 (12.3)
Intermediate (n/%)	39 (53.4)
Poor (n/%)	23 (31.5)
Unknown (n/%)	2 (2.7)
ECOG performance score
0 (n/%)	21 (28.8)
1 (n/%)	46 (63.0)
2 (n/%)	6 (8.2)
BM blasts at salvage therapy (%)	42.1 (24.2–64.9)
CR at first induction
Yes (n/%)	32 (43.8)
No (n/%)	41 (56.2)
Previous allo-HSCT
Yes (n/%)	15 (20.5)
No (n/%)	58 (79.5)
Lines of salvage therapy
First salvage therapy (n/%)	57 (78.1)
Second or higher salvage therapy (n/%)	16 (21.9)
CLAG regimen (n/%)	27 (37.0)
CLAG alone (n/%)	23 (31.5)
CLAG $+$ mitoxantrone (n/%)	4 (5.5)
FLAG regimen (n/%)	46 (63.0)
FLAG alone (n/%)	41 (56.2)
FLAG $+$ idarubicin (n/%)	5 (6.8)

Data were presented as a mean value $\pm$ standard deviation, median value (1/4–3/4 quartiles), or count (percentage). R/R AML, relapsed or refractory acute myeloid leukemia; ECOG, Eastern Cooperative Oncology Group; BM, bone marrow; CR, complete remission; allo-HSCT, allogeneic hematopoietic stem-cell transplantation.

2.7 Statistics

Statistical analysis was carried out by SPSS 21.0 software (IBM, USA) and Graphpad Prism 6.01 software (GraphPad Software Inc, USA). Data was mainly presented as mean value $\pm$ standard deviation, median value (1/4–3/4 quartiles), or count (percentage). Comparison between two groups was evaluated by t-test, Wilcoxon rank sum test or Chi-square test. Comparison among three or above groups was determined by Kruskal-Wallis H rank sum test. Kaplan-Meier (K-M) curves and log-rank test were applied to compare OS. Univariate logistic regression analysis and multivariate logistic regression analysis were conducted to assess the factors affecting CR and ORR, meanwhile, univariate and multivariate Cox’s proportional hazards regression analyses were performed to evaluate factors affecting OS. $P<$ 0.05 was considered significant.

3. Results

3.1 Baseline characteristics

Figure 1.

Comparison of lncRNA TUG1 relative expression between R/R AML patients and controls. LncRNA TUG1 relative expression was remarkably higher in R/R AML patients compared with the controls. Comparison between groups was performed by Wilcoxon rank sum test. $P<$ 0.05 was considered significant. LncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; R/R AML, refractory or relapsed acute myeloid leukemia.

Totally 73 R/R AML patients with mean age of 48.73 $\pm$ 18.84 years were enrolled in this study, including 36 males and 37 females (Table 1). Among these, 45 (61.6%) and 28 (38.4%) patients were diagnosed as relapsed disease and refractory disease respectively, meanwhile, 57 (78.1%) and 16 (21.9%) patients were with primary diseases and secondary diseases respectively. According to risk stratification, 9 (12.3%), 39 (53.4%) and 23 (31.5%) patients were categorized as favorable, intermediate and poor risk stratification respectively, while 2 (2.7%) patients were not evaluated for risk stratification. Additionally, all cases were assessed according to ECOG performance, which showed that 21 (28.8%), 46 (63.0%) and 6 (8.2%) patients were with ECOG performance score 0, 1 and 2 respectively. In total patients, the BM blasts at salvage therapy were 42.1% (24.2%–64.9%). As to treatment history: 32 (43.8%) patients had CR at first induction; 15 (20.5%) patients had received the previous allo-HSCT; 57 (78.1%) patients were at first salvage therapy; 16 (21.9%) patients were at second or higher salvage therapy. Furthermore, there were 27 (37.0%) patients received CLAG regimen (including 23 (31.5%) patients received CLAG alone and 4 (5.5%) patients received CLAG $+$ Mitoxantrone) and 46 (63.0%) patients received FLAG regimen (including 41 (56.2%) patients received FLAG alone and 5 (6.8%) patients received FLAG $+$ Idarubicin) respectively.

3.2 LncRNA TUG1 expressions in R/R AML patients

Comparison of lncRNA TUG1 relative expression was carried out between R/R AML patients ( $N=$ 73) and controls ( $N=$ 37). As exhibited in Fig. 1, expression of lncRNA TUG1 was remarkably higher in R/R AML patients compared to the controls ( $P<$ 0.001).

Figure 2.

Analysis of lncRNA TUG1 expression in subgroups. Expression of lncRNA TUG1 was elevated in patients with age $>$ 60 years (A) and patients with secondary AML (D) (vs primary AML), in addition, it was numerically lower in patients with relapsed AML without significant difference (C) (vs refractory AML). No difference of lncRNA TUG1 expression was found in other subgroups including: gender (B), risk stratification (E), ECOG score (F), BM blasts (G), CR at first induction (H), previous allo-HSCT (I) and lines of salvage therapy (J). Comparison between groups was performed by Wilcoxon rank sum test. $P<$ 0.05 was considered significant. LncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; AML, acute myeloid leukemia; ECOG, Eastern Cooperative Oncology Group; BM, bone marrow; CR, complete remission; allo-HSCT, allogeneic hematopoietic stem-cell transplantation.

Figure 3.

Comparison of CR and ORR between patients with lncRNA TUG1 high expression and lncRNA TUG1 low expression. Patients with lncRNA TUG1 high expression achieved numerically lower CR than patients with lncRNA TUG1 low expression without significant difference (A). Patients with lncRNA TUG1 high expression obtained decreased ORR compared with patients with lncRNA TUG1 low expression(B). Comparison between groups was performed by Chi-square test. $P<$ 0.05 was considered significant. CR, complete remission; ORR, overall response rate (ORR); lncRNA TUG1, long non-coding RNA taurine-upregulated gene 1.

Table 2

Factors affecting CR achievement by logistic regression analysis

Parameters	Univariate logistic regression				Multivariate logistic regression
	$P$ value	OR	95% CI		$P$ value	OR	95% CI
			Lower	Higher			Lower	Higher
LncRNA TUG1 high expression	0.056	0.356	0.123	1.027	0.267	0.461	0.118	1.807
CLAG treatment (vs FLAG treatment)	0.471	1.467	0.517	4.157	0.198	2.585	0.609	10.976
Age $\geqslant$ 60 years	0.043	0.298	0.092	0.965	0.265	0.410	0.086	1.964
Male (vs female)	0.267	0.559	0.200	1.560	0.738	1.285	0.295	5.608
Relapsed disease (vs refractory)	0.471	0.682	0.241	1.933	0.513	0.627	0.155	2.536
Secondary disease (vs primary)	0.363	0.545	0.148	2.012	0.827	0.823	0.143	4.741
Poorer risk stratification	0.218	0.594	0.259	1.361	0.775	0.854	0.290	2.514
Higher ECOG performance score	0.129	2.318	0.783	6.858	0.108	3.126	0.778	12.560
BM blasts $\geqslant$ 42.1%	0.466	0.686	0.249	1.889	0.171	0.384	0.098	1.513
CR at first induction	0.568	1.347	0.485	3.741	0.747	0.775	0.165	3.641
Previous allo-HSCT	0.132	0.339	0.083	1.385	0.247	0.381	0.074	1.951
Second or higher salvage therapy (vs first)	0.042	0.189	0.038	0.945	0.125	0.217	0.031	1.526

Data were presented as $P$ value, OR and 95% CI. Factors affecting CR achievement were determined by univariate logistic regression analysis and multivariate logistic regression analysis. A $P$ value $<$ 0.05 was considered significant. Risk stratification was scored as 1 – Favorable;2 – Intermediate; 3 – Poor. ECOG performance score was scored as 0; 1; 2. The analysis was based on these definitions. CR, complete remission; OR, odds ratio; CI, confidence interval; ECOG, Eastern Cooperative Oncology Group; BM, bone marrow; allo-HSCT, allogeneic hematopoietic stem-cell transplantation.

Table 3

Factors affecting ORR achievement by logistic regression analysis

Parameters	Univariate logistic regression				Multivariate logistic regression
	$P$ value	OR	95% CI		$P$ value	OR	95% CI
			Lower	Higher			Lower	Higher
LncRNA TUG1 high expression	0.030	0.313	0.109	0.894	0.022	0.170	0.037	0.777
CLAG treatment (vs FLAG treatment)	0.382	1.607	0.554	4.659	0.165	2.974	0.639	13.838
Age $\geqslant$ 60 years	0.086	0.389	0.132	1.144	0.354	0.492	0.110	2.203
Male (vs female)	0.934	1.044	0.380	2.870	0.151	3.202	0.653	15.697
Relapsed disease (vs refractory)	0.850	1.106	0.390	3.136	0.608	1.472	0.336	6.455
Secondary disease (vs primary)	0.115	0.363	0.103	1.278	0.497	0.557	0.103	3.010
Poorer risk stratification	0.934	0.966	0.428	2.180	0.605	1.325	0.456	3.847
Higher ECOG performance score	0.615	0.768	0.274	2.149	0.622	0.710	0.181	2.777
BM blasts $\geqslant$ 42.1%	0.829	1.118	0.407	3.067	0.966	0.971	0.251	3.759
CR at first induction	0.323	1.690	0.598	4.780	0.478	0.543	0.100	2.940
Previous allo-HSCT	0.331	0.543	0.158	1.862	0.356	0.477	0.099	2.297
Second or higher salvage therapy (vs first)	0.008	0.145	0.035	0.603	0.023	0.107	0.016	0.731

Data were presented as $P$ value, OR and 95% CI. Factors affecting ORR achievement were determined by univariate logistic regression analysis and multivariate logistic regression analysis. A $P$ value $<$ 0.05 was considered significant. Risk stratification was scored as 1 – Favorable;2 – Intermediate; 3 – Poor. ECOG performance score was scored as 0; 1; 2. The analysis was based on these definitions. ORR, overall remission rate; OR, odds ratio; CI, confidence interval; ECOG, Eastern Cooperative Oncology Group; BM, bone marrow; CR, complete remission; allo-HSCT, allogeneic hematopoietic stem-cell transplantation.

3.3 Subgroup analysis of lncRNA TUG1 expression

Comparison of lncRNA TUG1 expression was performed between subgroups divided by clinicopathological features (Fig. 2). LncRNA TUG1 expression was increased in patients with age $\geqslant$ 60 years than that in patients with age $<$ 60 years (Fig. 2A, $P=$ 0.030), and it was also elevated in patients with secondary AML compared with patients with primary AML (Fig. 2D, $P=$ 0.035). Additionally, a downtrend of lncRNA TUG1 expression was existed in relapsed AML patients compared with refractory AML patients (Fig. 2C, $P=$ 0.092). No difference of lncRNA TUG1 expression was observed in other subgroups, which were divided by gender (Fig. 2B, $P=$ 0.129), risk stratification (Fig. 2E, $P=$ 0.126), ECOG score (Fig. 2F, $P=$ 0.391), BM blasts (Fig. 2G, $P=$ 0.156), CR at first induction (Fig. 2H, $P=$ 0.153), previous allo-HSCT (Fig. 2I, $P=$ 0.900) or lines of salvage therapy respectively (Fig. 2J, $P=$ 0.161).

3.4 Treatment response in patients with high lncRNA TUG1 expression and patients with low lncRNA TUG1 expression

Totally 62 patients were included in treatment remission assessment and were divided into lncRNA TUG1 high expression group and low expression group according to the median value of lncRNA TUG1. Patients with lncRNA TUG1 high expression achieved CR of 28.6%, which was numerically lower than that in patients with lncRNA TUG1 low expression (52.9%), while no significant difference was observed ( $P=$ 0.053, Fig. 3A). Meanwhile, ORR was reduced in patients with lncRNA TUG1 high expression compared with patients with lncRNA TUG1 low expression (42.9% vs. 70.6%, $P=$ 0.028) (Fig. 3B).

3.5 Analysis of factors affecting CR

Factors affecting CR were assessed by univariate logistic regression and exhibited in Table 2, which revealed that lncRNA TUG1 high expression was numerically correlated with a lower possibility of CR ( $P=$ 0.056). Besides, age $\geqslant$ 60 years ( $P=$ 0.043) and second or higher salvage therapy ( $P=$ 0.042) predicted the absence of CR. Furthermore, no independent predictive factor for CR was observed in the multivariate model analysis.

Table 4
Factors affecting OS by Cox’s proportional hazards regression analysis

Parameters	Univariate Cox’s regression				Multivariate Cox’s regression
	$P$ value	HR	95% CI		$P$ value	HR	95% CI
			Lower	Higher			Lower	Higher
LncRNA TUG1 high expression	$<$ 0.001	2.629	1.540	4.486	0.001	2.787	1.551	5.011
CLAG treatment (vs FLAG treatment)	0.171	0.681	0.392	1.181	0.343	0.706	0.343	1.451
Age $\geqslant$ 60 years	0.288	1.330	0.786	2.249	0.206	0.636	0.315	1.282
Male (vs female)	0.171	0.681	0.392	1.181	0.095	1.810	0.903	3.628
Relapsed disease (vs refractory)	0.162	0.681	0.398	1.166	0.487	0.777	0.382	1.582
Secondary disease (vs primary)	0.028	1.977	1.078	3.625	0.082	1.967	0.917	4.217
Poorer risk stratification	0.067	1.473	0.973	2.229	0.338	1.255	0.788	1.999
Higher ECOG performance score	0.202	1.399	0.836	2.341	0.090	1.622	0.928	2.835
BM blasts $\geqslant$ 42.1%	0.160	1.458	0.862	2.468	0.028	2.154	1.088	4.262
CR at first induction	0.137	0.669	0.394	1.136	0.868	0.936	0.428	2.045
Previous allo-HSCT	0.916	0.966	0.511	1.826	0.846	1.079	0.502	2.318
Second or higher salvage therapy (vs first)	0.004	2.409	1.318	4.406	0.003	3.537	1.542	8.114

Data were presented as $P$ value, HR and 95% CI. Factors affecting OS were determined by univariate and multivariate Cox’s proportional hazards regression analysis. A $P$ value $<$ 0.05 was considered significant. Risk stratification was scored as 1-Favorable; 2-Intermediate; 3-Poor. ECOG performance score was scored as 0; 1; 2. The analysis was based on these definitions. HR, hazards ratio; CI, confidence interval; ECOG, Eastern Cooperative Oncology Group; BM, bone marrow; CR, complete remission; allo-HSCT, allogeneic hematopoietic stem-cell transplantation.

Figure 4.

OS in patients with lncRNA TUG1 high expression and patients with lncRNA TUG1 low expression. OS of patients with lncRNA TUG1 high expression was greatly lower than that in patients with lncRNA TUG1 low expression. K-M curves were used to display OS and log-rank test was conducted to compare OS between two groups. $P<$ 0.05 was considered significant. OS, overall survival; lncRNA TUG1, long non-coding RNA taurine-upregulated gene 1; K-M curves, Kaplan-Meier curves.

3.6 Analysis of factors affecting ORR

As shown in Table 3, lncRNA TUG1 high expression predicted unfavorable ORR ( $P=$ 0.030), as well as second or higher salvage therapy (vs. first salvage therapy, $P=$ 0.008). Furthermore, age $>$ 60 years ( $P=$ 0.086) was numerically correlated with unfavorable ORR, whereas no significant difference was found. In order to assess the independent predictive factors, subsequent multivariate logistic regression was executed, and lncRNA TUG1 high expression ( $P=$ 0.022) and second or higher salvage therapy (vs. first salvage therapy, $P=$ 0.023) were verified to be independent factors predicting the absence of ORR in R/R AML patients. Additionally, no other factors predicting ORR were observed.

3.7 Comparison of OS between lncRNA TUG1 high expression and low expression groups

As shown in Fig. 4, the median value of OS in patients with lncRNA TUG1 high expression was 6.0 (95% CI 3.6–8.4) months, which was remarkably reduced compared with patients with lncRNA TUG1 low expression (14.0 (95% CI 8.9–19.1) months) ( $P<$ 0.001).

3.8 Analysis of factors affecting OS

Univariate Cox’s proportional hazards regression was used to assess the factors influencing OS, which disclosed that lncRNA TUG1 ( $P<$ 0.001) was a predictive factor for worse OS, so did secondary disease (vs. primary disease, $P=$ 0.028) and secondary or higher salvage therapy (vs. first salvage therapy, $P=$ 0.004) (Table 4). Additionally, poorer risk stratification ( $P=$ 0.067) represented a trend of shorter OS. According to the subsequent multivariate Cox’s regression analysis, lncRNA TUG1 was an independent predictive factor for worse OS ( $P=$ 0.001), moreover, BM blasts $\geqslant$ 42.1% ( $P=$ 0.028) and second or higher salvage therapy (vs. first salvage, $P=$ 0.003) were also independent factors predicting poor OS.

4. Discussion

In the present study, we compared the treatment outcomes between R/R AML patients with high lncRNA expression and R/R AML patients with low lncRNA expression receiving CLAG or FLAG chemotherapy. The major results we obtained were as follows: (1) R/R AML patients possessed elevated lncRNA TUG1 expression compared with controls; (2) Patients with lncRNA TUG1 high expression achieved lower ORR and numerically decreased CR, as well as shorter OS compared to patients with lncRNA TUG1 low expression. When further evaluated by multivariate analysis, lncRNA TUG1 high expression was an independent predictive factor for the absence of ORR and reduced OS.

LncRNA is the common RNA molecule comprising 200 nucleotides to 100 kilobases in length [21, 22, 23]. Although lncRNA is not able to encode functional proteins, it plays a key role in maintaining physiological processes such as proliferation, differentiation and apoptosis of cells through various mechanisms, including: (1) reconstruction of chromatins; (2) regulation of recruiting transcription factors or co-activators; (3) negative modulation of the RNA polymerase II activity; (4) alternative mosaicking of pre-mRNAs; (5) disruption of the mRNA stability; (6) isolation of microRNAs [24, 25, 26]. Accumulating evidence has revealed that lncRNA is considered as a biomarker in predicting treatment outcomes in various cancers, such as B-cell malignancies, bladder cancer, hepatocellular carcinoma and osteosarcoma and AML [20, 27, 28]. As to AML, a previous study explores the influence of lncRNA growth arrest-specific 5 (lncRNA GAS5) expression on OS in 313 AML patients, which demonstrates that relatively higher expression of lncRNA GAS5 predicts reduced OS in AML patients compared with the relatively lower expression of lncRNA GAS5 [29]. Another study investigating the role of lncRNA HOTAIRM1 in 215 immediate-risk AML patients reveals that increased HOTAIRM1 expression (vs. lower HOTAIRM1 expression) is independently correlated with worse OS and LFS as well as increased incidence of relapse [30]. These previous studies indicate dysregulated lncRNA contributes to the disease progress as well as prognosis in AML patients.

Among cancer-related lncRNAs, lncRNA TUG1 is a rising star, which is recently identified as oncogenic lncRNA in some cancers such as B-cell malignancies, oesophageal squamous cell carcinoma, bladder cancer, hepatocellular carcinoma and osteosarcoma [20, 27, 28]. A study carried out by Tan et al. shows that lncRNA TUG1 is upregulated in bladder cancer cell line compared to normal human bladder cell line, and its downregulation suppresses the metastasis of cancer cells [31]. Also, raised lncRNA TUG1 level in osteosarcoma patients is associated with worse OS compared with lower lncRNA TUG1 level [32]. Aforementioned studies demonstrate lncRNA TUG1 is an oncogenic gene in various cancers and associates with poor prognosis. However, limited studies evaluate the predictive value of lncRNA TUG1 in AML patients. A study conducted by Isin et al. discovers that, compared with healthy controls, lncRNA TUG1 presents with remarkably higher expression in patients with multiple myeloma, a hematological malignancy, while only 62 patients with multiple myeloma are enrolled in this previous study, and whether lncRNA TUG1 high expression predicts treatment outcomes is not clarified [33]. Regarding R/R AML, no study has explored the effect of lncRNA TUG1. Hence, our study investigated the expression and prognostic effect of lncRNA TUG1 in R/R AML patients, and we discovered that lncRNA TUG1 expression was elevated in R/R AML patients, moreover, its high expression was correlated with lower ORR and numerically decreased CR as well as worse OS. The possible reasons were as follows: (1) lncRNA TUG1 bound to polycomb repressive complex 2 (PRC2), which was able to hinder the transcription of cell cycle regulatory genes, thereby decreasing malignant cellular proliferation in AML [34]; (2) overexpression of lncRNA TUG1 might contribute to chemoresistance through regulating multiple cell-survival pathways, and finally it led to worse treatment outcomes [35].

Our study also existed some limitations: (1) sample size was relatively small in this study, which just enrolled 73 R/R AML patients, so the statistical power was diminished to some extent; (2) the median follow-up duration in this study was only 10.0 months (1/4–3/4 quartiles: 4.0–15.5 months), which was relatively short, thus the effect of lncRNA TUG1 on long-term treatment outcomes was not evaluated, however, R/R AML presented with short OS, so long-term follow-up duration was hard to be executed as well; (3) detailed mechanisms of lncRNA TUG1 in AML was not investigated.

Footnotes

Acknowledgments

This study was supported by General Project Foundation for Natural Science of Sichuan Provincial Education Department (No. 17ZB0167).

Conflict of interest

None.

References

Dohner

Weisdorf

D.J.

and Bloomfield

C.D.

, Acute myeloid leukemia, N Engl J Med 373 (2015), 1136–52.

Estey

and Dohner

, Acute myeloid leukaemia, Lancet 368 (2006), 1894–907.

Martin

M.G.

Augustin

K.M.

G.L.

Welch

J.S.

Hladnik

Goyal

Tiwari

Monahan

R.S.

Reichley

R.M.

Cashen

A.F.

Stockerl-Goldstein

Westervelt

Abboud

C.N.

Dipersio

J.F.

and Vij

, Salvage therapy for acute myeloid leukemia with fludarabine, cytarabine, and idarubicin with or without gemtuzumab ozogamicin and with concurrent or sequential G-CSF, Am J Hematol 84 (2009), 733–7.

Ramos

N.R.

C.C.

Karp

J.E.

and Hourigan

C.S.

, Current approaches in the treatment of relapsed and refractory acute myeloid leukemia, J Clin Med 4 (2015), 665–95.

Walter

R.B.

Othus

Burnett

A.K.

Lowenberg

Kantarjian

H.M.

Ossenkoppele

G.J.

Hills

R.K.

Ravandi

Pabst

Evans

Pierce

S.R.

Vekemans

M.C.

Appelbaum

F.R.

and Estey

E.H.

, Resistance prediction in AML: Analysis of 4601 patients from MRC/NCRI, HOVON/SAKK, SWOG and MD Anderson Cancer Center, Leukemia 29 (2015), 312–20.

Breems

D.A.

Van Putten

W.L.

Huijgens

P.C.

Ossenkoppele

G.J.

Verhoef

G.E.

Verdonck

L.F.

Vellenga

De Greef

G.E.

Jacky

Van der Lelie

Boogaerts

M.A.

and Lowenberg

, Prognostic index for adult patients with acute myeloid leukemia in first relapse, J Clin Oncol 23 (2005), 1969–78.

Kim

Lee

J.H.

Joo

Y.D.

Bae

S.H.

Lee

J.H.

Kim

D.Y.

Lee

W.S.

Ryoo

H.M.

J.C.

Choi

Lee

K.H.

and CoOperative Study Group

A.F.H.

, A prospective, multicenter phase II study of continuous infusion of FLAG for patients older than 60 yr with resistant acute myeloid leukemia: A comparison with intensive younger patients’ trial, Eur J Haematol 96 (2016), 188–97.

Luger

S.M.

, How can one optimize induction therapy in AML? Best Pract Res Clin Haematol 30 (2017), 301–305.

Delia

Pastore

Carluccio

Pasciolla

Ricco

Rossi

A.R.

Casieri

Mestice

Albano

and Specchia

, FLAG-Ida regimen as bridge therapy to allotransplantation in refractory/relapsed acute myeloid leukemia patients, Clin Lymphoma Myeloma Leuk 17 (2017), 767–773.

10.

Mirza

A.S.

Lancet

J.E.

Sweet

Padron

Pinilla-Ibarz

Nardelli

Cubitt

List

A.F.

and Komrokji

R.S.

, A phase II study of CLAG regimen combined with imatinib mesylate for relapsed or refractory acute myeloid leukemia, Clin Lymphoma Myeloma Leuk 17 (2017), 902–907.

11.

Lee

S.R.

Yang

D.H.

Ahn

J.S.

Kim

Y.K.

Lee

J.J.

Choi

Y.J.

Shin

H.J.

Chung

J.S.

Cho

Y.Y.

Chae

Y.S.

Kim

J.G.

Sohn

S.K.

and Kim

H.J.

, The clinical outcome of FLAG chemotherapy without idarubicin in patients with relapsed or refractory acute myeloid leukemia, J Korean Med Sci 24 (2009), 498–503.

12.

Virchis

Koh

Rankin

Mehta

Potter

Hoffbrand

A.V.

and Prentice

H.G.

, Fludarabine, cytosine arabinoside, granulocyte-colony stimulating factor with or without idarubicin in the treatment of high risk acute leukaemia or myelodysplastic syndromes, Br J Haematol 124 (2004), 26–32.

13.

NCCN Clinical Practice Guidelines in Oncology: Acute Myeloid Leukemia (Version 1. 2016).

14.

Duan

M.H.

Zhang

Han

Zhang

Yang

Feng

Zhang

Tian

L.P.

Zhang

and Zhou

D.B.

, Efficacy and safety analysis of the combination of cladribine, cytarabine, granulocyte colonystimulating factor (CLAG) regime in patients with refractory or relapsed acute myeloid leukemia, Zhonghua Xue Ye Xue Za Zhi 37 (2016), 571–5.

15.

Zheng

Chen

Cai

Shi

Lin

and Chen

, Overexpression of long non-coding RNA HOTAIR predicts a poor prognosis in patients with acute myeloid leukemia, Oncol Lett 10 (2015), 2410–2414.

16.

Cao

Xiao

P.F.

Tao

Y.F.

S.Y.

Zhao

W.L.

Z.H.

Wang

N.N.

Wang

Feng

Chai

Y.H.

Pan

and Gu

G.X.

, Microarray profiling of bone marrow long non-coding RNA expression in Chinese pediatric acute myeloid leukemia patients, Oncol Rep 35 (2016), 757–70.

17.

Zhou

Wang

and Zhang

, An emerging understanding of long noncoding RNAs in kidney cancer, J Cancer Res Clin Oncol 140 (2014), 1989–95.

18.

Cai

Xue

Wang

Shang

and Liu

, The long noncoding RNA TUG1 regulates blood-tumor barrier permeability by targeting miR-144, Oncotarget 6 (2015), 19759–79.

19.

Rapicavoli

N.A.

and Blackshaw

, New meaning in the message: Noncoding RNAs and their role in retinal development, Dev Dyn 238 (2009), 2103–14.

20.

Shen

Chan

M.T.

and Wu

W.K.

, TUG1: A pivotal oncogenic long non-coding RNA of human cancers, Cell Prolif 49 (2016), 471–5.

21.

Wang

K.C.

Yang

Y.W.

Liu

Sanyal

Corces-Zimmerman

Chen

Lajoie

B.R.

Protacio

Flynn

R.A.

Gupta

R.A.

Wysocka

Lei

Dekker

Helms

J.A.

and Chang

H.Y.

, A long noncoding RNA maintains active chromatin to coordinate homeotic gene expression, Nature 472 (2011), 120–4.

22.

Chen

L.L.

, Linking long noncoding RNA localization and function, Trends Biochem Sci 41 (2016), 761–72.

23.

Kang

Shi

and Chen

, Prognostic value of the long noncoding RNA HOTTIP in human cancers, Oncotarget 8 (2017), 59563–59569.

24.

Sayad

Hajifathali

Hamidieh

A.A.

Roshandel

and Taheri

, HOTAIR long noncoding RNA is not a biomarker for acute myeloid leukemia (AML) in Iranian patients, Asian Pac J Cancer Prev 18 (2017), 1581–1584.

25.

Miller

C.R.

Ruppert

A.S.

Fobare

Chen

T.L.

Liu

Lehman

Blachly

J.S.

Zhang

Lucas

D.M.

Grever

M.R.

Tallman

M.S.

Flinn

I.W.

Rassenti

L.Z.

Kipps

T.J.

Sampath

Coombes

K.R.

and Hertlein

E.K.

, The long noncoding RNA, treRNA, decreases DNA damage and is associated with poor response to chemotherapy in chronic lymphocytic leukemia, Oncotarget 8 (2017), 25942–25954.

26.

and Li

, Long non-coding RNA HOTAIR: A novel oncogene (Review), Mol Med Rep 12 (2015), 5611–8.

27.

Yin

Yan

Song

Zhu

Zhao

Shen

Huang

Zhu

Orkin

S.H.

and Shen

, Opposing roles for the lncRNA haunt and its genomic locus in regulating HOXA gene activation during embryonic stem cell differentiation, Cell Stem Cell 16 (2015), 504–16.

28.

and Li

, Long non-coding RNA TUG1: A novel therapeutic target in small cell lung cancer, J Thorac Dis 9 (2017), E644–E645.

29.

Yan

Zhang

D.Y.

Yuan

X.Q.

Yang

Y.L.

Zhu

K.W.

Zeng

X.L.

Cao

Zhou

H.H.

Zhang

and Chen

X.P.

, Long non-coding RNA GAS5 polymorphism predicts a poor prognosis of acute myeloid leukemia in Chinese patients via affecting hematopoietic reconstitution, Leuk Lymphoma 58 (2017), 1948–1957.

30.

Diaz-Beya

Brunet

Nomdedeu

Pratcorona

Cordeiro

Gallardo

Escoda

Tormo

Heras

Ribera

J.M.

Duarte

de Llano

M.P.

Bargay

Sampol

Nomdedeu

Risueno

R.M.

Hoyos

Sierra

Monzo

Navarro

Esteve

and Cooperative

A.M.L.G.C.

, The lincRNA HOTAIRM1, located in the HOXA genomic region, is expressed in acute myeloid leukemia, impacts prognosis in patients in the intermediate-risk cytogenetic category, and is associated with a distinctive microRNA signature, Oncotarget 6 (2015), 31613–27.

31.

Tan

Qiu

and Liang

, Double-negative feedback loop between long non-coding RNA TUG1 and miR-145 promotes epithelial to mesenchymal transition and radioresistance in human bladder cancer cells, FEBS Lett 589 (2015), 3175–81.

32.

Zhang

Huang

Lei

J.B.

G.H.

Liu

C.X.

Lai

Q.W.

Chen

Q.Q.

and Wang

Y.L.

, Upregulation of long non-coding RNA TUG1 correlates with poor prognosis and disease status in osteosarcoma, Tumour Biol 37 (2016), 4445–55.

33.

Isin

Ozgur

Cetin

Erten

Aktan

Gezer

and Dalay

, Investigation of circulating lncRNAs in B-cell neoplasms, Clin Chim Acta 431 (2014), 255–9.

34.

Khalil

A.M.

Guttman

Huarte

Garber

Raj

Rivea Morales

Thomas

Presser

Bernstein

B.E.

van Oudenaarden

Regev

Lander

E.S.

and Rinn

J.L.

, Many human large intergenic noncoding RNAs associate with chromatin-modifying complexes and affect gene expression, Proc Natl Acad Sci USA 106 (2009), 11667–72.

35.

Jiang

Wang

Sun

Ren

Sheng

Gao

Wang

and Yu

, High TUG1 expression is associated with chemotherapy resistance and poor prognosis in esophageal squamous cell carcinoma, Cancer Chemother Pharmacol 78 (2016), 333–9.

Long non-coding RNA taurine-upregulated gene 1 correlates with unfavorable prognosis in patients with refractory or relapsed acute myeloid leukemia treated by purine analogue based chemotherapy regimens

Abstract

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Methods

2.1 Participants

2.2 Sample collection

2.3 Quantitative polymerase chain reaction (qPCR) assay for lncRNA TUG1 expression

2.4 Treatments

2.5 Assessments

2.6 Definitions

Table 1 Baseline characteristics of R/R AML patients

3. Results

3.1 Baseline characteristics

3.4 Treatment response in patients with high lncRNA TUG1 expression and patients with low lncRNA TUG1 expression

3.5 Analysis of factors affecting CR

Table 4 Factors affecting OS by Cox’s proportional hazards regression analysis

3.7 Comparison of OS between lncRNA TUG1 high expression and low expression groups

3.8 Analysis of factors affecting OS

4. Discussion

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Baseline characteristics of R/R AML patients

Table 4
Factors affecting OS by Cox’s proportional hazards regression analysis