This article is mentioned in:

Introduction

In Southeast China, nasopharyngeal carcinoma (NPC) is one of the most common malignancies of the head and neck that can be effectively treated by radiotherapy (1,2). However, a high proportion of patients with NPC exhibit radioresistance, which is the main risk factor that contributes to poor prognosis (3). Studies have revealed that increased radioresistance may be associated with various factors that participate in tumor development (4). Thus, the molecular mechanisms of radioresistance should be understood to provide opportunities for enhancing radiosensitivity and to develop a more effective anticancer strategy of NPC radiotherapy (5).

Curcumin (diferuloylmethane; Cur), a polyphenol from Curcuma longa rhizomes, is the major constituent of the yellow spice turmeric, a flavoring agent commonly used in Asian cooking (6). Cur also inhibits proliferation and angiogenesis in tumor cells to induce apoptosis or cell cycle arrest and cause tumor regression in pre-clinical models (7–9). In NPC, Cur has potent antitumor activity and radiosensitivity (10,11); however, the exact molecular mechanism remains unclear.

Long non-coding RNAs (lncRNAs) are non-protein-coding transcripts that are longer than 200 nucleotides (12). They are pervasively transcribed with spatially and temporally regulated expression patterns (13). lncRNAs have important functions in gene expression regulation, dosage compensation, genomic imprinting, nuclear organization and compartmentalization, and nuclear-cytoplasmic trafficking (14–18). lncRNAs have continuously emerged as new contributing factors in cancer because they are involved in diverse biological processes and aberrantly expressed in various human cancers (17). lncRNAs also have potential functions in oncogenic (18) and tumor-suppressive pathways (19). lncRNAs also regulate gene expression at transcriptional, post-transcriptional, and epigenetic levels (20–22). Altered lncRNA expression may potentially enhance oncogenesis by altering some of these functions (14,23). The differential lncRNA expressions can also indicate disease progression and function as predictors of patient outcomes.

In the present study, we demonstrated that Cur enhanced the radiosensitivity in NPC cell line CNE2 at an appropriate MTT concentration or with a clonogenic survival test. To determine the mechanism of radiosensitization, we performed a chip assay for CNE2 treated with irradiation (IR) and/or Cur. Numerous differentially expressed lncRNAs were identified, in which six lncRNAs were verified by qPCR. We observed that this response altered by IR was reversed by Cur in NPC cells. Our findings provide novel information on lncRNA expression profiles, in which Cur protected the cells from radiation toxicity, suggesting that this natural product may be an effective radiosensitizer or radioenhancer for managing patients with NPC.

Materials and methods

Cell culture

The present study was performed in human NPC cell lines [CNE-2; obtained from Sun Yat-sen University and had been described before (24)]. CNE-2 was maintained in Roswell Park Memorial Institute 1640 medium (RPMI-1640) supplemented with 10% fetal bovine serum (FBS; Invitrogen, USA) at 37°C in 5% carbon dioxide. Cur (Sigma-Aldrich, USA) was dissolved in 0.5% dimethyl sulphoxide (Sigma-Aldrich) and diluted with RPMI-1640 medium to the desired concentrations before use. The cells were divided into three groups [control group (CN); IR group (CX); and IR + Cur group (JX)] and irradiated linearly with X-rays at 6 MV to deliver the indicated doses (2 Gy) at room temperature. The compensators used were 1.5 cm bolus. For the microarray, the sample was pooled in each group and the experiment was performed in triplicate.

Isolation of RNA

Total RNA was extracted using the TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. The RNA integrity number was checked to inspect RNA integration by an Agilent Bioanalyzer 2100 (Agilent Technologies, USA). Qualified total RNA was further purified using RNeasy mini kit (Qiagen, Germany) and RNase-free DNase set (Qiagen).

Preparation of array hybridization

The SBC 8×60K human lncRNA microarrays were custom designed using the Agilent eArray program according to the manufacturer’s recommendations (https://earray.chem.agilent.com/earray). The microarray contained 31,171 mRNA probes, which were derived from the probe sequence for mRNA in Agilent 8×60K Whole Human Genome Oligo Microarray, and 29,971 lncRNA probes, which were designed by using an eArray-based system. The lncRNA sequence was derived from six databases, including LNCRNA-DB, NCBI_refseq, Ensembl, UCSC, NCBI_unigene, and ncRNASCAN. After purification of labeled cRNAs, each slide was hybridized and washed according to the manufacturer’s instructions (Agilent Technologies).

Data anaysis

Raw data were normalized by quantile algorithm on the Gene Spring 11.0 software (Agilent Technologies). lncRNAs and mRNAs with ‘Present’ or ‘Marginal’ (All Targets Value) flags in all of the groups were further subjected to data analysis. Differentially expressed lncRNAs and mRNAs were identified by fold change. Clustering was analyzed using the multi-experimental viewer (MeV) 4.6 and functional enrichment analysis was performed using DAVID’s Functional Annotation Tool (http://david.abcc.ncifcrf.gov) (25).

Confirmation test of real-time quantitative RT-PCR

Total RNAs from tissues were extracted using TRIzol (Invitrogen) according to the manufacturer’s instructions. Qualified total RNA was further purified by RNase-free DNase set (Qiagen). Reverse transcription was performed using a gene-specific primer and quantification was performed using the Quantitect SYBR Green PCR kit (Stratagene, USA) with an MX3005P multiplex quantitative PCR (qPCR) system (Stratagene) according to the manufacturer’s instructions. GAPDH, the human housekeeping gene, was used for normalization. The relative lncRNA expression levels were calculated using the comparative ΔΔCt method as previously described (26). The fold changes were calculated according to 2−ΔΔCt equation. All of the primers used are listed in Table I.

Table I. Primers for QPCR or cloning.

Table I.

Primers for QPCR or cloning.

Gene	Primer (5′-3′)	Product (bp)
GAPDH	Forward: ATCATCAGCAATGCCTCCTG	102
	Reverse: ATGGACTGTGGTCATGAGTC
AF086415	Forward: AGCGCGACTTCTCTGTCTCT	115
	Reverse: GCAGAGGAGGAGACGCTGA
AK095147	Forward: ACGAGTGACCGAAGCTGAAC	117
	Reverse: GCACCATCCAGAGGGATTTA
RP1-179N16.3	Forward: CGCGTTAGGAGATTCTGGAG	105
	Reverse: AGGGTGGATACAGGCTCCTT
AK056098	Forward: GGCCTCGGGGTAGAACTTAC	143
	Reverse: CAAGCCTCCTGGTCTTTCTG
AK294004	Forward: GTGCAACCAGAAATGCACAG	168
	Reverse: ACGCTTTGTCTGTCGTGATG
MUDENG	Forward: ACTTTGTGGCACCGTGAGAT	194
	Reverse: GGCCCACTAAATGCAGAGTC
CCND1	Forward: GATCAAGTGTGACCCGGACT	129
	Reverse: TCCTCCTCCTCTTCCTCCTC
Primer A	Forward: gattatagaTGTAATTCTTGTAATTTTT
	Reverse: gattatagaGCAGCAAACAATGTGAAAGA
Primer B	Forward: gattatagaTCAACCATCCTGGCTGCGGCGT
	Reverse: gattatagaTGCCGGTTACATGTTGGTGCT
Primer C	Forward: gattatagaTGTAATTCTTGTAATTTTT
	Reverse: gattatagaTGCCGGTTACATGTTGGTGCT
Primer D	Forward: gatggtaccGCCGGTTACATGTTGGTGCT
	Reverse: gatctcgagTGTAATTCTTGTAATTTTT

RNA interference

In all, 20–30% confluent CNE2 cells were transfected with 50 nM of siRNAs using Lipofectamine 2000 (Invitrogen) following the manufacturer’s direction. Two individual small interfering RNA (siRNAs) and scrambled negative control siRNA (siRNA-NC) were obtained from Invitrogen. The target sequences of AK294004 are the following: siRNA-1, 5-CUCCCUUCAACACUUCCUAAAUA-3 and siRNA-2, 5-AGCAGCAAACAAUGUGAAAGAGA-3. Thirty-six hours after transfection, cells were harvested for qRT-PCR (27).

Luciferase reporter assay

A 1,334-bp (1,794–3,127) fragment of CCND1 3′UTR containing whole complementry sequences of AK294004 two exons was amplified using the primer pairs A, the sequence (1,794–2,486 bp) of CCND1 3′UTR containing the complementry sequence of AK294004 exon 2 was amplified using the primer pairs B, and the sequence (2,876–3,127 bp) of CCND1 3′UTR containing the complementry sequence of AK294004 exon 1 was amplified using the primer pairs C (Table I). Each fragment was respectively cloned downstream of the Renilla luciferase gene at the XbaI site in the pGL-3 promoter plasmid (Promega, USA). The entire 497-bp fragment of AK294004 was amplified using the the primer pairs D and was cloned at the KpnI and XhoI sites in the pcDNA3.1+ plasmid (Promega). The pGL3 constructs were designated as pGL3-W (whole sequence), pGL3-E1 (completed to exon 1) and pGL3-E2 (completed to exon 2) and the pcDNA3.1 construct was designated as pcDNA3-AK.

To facilitate cloning into each expression plasmid, the primers were designed to incorporate XbaI, KpnI and XhoI sites at the 5′ end (underlined in the primers above). HEK293 cells were co-transfected with 30 pmol of either pcDNA3-AK or pcDNA3-NC (empty vector control) and each pGL-construct using Lipofectamine 2000 (Invitrogen). Transfection efficiency was normalized by co-transfection with a firefly luciferase expressing plasmid. Luciferase activity was measured using the Promega dual-luciferase assay kit, in accordance with the instructions of the manufacturer. Relative protein levels were expressed as Renilla/firefly luciferase ratios. Each transfection was repeated twice in triplicates (27).

Results

LncRNA and mRNA microarray data

Array hybridization was performed using the SBC 8×60K human lncRNA microarrays. After quantile normalization of the raw data, the expression profiles of 29,971 lncRNAs and 31,171 mRNAs were obtained from the cells in the three groups. We identified differentially expressed genes among the matched groups with a fold change >2. Table II summarizes the differentially expressed genes in each group.

Table II. Summary of differently expressed genes.

Table II.

Summary of differently expressed genes.

	CX vs. CN	JX vs. CN	JX vs. CX
Up	865 (592)	623 (445)	734 (615)
Down	777 (712)	579 (588)	859 (832)
Total	1,642 (1,304)	1,202 (1,033)	1,593 (1,447)

[i] CN, normal control; CX, ionizing radiation (IR); JX, IR + curcumin. LncRNA gene number (mRNA gene number).

Altered and reversed lncRNAs and mRNA expression

After clustering analysis was performed, JX and CN revealed seemingly similar expression signatures. The lncRNAs were altered by IR and Cur reversed this response and a similar trend was observed in the mRNA expression profile. Fig. 1 shows the heat maps of the expression ratios of lncRNAs and mRNAs among the JX, CX and CN groups. In addition, we focused on those altered expression genes induced by irradition while reversed by curcumin by a stronger raw signal screening. We obtained 116 lncRNAs, in which 76 were upregulated and 40 were downregulated in the CX group compared with those in the CN group. lncRNAs in the JX group were completely or partially reversed. We used the same screening method and obtained 178 differentially expressed mRNAs, in which 59 mRNAs were upregulated and 119 mRNAs were downregulated. Functional enrichment analysis of these differentially expressed genes was performed and clustered by DAVID’s Functional Annotation Chart (25). Fig. 2 shows the functional annotation terms of the analysis.

Figure 1. Heat maps of the expression ratios (log2 scale) of lncRNAs (A) and mRNAs (B) in different groups of CNE2 cells. Red and green indicate high and low relative expressions, respectively. CN, control group; CX, IR group; and JX, IR + Cur group.

Figure 2. Functional enrichment analysis on IR-induced and Cur-reversed differentially regulated genes. Annotation of 116 lncRNAs (A) and 178 mRNAs (B) was performed by utilizing the DAVID Functional Annotation Chart (25).

Confirmation of some differentially expressed lncRNAs

We performed qPCR assays to confirm the expression patterns of some differentially expressed lncRNAs. qPCR results were consistent with the microarray analysis results of six lncRNAs (AF086415, AK095147, RP1-179N16.3, MUDENG, AK056098 and AK294004) in terms of regulation direction and significance. In particular, 0.29-fold downregulation in CX and 0.78-fold reversal in JX were observed in AF086415 (0.38- and 0.59-fold in microarray analysis, respectively). For AK095147, 0.31-fold downregulation in CX and 0.92-fold of reversal in JX were observed (0.44- and 1.04-fold in microarray analysis, respectively). For RP1-179N16.3, 5.25-fold upregulation in CX and 1.54-fold reversal in JX were observed (6.96- and 1.04-fold in microarray analysis, respectively). For MUDENG, 4.01-fold upregulation in CX and 1.34-fold reversal in JX were observed (4.98- and 0.85-fold in microarray analysis, respectively). For AK056098, 3.81-fold upregulation in CX and 1.08-fold reversal in JX were observed (3.14- and 1.16-fold in microarray analysis, respectively). For AK294004, 2.88-fold upregulation in CX and 1.21-fold reversal in JX were observed (2.86- and 0.91-fold in microarray analysis, respectively. All of the fold changes were compared with CN (Fig. 3).

Figure 3. Comparison between microarray data and qPCR results. AF086415, AK095147, RP1-179N16.3, MUDENG, AK056098 and AK294004 differentially expressed in CX and reversed in JX by microarray (B) were validated by qPCR (A). The heights of the columns in the chart represent the fold changes. The columns below the x-axis represent downregulation. qPCR results show mean ± SD (n=3; t-test; *P<0.05, compared with CN).

Effect of interaction between AK294004 and 3′UTR of CCND1

AK294004 exhibited a negative effect on CCND1. For AK294004, 2.88-fold upregulation in CX and 1.21-fold reversal in JX were observed by qPCR, and for CCND1, 0.26-fold downregulation in CX and 0.61-fold reversal in JX were observed (compared with CN, Fig. 4).

Figure 4. Negative relationship between AK294004 and CCND1 expression by qPCR. The heights of the columns in the chart represent the relative fold changes. Results show mean ± SD (n=3; t-test; *P<0.05, compared with CN). No absolute relationship exists between lncRNA and mRNA columns.

To investigate the functional effects of AK294004 in NPC cells, we modulated its expression through RNA interference and overexpression experiments. pCDNA3-AK and two individual AK294004 siRNAs were transfected into CNE2 cells. qPCR analysis of AK204004 and CCND1 levels was performed 36 h post-transfection. As shown in Fig. 5, for overexpression experiments, AK204004 expression was increased 3.16-fold while CCND1 expression was decreased by 46% in pCDNA3-AK cells, compared with control cells (pCDNA3-NC). For RNA interference experiments, when compared with control cells (siRNA-NC), AK294004 expression was knocked down 67% by siRNA-2, and 75% by siRNA-1, while CCND1 expression was increased 1.34- and 1.48-fold, respectively.

Figure 5. Modulated expression levels of AK294004 and CCND1 by qPCR. pCDNA3-AK and two individual AK294004 siRNAs were transfected into CNE2 cells. qPCR analysis of AK204004 and CCND1 levels was performed 36 h post-transfection. Data represent mean ± SD of the three independent experiments and each experiment was performed in triplicate. *P<0.05.

A luciferase-based reporter was constructed to evaluate the effect of AK294004 direct binding to the putative target sites on the 3′UTR of CCND1. To substantiate the assumption that AK294004 can directly repress CCND1, the reporter construct pGL3-vector or pGL3-W, pGL3-E1 and pGL3-E2 was co-transfected with pcDNA3-AK and pcDNA3-NC to HEK293 cells. Luciferase activity was then assayed. As shown in Fig. 6, for pGL3-W or pGL3-E1 construct, pcDNA3-AK significantly lowered luciferase activity compared with pcDNA3-NC. There was no different luciferase activity observed between the pGL3-vector and pGL3-E2 constructs.

Figure 6. Effect of the putative lncRNA binding site derived from the CCND1 3′UTR on luciferase expression. (A) Schematic of the potential AK294004 binding site containing CCND1 3′UTR. (B) Luciferase activity in HEK293 cells transfected with pcDNA3-NC and pcDNA3-AK along with pGL3-vector or pGL3-W, pGL3-E2 and pGL3-E2 36 h post-transfection. Luciferase activity was measured using the Promega dual-luciferase assay kit, according to the instructions of the manufacturer. Data represent mean ± SD of the three independent experiments and each experiment was performed in triplicate. *P<0.05.

These findings support the hypothesis that lncRNA AK294004 directly targets CCND1 expression by its exon 1 part, but not the exon 2 part, thus leading to the decreased CCND1 expression through some inhibition mechanism.

Discussion

Radiotherapy is considered one of the most effective treatments for patients with NPC, and radioresistance is the main risk factor that contributes to poor prognosis (2). Radioresistance occurs in primary IR treatment and the survived cells may be more resistant to the second IR treatment, thereby leading to the failure of radiotherapy (2,28,29). In this regard, the exact molecules and signaling pathway involved in radiosensitization should be determined to develop target therapy and enhance the efficacy of radiation. In this study, we observed that IR-induced differentially expressed lncRNAs were almost reversed by Cur. This result is consistent with our hypothesis, in which Cur enhances radiosensitivity through the reversal of effective molecules (7,30,31). For example, AK294004, a natural antisense lncRNA, exhibited 2.86-fold upregulation in CX (compare with CN), whereas a reversal at 0.32-fold downregulation by Cur was observed in JX (compare with CX). This finding was further confirmed by qPCR. lncRNA may have an important function in IR-induced radioresistance.

Cur regulates the gene expression involved in survival, proliferation, angiogenesis, invasion and metastasis. This phytochemical also modulates various mechanisms that are associated with radioresistance, including the following: downregulating COX-2, MRP, Bcl-2, and survivin expression; inhibiting PI3K/AKT activation; suppressing growth factor signaling pathways; and inhibiting STAT3 activation (32–34). In this study, we demonstrated that Cur enhanced radiosensitivity in the NPC cell line CNE2 at 10 μmol/l by MTT or clonogenic survival test (35) before we performed the array test (data not shown), although Cur exhibited higher anti-proliferative effects when used alone at a concentration of 20 or 40 μmol/l. Considering the cytotoxicity of Cur and IR, a concentration of 10 μmol/l was more suitable as a radioenhancer. Therefore, no significant data were obtained by conjoint analysis with other groups at particular time-points, although the array of JN (Cur group) was performed (data not shown). Further analysis need to be performed to reveal other chemical mechanisms for Cur. Furthermore, the optimal IR dosage of 2 Gy and the Cur pretreatment time of 6 h were confirmed for the succeeding study (data not shown).

The mammalian genome clearly encodes numerous lncRNAs that are highly conserved and biologically functional (36). Expression patterns have suggested that these lncRNAs are involved in diverse biological processes, including cell cycle regulation, innate immunity, and pluripotency (37), but current understanding on the functions of lncRNAs is limited. In this study, 116 differentially expressed lncRNAs were expressed site-specifically, such as intergenic, intronic antisense, natural antisense, bidirectional, and intron sense overlapping. We used the DAVID Functional Annotation Chart (25) for the functional enrichment analysis of these differentially expressed genes. In this study, the most significant functional annotation terms of 116 lncRNAs were transcription regulation, DNA binding, transcription factor binding, activator and nucleus (Fig. 2A). For 178 mRNAs, the functional annotation terms were cell cycle, organelle fission, nuclear division, mitosis, mitotic cell cycle and M phase of the mitotic cell cycle (Fig. 2B). No direct relationship was found between the altered lncRNA and mRNA expressions, indicating that lncRNA performed a biological function via a complex regulatory mechanism instead of directly targeting mRNA during the Cur-induced radiosensitization involved in NPC.

AK294004, a natural antisense lncRNA that completely complements the terminal end of the 3′ untranslated region of CCND1 mRNA, exhibited a negative effect on CCND1, an important molecule of the cell cycle and DNA repair. CCND1 is downregulated during IR-induced DNA damage (17,38). In this study, we observed the IR-induced altered regulation and the Cur-induced reversal of either AK294004 or CCND1 that were consequently confirmed by qPCR (Fig. 4). Moreover, luciferase reporter assay and modulated expression experiments indicated that CCND1 might be a direct target of AK294004, however, it needed to be further determined how these lcnRNAs and mRNAs interact with one another.

In general, the cells respond to IR-induced biological process, such as DNA damage repair, cell cycle arrest, and so on (33,39–41). In this study, we performed the microarray assay at 3 h post-IR. We also performed qPCR to validate the altered lncRNA expression at different time-points until 48 h post-IR was reached in parallel cell groups (data not shown). The microarray results were consistent with the qPCR data, particularly at 3–12 h checkpoint but slightly differed after 24 h. These differences in responses may be attributed to different mechanisms of lncRNA performance.

In conclusion, we demonstrated the mechanism by which Cur enhanced radiosensitivity in NPC cells that involved differentially expressed lncRNAs and provided better understanding of chemically-mediated radiosensitization. The function of Cur-induced lncRNA reversal should be fully understood to provide a new and more effective radiotherapeutic treatment for patients with NPC by using natural products.

Acknowledgements

This study was supported by National Natural Science Foundation of China (grant no. 81173616) and New Technology Project of Nanfang Hospital (grant no. 201103). We thank medical personnel Jiabin Liu and Huarui Niu of NanFang Hospital for providing X-ray radiation equipment.

References