Introduction

Esophageal cancer is one of the most common digestive system neoplasms, especially in East Asia, and always has a poor prognosis. The current therapeutic modalities benefit little on the survival of the patients [1]. Therefore, novel strategies are highly required for esophageal cancer treatment.

TNF-related apoptosis-inducing ligand (TRAIL) is a strong apoptosis-inducing cytokine. It can recognize and bind cellular receptors, DR-4 and DR-5, and in turn activate the downstream apoptotic pathway [2]. TRAIL has been demonstrated to induce apoptosis in a wide range of cancer cells including esophageal cancer cells [3, 4]. However, many evidences also indicated that TRAIL had cytotoxicity to normal cells by inducing apoptosis [5, 6]. Thus, there is a possibility that the application of TRAIL for esophageal cancer treatment may lead to hepatotoxicity. So far, there has been no strategy that can confer TRAIL expression with specificity to esophageal cancers.

MicroRNAs (miRNAs) are a class of small noncoding RNA existing from yeast to mammals. They recognize and bind miRNA response element (MRE) that is located within the 3′ untranslated region (3′ UTR) of target messenger RNA (mRNA), and degrade the targeted mRNA molecules or inhibit their translation in cooperation with other protein complex [7]. Numerous evidences have demonstrated that the deregulation of miRNAs is closely complicated with the formation and progression of cancers [8]. For esophageal cancer, many miRNAs have been found to have differential expression levels between malignant tumors and their normal counterparts [9]. Among the miRNAs deregulated in esophageal cancer, miR-143 has been intensively studied for its role in the biology of tumor cells. miR-143 has been identified to be downregulated in esophageal cancers [10, 11] and suppress the malignant phenotypes of esophageal cancer cells by targeting multiple oncogenes [12, 13].

In this study, we constructed an adenoviral vector that carries MRE of miR-134 and miR-122 to restrict TRAIL expression in esophageal cancer cells and to prevent liver tissues from TRAIL action (miR-122 is a liver-enriched miRNA). Subsequent experiments confirmed that this strategy confers TRAIL expression with selectivity to esophageal cancers.

Materials and methods

Cell culture

Human esophageal cancer cell lines, TE-1, TE-10, TE-11, and Eca109, and human normal esophagus epithelial cells, Het-1A, normal lung fibroblast, MRC-5, and normal human liver cells, L-02, were purchased from the Shanghai Cell Collection (Shanghai, China). Human embryonic kidney cell line HEK293 was obtained from Microbix Biosystems (Toronto, ON, Canada). The cells were cultured in the recommended media supplemented with 10 % fetal bovine serum (FBS; GIBCO-BRL) at 37 °C under a 5 % CO2 condition.

Constructions of plasmids and adenoviruses

This study involved two plasmids (psiCheck2 and psiCheck2-143-122) and three adenoviral vectors (Ad-EGFP, Ad-TRAIL, and Ad-TRAIL-143-122). Their structures were schemed in Fig. 1.

Fig. 1
figure 1

The structures of plasmids and adenoviruses. The structures of the involved plasmids and adenoviruses were schemed here. Cytomegalovirus early promoter (CMV) early promoter was used in all of the constructs to drive the expression of the inserted genes such as luciferase, EGFP, and TNF-related apoptosis-inducing ligand (TRAIL). The grey and black bars are MREs of miR-143 and miR-122, respectively. psiCheck2-143-122, a luciferase expression vector under the regulation of miR-143 and miR-122, was generated based on psiCheck2 vectors. In adenoviral vectors, arrows indicated the inverse terminal repeat (ITR) sequences of adenovirus. The details of construction processes were described in “Materials and methods” section

Full size image

psiCheck2, Ad-EGFP, and Ad-TRAIL were kindly gifted from Dr. Zhao Youguang (Department of Urology, Chengdu Military Area General Hospital, Chengdu, China). psiCheck2-143-122 was constructed as the following protocols. A linker that contains two copies of MREs of miR-143 and miR-122 (CTCGAGGATATCACAAACACCTCATCTCACAAACACCTCATCTCACAAACACCACACTCCACAAACACCACACTCCACAAACACCGATATCGCGGCCGC) was synthesized, digested, and inserted into a psiCheck2 vector by XhoI and NotI sites, generating psiCheck2-143-122. Ad-TRAIL-143-122 was constructed as following. A MREs of miR-143 and miR-122-containing fragment was released from psiCheck2 by the digestion of EcoRV and inserted into the same site of pShuttle-CMV-TRAIL (also gifted by Dr. Zhao) to obtain pShuttle-CMV-TRAIL-143-122. Then pShuttle-CMV-TRAIL-143-122 and pAdEasy were co-transfected into Escherichia coli BJ5183 to obtain plasmid pAd-TRAIL-143-122. After the identification, pAd-TRAIL-143-122 was transfected into HEK293 cells for the production of recombinant adenovirus Ad-TRAIL-143-122.

The adenoviruses were harvested and purified with the CsCl gradient centrifugation method. The titers of Ad-EGFP, Ad-TRAIL, and Ad-TRAIL-143-122 were quantified through TCID50 assay on HEK293 cells.

Patients and specimens

The esophageal cancer specimens and the matched adjacent normal tissue (n = 22) were obtained from the patients undergoing esophageal carcinoma in the Department of Thoracic Surgery, the First Affiliated Hospital of Zhengzhou University, with their written consent according to the procedures approved by the Ethics Committee in Zhengzhou University (Zhengzhou, China).

Quantitative PCR

Total RNA was extracted from the specimens and cells using Trizol (Invitrogen) according to the manufacturer’s instructions.

For microRNA detection, reverse transcriptase reaction contained total RNA, 50 nM stem-loop RT primer (RNU6B P/N: 4373381, hsa-miR-143 P/N: 4373134, hsa-miR-122a P/N: 4373151, Applied Biosystems, USA), 1 × RT buffer (P/N: 4319981, Applied Biosystems), 0.25 mM each of dNTPs, 3.33 U/μl MultiScribe reverse transcriptase (P/N: 4319983, Applied Biosystems), and 0.25 U/μl RNase inhibitor (P/N: N8080119; Applied Biosystems). The 7.5 μl reactions were incubated in an Applied Biosystems 9700 Thermocycler in a 96-well plate for 30 min at 16 °C, 30 min at 42 °C, 5 min at 85 °C and then held at 4 °C. All reverse transcriptase reactions, including no-template controls and RT minus controls, were run in duplicate. Quantitative polymerase chain reaction (qPCR) was performed according to the protocols provided by the manufacturers on Applied Biosystems 7300 real time PCR System. The 10 μl PCR included 0.67 μl reverse transcription PCR product, 1 × TaqMan® Universal PCR Master Mix (P/N: 4324018, Applied Biosystems), 0.2 μM TaqMan® probe, 1.5 μM forward primer, and 0.7 μM reverse primer. The reactions were incubated in a 96-well plate at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min.

For mRNAs detection, reverse transcription PCR was performed according to the protocol of RevertAid™ First-Strand cDNA Synthesis Kits (Fermentas); qPCR was performed using SYBR premix Ex Taq (TaKaRa) and Applied Biosystems 7300 Real-Time PCR System (Applied Biosystems) supplied with analytical software. GAPDH mRNA levels were used for normalization. The oligonucleotides used as PCR primers were: TRAIL-forward: 5′-GACCTGCGTGCTGATC-3′; TRAIL-reverse: 5′-TAAAAGAAGATGACAG-3′; GAPDH-forward: 5′-TCAGTGGTGGACCTGA-3′, and GAPDH-reverse: 5′-TGCTGTAGCCAAATTC-3′.

All reactions were run for three times. Threshold cycle is defined as the fractional cycle number at which the fluorescence passes the fixed threshold.

Immunoblot assay

Total proteins were extracted from the tissues and cells with M-PER® Mammalian Protein Extraction Reagent (Thermo Scientific), separated on 10–12 % polyacrylamide gels, transferred onto 0.45 μm nitrocellulose in a buffer containing 25 mM Tris–HCl (pH 8.3), 192 mM glycine, 20 % methanol, and blocked with 5 % fat-free dry milk in PBS for 2 h. The membranes were incubated with primary antibodies and detected by the addition of anti-rabbit infrared (IR) dye 700 or antimouse IR dye 800 (Li-Cor, Nebraska, USA; 1:1000). The fluorescent signal was revealed through the Odyssey infrared imaging system (Li-Cor). The primary antibodies used in our study are described below. TRAIL antibody (Santa Cruz), cleaved caspase-3 (Asp175) antibody (Cell Signaling Technology), cleaved PARP antibody (Asp214) (Cell Signaling Technology), and GAPDH antibody (Santa Cruz).

The mimics and inhibitors of miRNAs

The mimics and inhibitors of miR-143 and miR-122 were purchased from GenePharma (Shanghai, China). miRNA mimics is a group of chemically modified double-stranded RNAs that mimic endogenous miRNAs. miRNA inhibitors are single-stranded RNAs designed to bind and suppress specific endogenous miRNA. The indicated cell lines were transfected with 200 nM control mimics (inhibitors) or the mixture of 100 nM miR-143 and 100 nM miR-122 mimics (inhibitors), using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Overnight, the cells were infected with indicated adenoviruses of 10 MOIs.

Cell viability assay

Cells (1.5 × 104) were grown in 96-well plates. Overnight, the cells were transduced with Ad-EGFP, Ad-TRAIL, or Ad-TRAIL-143-122 of indicated MOIs. After 48 h, the viability of the infected cell was measured on a model 550 microplate reader at 570 nm with a reference wavelength of 655 nm after adding 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT, 1 mg/ml). Cell viability was calculated as the following formula: Viability (percent) = absorbance value of adenovirus-infected cells / absorbance value of control cells.

Cytometric analysis of apoptotic rates

Cells (2 × 105) were cultured in each well of six-well plates. Overnight, the cells were infected with Ad-EGFP, Ad-TRAIL, or Ad-TRAIL-143-122 of 10 MOI. The cells were harvested with trypsin, fixed with 70 % cold alcohol and stained with Propidium iodide (1 μg/ml), 48 h after infection. The percentage of apoptotic cells was analyzed through flow cytometry (FAS Aria II, BD Biosciences) and modtif LT software (BD Biosciences).

Animal experiments

The procedures of animal experiment have been approved by the Committee on the Use and Care on Animals in Zhengzhou University (Zhengzhou, China). Eca109 esophageal cancer xenograft was established by subcutaneously inoculating 5 × 106 cells at the flanks of 5-week-old male BALB/c nude mice. There were 24 mice randomly divided into four groups (n = 6). The mice were intratumorally injected with PBS (100 μl), Ad-EGFP (1 × 109 pfu), Ad-TRAIL (1 × 109 pfu), and Ad-TRAIL-143-122 (1 × 109 pfu). Tumor diameters were periodically measured with calipers. Tumor volume (mm3) = maximal length (mm) × perpendicular width (mm)2 / 2.

Hepatotoxicity detection

There were 20 5-week-old male BALB/c mice randomly divided into four groups (n = 5) and intravenously injected with PBS (100 μl), Ad-EGFP (1 × 109 pfu), Ad-TRAIL (1 × 109 pfu), and Ad-TRAIL-143-122 (1 × 109 pfu) every other day for three times. After 12 days, blood (600 ml) was harvested by cardiac puncture and incubated with heparin (12 U). Alanine aminotransferase (ALT) was determined with ALT Activity Assay Kit (ab105134, Abcam).

Histological staining

Eca109 xenograft was resected from a sacrificed mouse of each group 7 days after the treatment of adenoviral vectors. Liver section was obtained from the tumor-free mice intravenously injected with different adenoviruses after they were sacrificed. These samples were fixed with 10 % formalin. Streptavidinbiotin peroxidase complex-based histological staining was performed to examine TRAIL expression with TRAIL antibody (1:150; sc-6079, Santa Cruz Biotechnology). The slides were then treated with hematoxylin (Sigma-Aldrich) to stain the nuclei.

Statistical analysis

Each experiment was performed for at least three times. All values were reported as means ± SD and compared at a given time point by unpaired, two-tailed t test. Data were considered to be statistically significant when p < 0.05 (*) and p < 0.01 (**).

Results

miR-143 and miR-122 downregulated in esophageal cancers

First of all, we investigated the expression profiles of miR-143 and miR-122 in esophageal cancer samples and cell lines, and normal tissues and cells, with qPCR assay. The results revealed that the abundances of miR-143 and miR-122 were reduced in the neoplasm specimens from esophageal cancer patients (n = 22), in comparison with normal esophageal tissues (n = 20; Fig. 2a). Furthermore, the expression levels of these two miRNA were also found to be lower in esophageal cancer cell lines, TE-1, TE-10, TE-11, and Eca109, than normal cell lines, Het-1A, MRC-5, and L-02 (Fig. 2b).

Fig. 2
figure 2

miR-143 and miR-122 level was reduced in esophageal cancer. a The expression of miR-143 and miR-122 in esophageal cancer specimens (T; n = 22) was indicated to be lower than noncancerous esophageal tissue (N; n = 20) by qPCR assay (p < 0.05). b miR-143 and miR-122 abundance was reduced in esophageal cancer cell lines, TE-1 TE-10, TE-11, and Eca109, compared with Het-1A, MRC-5, and L-02 cells (p < 0.01 or 0.05)

Full size image

MREs of miR-143 and miR-122 confer luciferase expression with specificity to esophageal cancer

The reduced abundance of miR-143 and miR-122 made it possible that the application of their MREs may suppress the expression of exogenous genes in normal cells, which highly express the two miRNAs, without affecting their expression within esophageal cancer cells. Luciferase assays indicated that luciferase activity was greatly suppressed in psiCheck2-143-122-transfected normal cell lines (Fig. 3a). However, there was no significant difference in luciferase expression between psiCheck2- and psiCheck2-143-122-transfected esophageal cancer cell lines (Fig. 3b).

Fig. 3
figure 3

MREs of miR-143 and miR-122 suppressed luciferase activity in normal cells. a The activity of luciferase was suppressed in psiCheck2-143-122-transfected normal cells, Het-1A, MRC-5, and L-02 (p < 0.01 or 0.05). Solid bars psiCheck2 and empty bars psiCheck2-143-122. b The activity of luciferase was not significantly affected in psiCheck2-143-122-transfected esophageal cancer cells, TE-1, TE-10, and Eca109 (p > 0.05). Solid bars psiCheck2 and empty bars psiCheck2-143-122

Full size image

Ad-TRAIL-143-122 selectively expresses TRAIL in esophageal cancer cells

Next, we employed qPCR and immunoblot assays to detect if Ad-TRAIL-143-122 was able to express TRAIL in esophageal cancer cells, rather than normal cell lines. In TE-1 and Eca109 cells, Ad-TRAIL-143-122 expressed TRAIL mRNA at a similar level to Ad-TRAIL (Fig. 4a). In contrast, TRAIL mRNA was almost not detected in the normal cell lines, Het-1A and L-02 cells, infected with Ad-TRAIL-143-122 (Fig. 4a). This selectivity of TRAIL expression mediated by Ad-TRAIL-143-122 was also observed at the level of proteins (Fig. 4b).

Fig. 4
figure 4

Ad-TRAIL-143-122 selectively expressed TRAIL in esophageal cancer cells. a qPCR assay revealed that Ad-TRAIL-143-122 highly expressed TRAIL mRNA in esophageal cancer cells, but not in normal cells (p < 0.05). Solid bars Ad-TRAIL, empty bars Ad-TRAIL-143-122, and grey bars Ad-EGFP. b Immunoblot analysis of TRAIL protein confirmed that TRAIL expression was suppressed in normal cells, L-02

Full size image

The expression TRAIL mediated by MRE-regulated adenovirus is regulated by the levels of miR-143 and miR-122

To confirm that TRAIL expression by this MRE-based strategy is dependent on the abundance of cellular miR-143 and miR-122, we changed the expression level of miR-143 and miR-122 in normal and esophageal cancer cells with miRNA inhibitors and mimics (Fig. 5a). Immunoblot assays revealed that TRAIL expression was restored when miR-143 and miR-122 was suppressed in L-02 cells (Fig. 5b). Consistently, the mimics of miR-143 and miR-122 lead to the reduction in TRAIL expression in esophageal cancer cells infected with Ad-TRAIL-143-122 (Fig. 5b).

Fig. 5
figure 5

Ad-TRAIL-143-122-mediated TRAIL expression was regulated by the levels of miR-143 and miR-122. a The mixed inhibitors of miR-143 and miR-122 were transfected into L-02, whereas Eca109 cells were transfected with miR-143 and miR-122 mimics. Then qPCR assays were performed to detect miR-143 and miR-122 expression levels. Solid bars control, empty bars mixed miR-143 and miR-122 inhibitors (for L-02 cells) or mimics (for Eca109 cells, p < 0.05). b The expression of TRAIL was partially restored in L-02 cells infected with Ad-TRAIL-143-122 by the suppression of miR-143 and miR-122 and was suppressed in Eca109 cells under the simultaneous treatment of Ad-TRAIL-143-122 and mimics of miR-143 and miR-122

Full size image

Ad-TRAIL-143-122 leads to selective apoptosis in esophageal cancer cells

TRAIL has been well documented to trigger apoptotic pathway in a wide range of cancer cells. Therefore, we wondered if Ad-TRAIL-143-122 specifically induced apoptosis in esophageal cancer cells. Immunoblot analysis of apoptosis-related proteins revealed that cleaved caspase 3 and PARP were highly expressed in Eca109 cells infected with both Ad-TRAIL and Ad-TRAIL-143-122 as well as Ad-TRAIL-transduced L-02 cells, but not in Ad-TRAIL-143-122-infected ones (Fig. 6a). Flow cytometric analysis also indicated that the percentages of apoptotic population were increased in normal cells infected with Ad-TRAIL, but not Ad-TRAIL-143-122 (Fig. 6b). There was no difference in apoptotic rates between Ad-TRAIL- and Ad-TRAIL-143-122-infected esophageal cancer cells (Fig. 6b).

Fig. 6
figure 6

Ad-TRAIL-143-122 induced apoptosis specifically in esophageal cancer cells. a The expression of cleaved caspase 3 and PARP was suppressed in Ad-TRAIL-143-122-transduced L-02 cells. b Cytometrical analysis suggested that apoptotic rates were much higher in esophageal cancer cell, TE-1 and Eca109, than normal cell, Het-1A and L-02, under the infection of Ad-TRAIL-143-122 (p < 0.05). Solid bars Ad-TRAIL, empty bars Ad-TRAIL-143-122, and grey bars Ad-EGFP

Full size image

Ad-TRAIL-143-122 decreased the viability of esophageal cancer cell lines without cytotoxicity to normal cells

To determine the selective TRAIL expression and its subsequent apoptosis also resulted in a selective reduction in the survival of esophageal cancer and normal cells, we employed MTT assay to evaluate the viability of cells infected with different adenoviral vectors. The data showed that Ad-TRAIL-143-122 displayed a selective cytotoxic activity to esophageal cancer cell lines, TE-1, and Eca109 (Fig. 7). In contrast, Ad-TRAIL exerted an indiscriminative cytotoxicity to normal and esophageal cancer cells (Fig. 7).

Fig. 7
figure 7

Ad-TRAIL-143-122 exerted selective cytotoxicity to esophageal cancer cells. MTT assays revealed that Ad-TRAIL-143-122 reduced the viability of esophageal cancer cells, TE-1 and Eca109, by a similar extent to Ad-TRAIL (p < 0.01 or 0.05). Solid squares Ad-TRAIL, empty squares Ad-TRAIL-143-122, and solid triangles Ad-EGFP

Full size image

Ad-TRAIL-143-122 suppressed the growth of Eca109 esophageal cancer xenograft in mice

To further confirm the anti-tumor activity of Ad-TRAIL-143-122, we established an esophageal cancer animal model by subcutaneously injecting Eca109 cells into the flanks of mice. PBS, Ad-EGFP, Ad-TRAIL, and Ad-TRAIL-143-122 were intratumorally injected followed by periodic measurement of tumor diameters. The data demonstrated that Ad-TRAIL-143-122 was able to inhibit the growth of Eca109 xenograft by an extent similar to Ad-TRAIL (Fig. 8a). Furthermore, immunohistological staining showed that TRAIL expression was detected in the tumor sections from the mice injected with Ad-TRAIL or Ad-TRAIL-143-122 (Fig. 8b).

Fig. 8
figure 8

Ad-TRAIL-143-122 suppressed the growth of Eca109 xenograft in mice. a The growth of Eca109 tumor xenografts (n = 6) was impeded by the injection of Ad-TRAIL and Ad-TRAIL-143-122, but not PBS and Ad-EGFP (p < 0.05). Empty triangles PBS, solid triangles Ad-EGFP, solid squares Ad-TRAIL, and empty squares Ad-TRAIL-143-122. b TRAIL expression was detected in tumor section from the mice injected with Ad-TRAIL and Ad-TRAIL-143-122, but not PBS and Ad-EGFP by immunohistological staining

Full size image

Ad-TRAIL-143-122 protected animal from hepatoxicity

To explore if this MRE-regulated adenoviral vector prevented liver from TRAIL expression and its induced toxicity, we employed another group of tumor-free mice. PBS, Ad-EGFP, Ad-TRAIL and Ad-TRAIL-143-122 were injected via tail vein, followed by blood harvesting for the detection of serum ALT levels. The data indicated that Ad-TRAIL induced a significant hepatotoxicity, while the other two adenoviral vectors did not affect the concentrations of ALT in mice (Fig. 9a). Immunohistological staining assay revealed that TRAIL expression was suppressed in the hepatic tissues from the mice injected with Ad-TRAIL-143-122. However, TRAIL proteins were extensively detected in the liver sections after Ad-TRAIL administration (Fig. 9b).

Fig. 9
figure 9

Ad-TRAIL-143-122 protected liver from toxicity induced by TRAIL. a Serum ALT levels were found to be suppressed in the mice injected with Ad-TRAIL-143-122, in comparison with Ad-TRAIL (p < 0.05). Empty triangles PBS, solid triangles Ad-EGFP, solid squares Ad-TRAIL, and empty squares Ad-TRAIL-143-122. b Immunohistological staining showed that the TRAIL expression was suppressed in the livers of mice injected with Ad-TRAIL-143-122

Full size image

Discussion

Lack of cancer selectivity is one of the major problems that impede the application of tumor cancer therapy. To solve this problem, many researchers have developed some strategies to confer the expression of inserted tumor suppressor genes with selectivity to cancer cells. For examples, tumor-specific promoters, such as hTERT promoter, have been used to drive the expression of inserted genes [14]. However, hTERT is a relatively week promoter, and thus the expression levels of exogenous genes under the control of hTERT promoter is quite low. The promoters originating from viruses have much higher transcription activity than tumor-specific promoters, such as CMV promoter, although they cannot discriminate between cancer and normal cells. Thus, how to simultaneously have potency and selectivity is a key issue for tumor gene therapy.

In our study, we constructed an adenoviral vector that contains an expression cassette with a CMV promoter and two copies of MREs of miR-143 and miR-122. After this modification, this adenoviral vector can highly express the genes of interest in esophageal cancer cells due to the potent activity of CMV promoter. Simultaneously, the expression of inserted genes is greatly suppressed in normal cells which express abundant miR-143 and miR-122. It appears that this strategy can confer gene therapy with efficiency and selectivity to esophageal cancer.

So far, several research groups have tested this MRE-based strategy of gene therapy for some types of cancers, such as melanoma [15], glioma [16], and bladder cancers [17]. Their data have verified the feasibility and effectiveness of the application of MREs of miRNAs that have reduced levels in cancer cells. However, the safety to normal tissues, especially liver, is limited, because MREs of miR-122 is not used in their constructions. miR-122 is a liver-enriched miRNA and plays multiple physiological roles in hepatic cells [18]. The expression of miR-122 has been identified to decline in various cancers including hepatocellular carcinoma [19]. Hence, the usage of MRE of miR-122 is expected to further prevent the expression of inserted gene in liver cells, but not to affect its expression in cancer cells. In our study, we provided evidence that MRE of miR-122 can efficiently protect hepatic tissue from toxicity induced by TRAIL-expressing adenovirus.

Collectively, we developed a strategy that applied MREs of tumor suppressor miRNAs to restrict TRAIL expression within esophageal cancer cells. The obtained data confirmed that this strategy is feasible, efficient, and biosafe. This MRE-based gene therapy may be a promising regimen against esophageal cancer.