Chicken gga-miR-103-3p Targets CCNE1 and TFDP2 and Inhibits MDCC-MSB1 Cell Migration

Marek’s disease (MD) is a highly contagious viral neoplastic disease caused by Marek’s disease virus (MDV), which can lead to huge economic losses in the poultry industry. Recently, microRNAs (miRNAs) have been found in various cancers and tumors. In recent years, 994 mature miRNAs have been identified through deep sequencing in chickens, but only a few miRNAs have been investigated further in terms of their function. Previously, gga-miR-103-3p was found downregulated in MDV-infected samples by using Solexa deep sequencing. In this study, we further verified the expression of gga-miR-103-3p among MDV-infected spleen, MD lymphoma from liver, noninfected spleen, and noninfected liver, by qPCR. The results showed that the expression of gga-miR-103-3p was decreased in MDV-infected tissues, which was consistent with our previous study. Furthermore, two target genes of gga-miR-103-3p, cyclin E1 (CCNE1) and transcription factor Dp-2 (E2F dimerization partner 2) (TFDP2), were predicted and validated by luciferase reporter assay, qPCR, and western blot analysis. The results suggested that CCNE1 and TFDP2 are direct targets of gga-miR-103-3p in chickens. Subsequent cell proliferation and migration assay showed that gga-miR-103-3p suppressed MDCC-MSB1 migration, but did not obviously modulate MDCC-MSB1 cell proliferation. In conclusion, gga-miR-103-3p targets the CCNE1 and TFDP2 genes, and suppresses cell migration, which indicates that it might play an important role in MD tumor transformation.

. The chicken and MDV-encoded miRNAs were first identified in MDV-infected chicken embryo fibroblasts (CEF) by Burnside et al. (2006). Since then, many viral miRNAs had been found, and are considered to play critical roles in MDV tumorigenesis (Morgan et al. 2008;Zhao et al. 2009Zhao et al. , 2011Luo et al. 2011;Xu et al. 2011;Hicks and Liu 2013;Z. Li et al. 2014;Yu et al. 2014;Chi et al. 2015;Teng et al. 2015). However, the functions of host miRNAs involved in chicken MD still need to be investigated (Burnside et al. 2008;Lian et al. 2012aLian et al. , 2012bLian et al. , 2015Tian et al. 2012;X. Li et al. 2014).
In a previous study (Lian et al. 2012a), we identified 187 known chicken miRNAs in MDV-infected samples by Solexa deep sequencing, among which, gga-miR-103-3p was downregulated in MDV-infected groups. In this study, we further verified its targets and conducted a preliminarily investigation into the role of gga-miR-103-3p in MDCC-MSB1 cell proliferation and migration.

Ethics statements
Animal experiments were specifically approved by the Animal Care and Use Committee of China Agricultural University (Approval ID: XXCB-20090209), and this study was carried out in strict accordance with the regulations and guidelines established by this committee. The animals were killed using T61 intravenously before dissection for sample collection. All efforts were made to minimize suffering.

Sample collection and experimental design
In our previous study, MDV challenge trial was conducted with 150 chickens from a White Leghorn specific pathogen-free line (BWEL) (Lian et al. 2010). A total of 100 1-d-old chickens was infected intraperitoneally with 2000 PFU of the MDV GA strain in CEF, and 50 chickens were injected with the same dosage of CEF cells (0.2 ml) as controls. The two groups were housed in separate isolators. The chickens were observed until 56 d postinfection (dpi). During this phase, the clinical signs of infected chickens were observed two to three times daily. Some chickens showed severe appetite loss, serious distress and depression, and impending death. These severely morbid chickens, and one to three age-matched noninfected chickens were killed using T61 intravenously (0.3 ml/kg) (Intervet, Ukkel, Belgium). The dead chickens were dissected, and their whole spleens and MD lymphoma in livers were collected. Eight MDV-infected tumorous spleens, eight MD lymphomas from liver, eight noninfected spleens, and eight noninfected livers collected during 30-56 dpi (Lian et al. 2012a) were used in this study. A schematic of the experimental design is shown in Supplemental Material, Figure S1. The binding sites of gga-miR-103-3p in the 3ʹ-UTR of target genes were confirmed by luciferase reporter assay, and the expression of target genes in mRNA, and protein levels, were validated using real-time PCR and western blot analysis, respectively; moreover, the biological function of gga-miR-103-3p in cell proliferation and migration was investigated.
RNA isolation and real-time PCR for quantifying gga-miR-103-3p expression in tissue Total RNA was isolated from tissue samples using a mirVana miRNA Isolation Kit (Ambion, Life Technologies, Carlsbad, CA), as described by the manufacturer, and reverse transcribed using a miRACLE cDNA Synthesis Kit (Genetimes Technology, Shanghai, China). Real-time PCR was performed on an ABI 7500 system (Applied Biosystems, Foster City, CA). A specific forward primer, and universal reverse primer, for gga-miR-103-3p, and a reference small noncoding RNA (chicken 5S rRNA), were designed and synthesized by Genetimes Technology, Inc (Table 1). Real-time PCR was performed using miRACLE qPCR miRNA Master Mix kit (Genetimes Technology). The optimum n thermal cycling parameters were 95°for 10 min, 40 cycles of 95°for 10 s, 60°for 20 s, 72°for 1 min, 95°for 15 s, 60°for 30 s, and 95°f or 15 s. All experiments were run in triplicate. The expression of gga-miR-103-3p was measured using the 2 -DDCt method.

Gain/loss function study
The MD lymphoma cell line MDCC-MSB1 [provided by Dr. C. Itakura (Akiyama and Kato 1974;Goryo et al. 1987)] was cultured in RPMI-1640 medium (Gibco) with 10% FBS. All cells were maintained at 37°in a humidified incubator containing 5% CO 2 . The gga-miR-103-3p mimics, mimics NC, gga-miR-103-3p inhibitor, and inhibitor NC were transfected into MDCC-MSB1 cells with FuGENE HD transfection reagent, respectively. Total RNA was isolated at 24 hr, 48 hr, and 72 hr after transfection, and real-time PCR was performed as described above for validating the expression of gga-miR-103-3p. To detect gene expression, total RNA was extracted from MDCC-MSB1 cells using E.Z.N.A. Total RNA Kit II (Omega, GA), and reverse transcribed into cDNA by EasyScript First-Strand cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). Each reaction was in a final volume of 15 ml, containing 1 ml of cDNA, 0.2 ml of each gene primer (100 nM, Table 4), and 1· PCR mix (Power SYBR Green PCR Master Mix, Applied Biosystems), and the optimum thermal cycling parameters included 95°for 10 min, 40 cycles of 95°for 10 s, 60°for 1 min, 95°for 15 s, 60°for 30 s, and 95°for 15 s. All experiments were run in triplicate. The expression of gga-miR-103-3p, and the two targets, were measured using the 2 -DDCt method.
n  Cell proliferation and migration assay A total of 1 · 10 4 MDCC-MSB1 cells per well were seeded into 96-well plates, and transfected with gga-miR-103-3p mimics, mimics NC, gga-miR-103-3p inhibitor, and inhibitor NC, respectively. The CCK-8 (Cell Counting Kit-8, Beyotime) was added to each well at 24 hr, 48 hr, and 72 hr post transfection, and incubated at 37°for 2 hr. The absorbance at 450 nm was measured using the microplate spectrophotometer. All experiments were performed in triplicate. The migration ability of MDCC-MSB1 cells was detected using Transwells (8 mm pore size, Corning Costar Corp, Corning, NY) at 48 hr after transfection. Transwells were put into the 24-well plates. A total of 1.5 · 10 4 cells/well were seeded in the upper chamber, and 500 ml of RPMI-1640 containing 20% FBS was added into the lower chambers. After incubating for 16 hr at 37°in a 5% CO 2 humidified incubator, the cells were counted under a microscope in three independent fields, with magnification of · 100. The mean of triplicate assays for each experimental condition was used. All experiments were performed in triplicate.

Data analysis
Data were expressed as the mean 6 SD. The data were analyzed using a two-tailed Student's t-test, and the differences were considered to be statistically significant at P , 0.05.

Data availability
File S1 contains the target genes of gga-miR-103-3p predicted by TargetScan and miRDB, and clustered by GO analysis.

Differential expression of gga-miR-103-3p between MDV-infected and noninfected groups
The expression of gga-miR-103-3p among tumorous spleen, MD lymphoma from liver, noninfected spleen, and noninfected liver, was detected by qPCR. The results showed that expression of gga-miR-103-3p was significantly downregulated in tumorous spleen and MD lymphoma from liver compared with the noninfected tissues (Figure 1).
Prediction and verification of gga-miR-103-3p target genes In total, 236 and 316 target genes were predicted by TargetScan and miRDB, respectively. Fifty-six genes were predicted in both databases, which were used for GO analysis on the DAVID platform. The 56 common genes were clustered to different kinds of biological processes, including energy metabolism, ion transport, and cellular processes. Cyclin E1 (CCNE1), and transcription factor DP-2 (E2F dimerization partner 2) (TFDP2), the two genes with the highest prediction score in both TargetScan and miRDB, are linked with the cell cycle (KEGG Pathway: map 04110), and TFDP2 is involved in regulation of transcription (GO:0045449) (File S1). Referring to the results of GO analysis, and our previous microarray study (Lian et al. 2012a), CCNE1 and TFDP2 were chosen for further validation. The luciferase reporter system was used to detect the interaction of miRNA and its predicted target genes. The results showed that relative luciferase activity of CCNE1 wt UTR reporter was significantly inhibited by 40% when cotransfected with gga-miR-103-3p mimics compared with mimics NC ( Figure 2B). The three mutant-type reporters, including CCNE1 mut (75-81) UTR, CCNE1 mut (248-254) UTR, and CCNE1 mut (475-481) UTR reporter, were cotransfected with gga-miR-103-3p mimics, and mimics NC, respectively. The results showed that the relative luciferase activities of CCNE1 mut (248-254) UTR reporter transfected with gga-miR-103-3p mimics was significantly decreased compared with mimic NC. CCNE1 mut (75-81) UTR and CCNE1 mut (475-481) UTR had no significant change, but the relative luciferase activities of CCNE1 (475-481) UTR displayed a downtrend when this reporter was cotransfected with gga-miR-103-3p mimics ( Figure 2B). Taken together, the results demonstrate strongly that the CCNE1 (248-254) UTR was not the binding site of gga-miR-103-3p. The CCNE1 (75-81) UTR and CCNE1 (475-481) UTR could bind gga-miR-103-3p, and the CCNE1 (75-81) UTR was the primary binding site of gga-miR-103-3p.
The relative luciferase activities of TFDP2 wt (1-1252) UTR and TFDP2 wt (5161-6415) UTR reporters cotransfected with gga-miR-103-3p mimics were both significantly decreased ( Figure 2D). However, the relative luciferase activities of the three mutant-type luciferase reporters exhibited no significant change after cotransfection ( Figure  2D). These results suggest that gga-miR-103-3p could bind to three sites in the target TFDP2 mRNA. Overall, CCNE1 and TFDP2 are the two candidate target genes for gga-miR-103-3p.
The RNA expression of gga-miR-103-3p, CCNE1, and TFDP2 in MDCC-MSB1 cells was examined at 24 hr, 48 hr, and 72 hr after transfecting gga-miR-103-3p mimics, mimics NC, gga-miR-103-3p inhibitor, and inhibitor NC, respectively. As shown in Figure 3, the expression level of gga-miR-103-3p was significantly higher in the gga-miR-103-3p mimics group, and lower in the gga-miR-103-3p inhibitor transfection group than that in corresponding negative control, Figure 1 Differential expression of gga-miR-103-3p among tumorous spleen, noninfected spleen, MD lymphoma from liver, and noninfected liver. Results were measured using the 2 -DDCt method. The fold change of TS was compared with that of NS, and NL was the control for LL. Ã P , 0.05. TS, tumorous spleen; NS, noninfected spleen; LL, MD lymphoma from liver; NL, noninfected liver. respectively ( Figure 3A). On the contrary, expression of CCNE1 was significantly lower at 48 hr after transfection with gga-miR-103-3p mimics compared with the mimics NC group. At 24 hr and 72 hr, the expression of CCNE1 in the mimics groups exhibited no marked change, but there was a visible downward trend. The expression of CCNE1 at 72 hr after transfecting MDCC-MSB1 cells with gga-miR-103-3p inhibitor was notably increased compared with the inhibitor NC group, which indirectly indicated that gga-miR-103-3p could bind with CCNE1 ( Figure 3B). However, there was no obvious change in TFDP2 mRNA level after transfection ( Figure 3C). The results demonstrated that gga-miR-103-3p could interfere with expression of CCNE1 mRNA, but did not regulate TFDP2 mRNA expression.
The protein expression levels of CCNE1 and TFDP2 were detected in MDCC-MSB1 cells at 72 hr after transfection with gga-miR-103-3p mimics, mimics NC, gga-miR-103-3p inhibitor, and inhibitor NC, respectively. The results showed that the protein levels of CCNE1 and TFDP2 were significantly reduced after transfection with gga-miR-103-3p mimics compared with mimics NC, and the relative protein expression of these two genes in gga-miR-103-3p inhibitor groups were notably increased compared to inhibitor NC groups, respectively (Figure 4). Results of target mRNA expression and western blot assay indicated that gga-miR-103-3p regulated CCNE1 gene expression at both mRNA and protein levels. However, gga-miR-103-3p could modulate TFDP2 at the protein but not the mRNA level.

The effects of gga-miR-103-3p on cell proliferation and migration
To explore the role played by gga-miR-103-3p in MDCC-MSB1 cells, we performed cell proliferation and migration assays after overexpressing or interfering with gga-miR-103-3p. The increasing cell viability at various time points indicated that cell proliferation increased over time. However, the cell viability exhibited no obvious change at 24 hr, 48 hr, Figure 2 Luciferase reporter assay. (A) Nucleotide sequences of CCNE1 wild type and mutant in the 3ʹ-untranslated region (3ʹ-UTR). Three potential binding sites were marked in bold italic. (B) The relative luciferase activities of CCNE1 wild type and mutant reporters normalized against mimics NC. (C) Nucleotide sequences of TFDP2 wild type and mutant in the 3ʹ-UTR. Three potential binding sites were marked with bold italic. (D) The relative luciferase activities of TFDP2 wild and mutant reporters normalized against mimics NC. Two wild luciferase reporters [TFDP2 wt (1-1252) UTR and TFDP2 wt (5161-6415) UTR] for the TFDP2 gene were constructed because of its long 3ʹ-UTR region. Ã P , 0.05. and 72 hr in gga-miR-103-3p mimics or gga-miR-103-3p inhibitor groups, compared with the corresponding NC group ( Figure 5A). This suggested that gga-miR-103-3p could not influence MDCC-MSB1 cell proliferation. As shown in Figure 5B, the migrated cell number in the gga-miR-103-3p mimics transfection group was notably decreased compared with the mimics NC group, suggesting that overexpression of gga-miR-103-3p could inhibit MDCC-MSB1 cell migration at 48 h. Collectively, gga-miR-103-3p could suppress MDCC-MSB1 cell migration, but has no obvious effects on its proliferation.

DISCUSSION
MicroRNAs are associated with a variety of diseases, including cancers and oncosis (Chen et al. 2008). It was reported that miR-103 was differentially expressed in lung cancer (Garofalo et al. 2012), colorectal cancer (Chen et al. 2012), human gastric carcinoma (Zhang et al. 2015), and breast cancer (Kleivi Sahlberg et al. 2015). Although, miR-103 has been investigated extensively in mammalian diseases, studies on gga-miR-103 in chickens are still limited. In recent years, only a few studies on chicken gga-miR-103 have been reported, including deep sequencing in chicken preadipocytes , hypothalamus tissue , chromium metabolism (Pan et al. 2013), abdominal fatness , and abdominal adipose tissues (Huang et al. 2015). As for MD, it was reported that gga-miR-103-3p was expressed differentially between MDV-infected and uninfected CEF cells (Burnside et al. 2008). In MDV-infected chickens of line 7 2 (MD susceptible line), the expression of gga-miR-103 was significantly downregulated compared with noninfected chickens at 21 dpi . Our results showed that gga-miR-103-3p was greatly decreased in MDV-infected tissues, which is consistent with the previous studies (Burnside et al. 2008;Tian et al. 2012). at 24 hr, 48 hr, and 72 hr. Results showed the fold change of expression using the 2 -DDCt method. Chicken 5S rRNA and GAPDH were the endogenous reference for gga-miR-103-3p and the two target genes, respectively. The gga-miR-103-3p mimics group was compared with mimics NC group, and inhibitor NC group was the control for gga-miR-103-3p inhibitor group. Ã P , 0.05, ÃÃ P , 0.01. Several studies have shown that miR-103 is involved in cancers, and in regulated cell proliferation and migration. A vital molecular relationship between miR-103 and PER3 (period circadian clock 3) in human colorectal cancer (CRC) cell lines was confirmed. The proliferation and migration of CRC cells could be regulated by miR-103 (Hong et al. 2014). MiR-103/107 overexpression or CDK5R1 [cyclin-dependent kinase 5, regulatory subunit 1 (p35)] silencing caused a reduction in SK-N-BE migration ability (Moncini et al. 2011). Our results suggested that gga-miR-103-3p suppressed cell migration in the MDCC-MSB1 cell line. miRNAs play critical roles in gene regulatory networks. An individual miRNA can target hundreds or thousands of different mRNAs; meanwhile, an individual gene can be coordinately suppressed by multiple various miRNAs (Krek et al. 2005;Friedman et al. 2009). The miR-103 family has been found in 23 species (Griffiths-Jones 2004;Griffiths-Jones et al. 2006). A total of 257 conserved sites, and 79 poorly conserved sites, were predicted using TargetScan, and the seed regions, which aligned to mRNA 3ʹ-UTR of targets, were the same in different species, suggesting that the targets were highly conserved among human, chicken, and some other species (http://www.targetscan. org/cgi-bin/targetscan/vert_61/targetscan.cgi?mirg=gga-miR-103). Considering the results of TargetScan, miRDB, and GO analysis, we selected two target genes, CCNE1 and TFDP2, for further validation. The product of the CCNE1 gene forms a complex with CDK2 (cyclin-dependent kinase 2), and functions as a regulatory subunit, which is required for cell cycle G1/S transition (Ekholm and Reed 2000;Mazumder et al. 2004;Sherr 2000;Sherr and Roberts 1995). Overexpression of this gene can cause chromosome instability in many tumors; thus, it may contribute to tumorigenesis. The expression level of CCNE1 begins to rise in mid-G1 phase, peaks during late G1, and declines around the G1/S transition. So the timing of its expression level plays a critical role in initiation of DNA replication as well as chromatin remodeling during tumorigenesis . CCNE1 is the target gene of miR-15/16 in human lung cancer (Bandi and Vassella 2011) and glioblastoma (Xia et al. 2009) during tumorigenesis (http:// www.kegg.jp/kegg-bin/show_pathway?hsa05206+898). From the point of view of cancer or tumor research, it was reported that miR-103 is part of the G1/S transition regulatory network, by targeting CCNE1, CDK2, and CREB1 (cAMP responsive element binding protein 1) during IGF-1 simulated proliferation in mouse crypt cells (Liao and Lonnerdal 2010). miR-107, as a paralog of miR-103, was identified in human nonsmall cell lung cancer, and the potential target CCNE1 was downregulated by transfection with miR-107 (Takahashi et al. 2009). In this study, we identified a binding site of gga-miR-103-3p in the 39UTR of CCNE1, and verified that CCNE1 was a target gene of gga-miR-103-3p.
The TFDP2 gene (E2F/DP2) belongs to the transcription factor dimerization partner (DP) family, and its encoded protein forms heterodimers with the E2F transcription factor, resulting in transcriptional activation of cell cycle regulated genes (Wu et al. 1995;Zhang et al. 1997). E2F complexes can be broadly classified into two subgroups: "activating" E2Fs (E2F1, E2F2, and E2F3), and "repressive" E2Fs (E2F4, E2F5, and E2F6) (Trimarchi and Lees 2002). Studies have suggested that E2F1 is the target gene of miR-20a (Sylvestre et al. 2007), miR-106b (Ambs et al. 2008), and miR-330 (Lee et al. 2009) in prostate cancer. These two target genes are involved in the cell cycle pathway (KEGG Pathway: map 04110), and affect cell cycle progression from G1 phase to S phase. Therefore, aberrant expressions of these two genes may disturb the normal cell cycle, and further cause tumorigenesis (http://www. genome.jp/kegg-bin/show_pathway?map04110). Meanwhile, the Cyclin E family and E2F factors participate in the pathways in cancer (KEGG Pathway: ko 05200), and can affect cell proliferation indirectly (http://www.genome.jp/kegg-bin/show_pathway?map05200). Additionally, TFDP2 could be targeted by miR-122, which inhibits c-Myc transcription in hepatocellular cancer indirectly (B. . Here, we verified that TFDP2 was another target gene of gga-miR-103-3p by luciferase reporter assay and western blot assay. In summary, we found that gga-miR-103-3p was downregulated in MDV-infected tissues, and it inhibits cell migration in MDCC-MSB1 cell lines. The two target genes CCNE1 and TPDP2 were identified. The effective binding sites of gga-miR-103-3p in the 3ʹ-UTR of CCNE1 and TFDP2 mRNA were confirmed. Gga-miR-103-3p with gga-miR-103-3p mimics, mimics NC, gga-miR-103-3p inhibitor, and inhibitor NC, respectively. OD, optical density. (B) The number of migrated MDCC-MSB1 cells at 48 hr after transfection with gga-miR-103-3p mimics, mimics NC, gga-miR-103-3p inhibitor, and inhibitor, respectively. Cell number is the mean of triplicate assays for each experimental condition. Ã P , 0.05. could suppress CCNE1 gene expression at both mRNA and protein levels, while the TPDP2 gene was regulated by gga-miR-103-3p at the protein but not mRNA level. Gga-miR-103-3p, working together with its targets, may play a potential role in MDV tumorigenesis.