Discovery of chicken microRNAs associated with lipogenesis and cell proliferation

Julie A. Hicks, Nares Trakooljul, Hsiao-Ching Liu

Abstract

The primary function of microRNA (miRNA, a class of small regulatory RNA) is to regulate gene expression. Studies of miRNA in mammals suggest that many liver-associated miRNAs are expressed, with a wide range of functions. To characterize miRNA expressed in the avian liver, we created two small RNA libraries from embryonic chick livers at embryonic day (E)15 and E20, a time at which the embryo begins to grow rapidly and so its energy demands increase. It is of interest to examine miRNAs expressed at these developmental stages because miRNAs involved in regulating metabolic pathways and cell proliferation are likely to be identified. The small RNA libraries were sequenced with 454 Life Sciences deep sequencing. Of the 49,937 sequences obtained, 29,390 represented known chicken miRNAs and 1,233 reads represented homologous miRNAs that have not been previously identified in chickens. Additionally, 1,032 reads represented 17 potential novel miRNAs not previously identified in any species. To further investigate the possible functions of avian liver miRNAs we identified the potential targets of two differentially expressed novel miRNAs, nc-miR-5 and nc-miR-33. These two miRNAs were predicted to target metabolic genes, including the lipid metabolism-associated gene fatty acid synthase (FAS), and genes involved in the control of cell proliferation, such as peroxisome proliferator-activated binding protein (Pparbp) and bone morphogenetic protein 4 (BMP4). Our findings demonstrate that a diverse group of miRNAs are expressed in developing avian livers. In addition, some of the identified miRNAs have been suggested to play a key role(s) in regulating metabolic pathways.

  • liver
  • targets

it is now well established that small regulatory RNAs are a major component of gene regulation in the vertebrate genome. An important member of the small regulatory RNA family is microRNA (miRNA). miRNAs are small noncoding molecules ranging in size from 19 to 24 nucleotides (nt). Mature miRNAs are processed from ∼70-nt-long precursor miRNA (pre-miRNA) by the endonucleases Drosha and Dicer before they associate with RNA-induced silencing complex (RISC) to silence/suppress the expression of their target mRNAs (4). Current estimates indicate that each miRNA could potentially target a couple of hundred mRNAs (21). In most cases the miRNA binding site is located within the 3′ untranslated region (UTR) of the targeted mRNA. Of particular importance in miRNA target site recognition is a perfect complementary match between the seed region of the miRNA and its target site on the mRNA (3), although this seed match complementarity may not be necessary for all miRNA recognitions (12). The seed region consists of 6–8 nt located at the 5′ terminus of the miRNA and the 3′ terminus of the mRNA target site. Thermodynamic stability of the miRNA:mRNA interaction is also a critical determinant of miRNA targeting (39). Several computational methods have been developed to predict miRNA target sites based on these characteristics in addition to other factors, such as target site conservation across species (18, 21, 24). These algorithms have predicted thousands of miRNA targets, but few have been experimentally validated. Both bioinformatics and experimental evidence suggest that miRNAs regulate many biological processes, including stem cell proliferation, differentiation, and developmental stage transitioning (for review see Ref. 20).

A large number of miRNAs are expressed in the vertebrate liver. Recent efforts, particularly in humans and mice, have revealed that these hepatic-associated miRNAs are involved in regulating a variety of genes and pathways. Use of a mouse model in which Dicer function was abated in the liver showed that hepatocytes appeared to function normally in the livers of young mice (15). However, long-term loss of miRNA function was detrimental, as the livers of older Dicer-knockout mice were enlarged and displayed decreased apoptosis and regeneration capabilities and extensive hepatocyte damage. These long-ranging effects of mature miRNA loss in the developing mouse liver confirm that miRNAs regulate many different hepatically expressed genes. Analysis of global changes in miRNA expression in the aging mouse liver identified a group of miRNAs whose expression increases as the liver ages (25). Target gene analysis of these miRNAs, which include miR-93, miR-214, miR-669c, and miR-704, revealed that they target mitochondrion-associated genes as well as genes involved with alcohol, carbohydrate, and glutathione metabolism. It was therefore suggested that increased expression of miRNAs in older livers may be associated with aging by targeting key members of the regenerative and oxidative defense pathways. Analysis of miRNA expression between the livers of obese and normal mice revealed the differential expression of 12 miRNAs (8). Eight miRNAs, miR-34a, miR-31, miR-103, miR-107, miR-194, miR-335–5p, miR-221, and miR-200a, are upregulated in the liver of obese mice, while four miRNAs, miR-29c, miR-451, miR-21, and miR-122, are downregulated. These miRNAs were predicted to target genes involved in both glucose and lipid metabolism, suggesting that aberrant expression of miRNAs in the obese liver leads to altered metabolic states.

A recent study found that miR-122 is downregulated in intrahepatic metastatic cancers (36). A disintegrin and metalloprotease 17 (ADAM17) was identified as a potential miR-122 target. Silencing of ADAM17 decreased in vivo hepatic tumorigenesis and invasion in nude mice, indicating that one possible function of miR-122 in the vertebrate liver is to control cell proliferation by targeting ADAM17 (36). Another miRNA linked to hepatic cancer is miR-21, which is overexpressed in cholangiocarcinomas (CCAs) (32). Inhibition of miR-21 expression in a CCA cell line leads to increased expression of programmed cell death 4 (PDCD4) and tissue inhibitor of metalloproteinase 3 (TIMP3), suggesting that miR-21 may function to promote cell proliferation by targeting proapoptotic genes such as PDCD4 and TIMP3. The miRNA miR-21 is also associated with hepatocellular carcinoma (HCC) (10). Analysis of miRNA expression in various hepatitis B virus-positive human and woodchuck HCC cell lines found that miR-21 and the polycistronic miR-17–92 miRNA family were overexpressed in all of the cell lines examined. Moreover, knockdown of these miRNAs with antisense oligonucleotides dramatically reduced proliferation and anchorage-independent growth of these HCC cell lines. Current knowledge suggests that a diverse group of miRNAs is expressed in the vertebrate liver and that these miRNAs are likely involved in regulating many different genes associated with metabolic pathways and the control of cell growth.

To better understand the role(s) of miRNAs in the avian liver, it is essential to first identify which miRNAs are expressed. Once this has been established, the focus can then be placed on determining the function of specific miRNAs in the liver. Profiling miRNA expression in the later developmental stages of the embryonic chick liver, a time at which the embryo is growing rapidly and has high energy demands, should reveal a miRNA population involved in regulating metabolic pathways. MiRNAs associated with gene regulation in the developing avian liver, such as those involved in regulating cell growth, will also be discovered. In the present study, small RNA libraries were constructed from the embryonic chick liver at embryonic day (E)15 and E20 and sequenced with the 454 Life Sciences approach. Over 100 miRNAs, including novel miRNAs, were identified. Many of these miRNAs were differentially expressed between the two time points, suggesting that different sets of miRNA may be key regulators during these stages of liver development. A group of miRNAs that displayed high and/or differential expression in our small RNA profiles were further characterized by determining their expression at additional embryonic and posthatch time points. Additionally, the miRNA target prediction algorithm miRanda was employed to predict the potential target genes of two of these miRNAs, which were then verified by means of a luciferase reporter assay. Our results suggest that liver-associated miRNAs are likely involved in regulating metabolic genes, such as those involved in lipid metabolism, as well as genes involved in the control of cell proliferation in the avian liver.

MATERIALS AND METHODS

Tissue collection.

Tissues were collected from specific pathogen-free white leghorn embryos (Charles River Laboratories) on E15, E18, and E20 and posthatch day 0, day 3, day 7, and day 14, snap frozen in liquid nitrogen, and stored at −80°C until analysis.

Small RNA isolation and sequencing.

Five hundred micrograms of total RNA (from 8 pooled livers at each time point) was purified with TRI Reagent (Sigma-Aldrich) according to the manufacturer's instructions. The RNA was precipitated overnight at −20°C and fractionated on a 15% denaturing acrylamide gel. The small RNA fraction (19–24 bases) was gel purified. A 3′ adapter (5′-AMP-CTGTAGGCACCATCAAT-ddC-3′, miRNA CloningLinker1; Integrated DNA Technologies) followed by a 5′ adapter [5′-ATCGTr (AGGCACCUGAAA)-3′; Integrated DNA Technologies] were ligated to the small RNA fraction. The small RNA fraction was reverse transcribed with primer (5′-ATTGATGGTGCCTACAG-3′) and a SuperScript RT kit (Invitrogen). The cDNA was PCR amplified with the forward primer (5′-ATCGTAGGCACCTGAAA-3′) and the reverse primer (5′-ATTGATGGTGCCTACAG-3′). Five nanograms of the amplicon was then subjected to additional PCR amplification with the 454 Life Sciences pyrosequencing protocol. PCR products were sequenced with 454 Life Sciences methodology.

Sequence analysis.

Identical sequencing reads were grouped with an in-house program. Next, the sequences were searched against the miRBase database of known miRNA sequences (http://miRNA.sanger.ac.uk/sequences/) and the Rfam database, a database of RNA families (http://www.sanger.ac.uk/Software/Rfam/). If no matches were found in either database then the sequences were BLAST searched against the Ensembl Gallus gallus database, using default parameters, to determine their genomic location. The secondary structure of the 300 bases flanking the small RNA was determined with the Mfold algorithm (45). If the small RNA and its surrounding sequence was mapped to a noncoding region of the genome and could form a hairpin secondary structure, it was then considered a potential novel miRNA for further analysis.

Northern blot.

Twenty micrograms of total RNA was isolated in the same manner as described above, separated on a 15% denaturing acrylamide gel, and transferred onto a BrightStar-Plus nylon membrane (Ambion). DNA probes (antisense to the mature miRNA sequence) were end labeled with [γ-32P]ATP (GE Healthcare Life Sciences) and a mirVana Probe & Marker Kit (Ambion). Prehybridization, hybridization, and washes were carried out at 40°C with ULTRAhyb-Oligo hybridization buffer according to the manufacturer's instructions (Ambion). Membranes were exposed on BioMax MS film (Kodak) at −80°C for 48 h. Ethidium-stained 5S rRNA was used as a loading control.

miRNA target prediction.

The chicken (G. gallus) Unigene database (NCBI) and the miRNA target prediction algorithm miRanda 3.1 (http://www.microrna.org/microrna/getDownloads.do) were employed to predict potential targets of two novel miRNAs, nc-miR-5 and nc-miR-33. For miRanda, default parameters were used with the following exceptions: the score was set to greater than or equal to 130, and the free energy was set to less than or equal to −16 kcal/mol. Additional parameters were set to include either a perfect match between bases 1 and 8 or between bases 2 and 7 as well as bases 13–16 of the miRNA and the target site of the targeted mRNA. G:U wobble base pairing was permitted. Potential target genes were selected because they had known metabolic functions and/or contained multiple miRNA binding sites.

miRNA target validation.

To validate the predicted targets of the novel miRNAs nc-miR-5 and nc-miR-33, a previously developed miRNA-expressing RCAS system was utilized to express each miRNA in chicken embryo fibroblast cell line DF1 cells (7). In this system, RCAS-miR virus-infected cells were transfected with a dual-luciferase vector in which each miRNA target site was inserted into the 3′ UTR of the Renilla luciferase reporter gene. The Gateway cloning system (Invitrogen) was utilized to insert the miRNA of interest into a modified RCAS virus. Briefly, forward and reverse primers were designed with a previously described program (6, 7). Upon annealing of the primers, a pre-miRNA sequence for each miRNA would be generated with the addition of restriction enzyme sites SphI and NgoMIV. In addition, a scrambled control RCAS-miR virus was constructed. The full primer list is given in Supplemental Table S1.1 For forward and reverse primer annealing, a 1 μM mixture of each primer was incubated at 95°C for 20 s and then allowed to cool to room temperature. The annealed product was then ligated into the pENTR3C-miR-SphNgo vector with the SphI and NgoMIV restriction sites. One hundred and fifty nanograms of each pRENTR3C-miR entry vector was recombined with 150 ng of the destination vector pRCASBP (A)-YDV (provided by Dr. Jerry Dodgson, Michigan State University) with the LR clonase mixture (Invitrogen) in Tris-EDTA (TE) buffer (pH 8.0) overnight at 25°C. For RCAS-miR virus production, RCAS vectors were transfected into DF1 cells with FuGENE 6 (Roche) according to the manufacturer's instructions. Infection was confirmed at 3 days postinfection (dpi) by immunofluorescence staining with the mouse monoclonal 3C2 antibody against viral gag protein (Developmental Studies Hybridoma Bank at University of Iowa) and FITC-conjugated goat anti mouse IgG (Invitrogen). The expression of nc-miR-5 and nc-miR-33 from RCAS-miR-infected cells was confirmed with the miScript (Qiagen) miRNA RT-PCR system (data shown in Supplemental Table S2).

The potential mRNA target sites listed in Supplemental Table S3 were cloned into the dual-luciferase vector psiCHECK2 (Promega). One hundred nanograms of each psiCHECK2 construct was transfected into RCAS-miR virus-infected DF1 cells 3 dpi with FuGENE 6 (Roche). Forty-eight hours after transfection cells were lysed in 30 μl of passive lysis buffer (Promega). Firefly luciferase and Renilla luciferase activities were determined with a VictorLight 1420 luminescence counter (PerkinElmer). The ratio of luminescence from Renilla luciferase to luminescence from firefly luciferase was determined and compared with the ratio determined for the scrambled control miRNA. Each target assay was performed in triplicate, and statistical significance was determined with a two-tailed t-test. Luciferase assays were repeated to confirm results.

RESULTS

Pyrosequencing read characteristics of chick embryonic liver small RNA libraries.

A total of 49,937 sequence reads were obtained from the chick E15 (24,260 reads) and E20 (25,677 reads) liver small RNA libraries. Of these, 29,390 represented miRNAs matching the miRBase G. gallus database and 1,233 were not found in the G. gallus database but were homologs of miRBase miRNAs from other species. A total of 1,032 represented previously unidentified novel miRNAs. Other small RNAs, such as tRNA and rRNA, represented 10,834 of the reads, and the remaining 7,448 reads represented other types of sequences, such as degraded mRNA and adapter artifacts (Fig. 1).

Fig. 1.

Composition of small RNA libraries obtained from the embryonic chick liver. Small RNA libraries were created from the embryonic day (E)15 and E20 chick liver and deep sequenced. Sequencing reads were classified into the following categories: known Gallus gallus microRNAs (miRNAs), homologous G. gallus miRNAs, novel G. gallus miRNAs, other types of small RNAs (tRNAs, rRNAs, mtRNAs, snRNAs), or other (mRNA degradation products, sequencing artifacts, etc.). miRNA represented over half (63%) of chick liver small RNA libraries.

A total of 114 different miRNAs were identified in the E15 liver library. Of these miRNAs, 86 represented known chicken miRNAs; the most frequently sequenced known miRNA was gga-miR-122, which represented 34.48% of reads. Other frequently sequenced miRNAs were gga-miR-126 (1.89%), gga-miR-17–5p (1.29%), and gga-miR-19b (1.24%) (Fig. 2 and Supplemental Table S4). Of the homologous miRNAs matching miRBase sequences from other species, miR-143 (0.94%), miR-1274b (0.42%), and miR-454–3p (0.38%) were the most commonly identified (Table 1). Fourteen miRNAs were identified in the E15 chick liver library (Table 2) and were classified as potential novel miRNAs because they fit the predetermined criteria discussed above, such as the ability to form a hairpin secondary precursor structure (Fig. 3). These novel miRNAs were given provisional names with the prefix NC (North Carolina) followed by a unique identification number. A complete list of identified known miRNAs is shown in Supplemental Table S4.

Fig. 2.

Frequently identified miRNAs in the small RNA profiles of the embryonic chick liver: a select group of miRNAs that are either the most highly and/or differentially sequenced in the 2 small RNA profiles of the embryonic chick liver. A complete list of miRNAs identified in this study can be found in Tables 1 and 2 and Supplemental Table S1.

View this table:
Table 1.

Homologous microRNAs identified in small RNA libraries of chick embryonic liver at E15 and E20

View this table:
Table 2.

Novel microRNA identified in small RNA libraries of chick embryonic liver at E15 and E20

Fig. 3.

Predicted precursor secondary structures of novel miRNAs: precursor microRNA (pre-miRNA) secondary structures (predicted with the Mfold algorithm) of selected novel chicken miRNAs identified in small RNA libraries of the embryonic chick liver. Gray text represents the mature miRNA sequence.

From the embryonic chick liver at E20, we identified 80 miRNAs in the G. gallus miRBase database (Supplemental Table S4). The most commonly sequenced (32.47%) miRNA was gga-miR-122, followed by gga-miR-126 (3.00%), gga-miR-451 (1.93%), and gga-miR-125b (1.87%) (Fig. 2). Of the miRBase sequences from species other than chicken, miR-143 was the most frequently identified miRNA (0.92%), followed by miR-1274b (0.67%) (Table 1). Sixteen potential novel miRNAs were identified in the chick E20 liver library (Table 2). These novel miRNA constituted a total of 574 reads.

Expression of selected miRNA.

Expression of selected known, homologous, and novel miRNAs in the liver, as well as in the lung and heart, at E15 and E20 was demonstrated by Northern blotting (Fig. 4). Overall, the Northern blot expression patterns of these miRNAs appear to correlate with the profiling results. The novel miRNA nc-miR-31 was equally represented in both profiles (0.08% of reads), whereas nc-miR-33 represented 0.73% of the E15 library reads and 0.56% of the E20 library. Northern blotting of these miRNAs agrees with the 454 Life Sciences sequencing results (Figs. 4 and 5). However, there is a small discrepancy between the two methods due to multiple factors, such as the exposure times required for Northern blots to visualize small RNA, which may affect band intensity. Our Northern blot indicated a higher expression of miR-363 than would be expected based on the profiling results (Fig. 4). To clarify this discrepancy quantitative RT-PCR analysis was used to verify the expression of miR-363 along with five other randomly selected miRNAs. In the miRNA expression profiles, miR-363 was represented by 1.8-fold more sequencing reads in the E15 liver, while RT-PCR analysis indicates that miR-363 was expressed ∼1.5-fold higher in the E15 liver (Supplemental Fig. S1). Overall, the results of RT-PCR analysis for the relative expression of miRNAs agree with the data gathered from the 454 Life Sciences deep sequencing and thus further indicate that our profiles accurately represent miRNA expression patterns in the chick liver.

Fig. 4.

Expression of miRNAs identified in a small RNA profile of the embryonic chick liver. Twenty micrograms of total RNA was separated on a 15% denaturing PAGE gel, transferred to a nylon membrane, and probed as described in materials and methods. Lanes from left to right: E15 and E20 liver, E15 and E20 lung, E15 and E20 heart. 5S rRNA was stained with ethidium bromide to serve as a loading control.

Fig. 5.

Expression of nc-miR-5, miR-1274b, nc-miR-33, miR-21, and miR-103 in the liver, fat, and muscle at both embryonic and posthatch stages of development. Twenty micrograms of total RNA was separated on a 15% denaturing PAGE gel, transferred to a nylon membrane, and probed as described in materials and methods. A: expression of nc-miR-5, miR-1274b, nc-miR-33, miR-21, and miR-103 at embryonic time points E18 and E20 and day (D)0. B: expression of nc-miR-5, miR-1274b, nc-miR-33, miR-21, and miR-103 at posthatch time points D3, D7, D14. 5S rRNA was stained with ethidium bromide to serve as a loading control.

To further characterize the expression patterns of miRNAs that were highly and/or differentially represented in the small RNA profiles of the embryonic chick liver, the expression of miRNAs in the liver, fat, and muscle at both embryonic and posthatch time points was determined (Fig. 5). These selected miRNAs also showed a wide range of expression patterns. The novel miRNA nc-miR-33 was most highly expressed in fat, particularly at the embryonic stages, whereas the known miRNA, miR-21, was expressed similarly in all tissues and time points.

miRNA target identification and validation.

To identify potential targets of two novel miRNAs, nc-miR-5 and nc-miR-33, the miRNA target prediction algorithm miRanda was used. These miRNAs were chosen for further analysis because they have not previously been identified and they are differentially expressed among tissues and stages of development (Fig. 5). A set of potential target genes were selected for validation of miRNA targeting because they have known roles in metabolism or cell proliferation and/or known multiple miRNA target sites. A complete list of the gene target sites is shown in Supplemental Table S3. Genes involved in the lipogenesis pathway, such as fatty acid synthase (FAS), and genes involved in cell proliferation, such as bone morphogenetic protein 4 (BMP4) and cAMP-responsive element binding protein 1 (CREB1), were identified as potential targets of these two miRNAs. Additionally, nc-miR-5 was predicted to target peroxisome proliferator-activated receptor-binding protein (Pparbp, also known as MED1), which is a critical coactivator of peroxisome proliferator-activated receptor α (PPARα) signaling.

To validate the miRNA-mRNA target pairs, a dual-luciferase reporter assay was used, the results of which are shown in Fig. 6. The ability of each miRNA to knock down luciferase activity varied among target sites. The binding of nc-miR-33 to its respective target site in the 3′ UTR of FAS reduced luciferase activity by 90% (P < 0.01). The 3′ UTR of CREB1 was predicted to contain five potential nc-miR-5 binding sites, two sites near the 5′ end of its 3′ UTR and three sites in the distal end of its 3′ UTR. The activity of a Renilla luciferase construct containing the first two predicted target sites was reduced by ∼70% in DF1 cells overexpressing nc-miR-5, while the luciferase activity of a construct containing the distal three nc-miR-5 target sites was reduced by ∼90% on the introduction of nc-miR-5 (Fig. 6). Our results also suggest that genes KFL1, NOS2, FHL2, and ERBB4 are genuine targets for nc-miR-5, whereas genes TGFB1, HOXA11, and GTAM are targets for nc-miR-33. However, the expression of nc-miR-5 and nc-miR-33 with the RCAS-miR system did not significantly reduce the Renilla luciferase constructs for several of the predicted targets sites, such as the nc-miR-5 target sites of FAS, E2F8, and ELOV5 and the nc-miR-33 predicted target site in CIP1. This would suggest either that these computationally predicted binding sites are not true binding sites of these miRNAs or that possibly these sites are only very weakly targeted by these miRNAs.

Fig. 6.

miRNAs expressed in the embryonic chick liver likely target genes involved in metabolic pathways and hepatocyte proliferation. Luciferase assays for the miRNAs nc-miR-5 and nc-miR-33 and the scrambled control (SC) for miRNA target gene validation are shown. DF1 cells infected with each RCAS-miR virus were transfected with appropriate predicted target 3′ untranslated region luciferase constructs. KLF11 served as a nontarget gene control. Relative Renilla luciferase activity is shown. *P < 0.05, **P < 0.01.

DISCUSSION

In the present study we identified 123 miRNAs. Of these miRNAs, 89 represented known chicken miRNAs, 17 represented homologous miRNAs, and 17 represented potential novel miRNAs. Among the obtained small RNA reads, miR-122 represented ∼30% of the total reads (Fig. 2), consistent with other studies that found miR-122 to be the most highly expressed miRNA in the liver (5, 19). miR-122 is highly conserved across vertebrates and is involved in regulating cholesterol synthesis and lipid metabolism (22). Analysis of miRNAs expression between the two libraries using Fisher's exact test with Bonferroni correction identified 31 (∼24%) differentially expressed miRNAs, suggesting that these miRNAs act as key regulators during liver development or hepatogenesis. Among these, the expression of miR-21 was significantly higher in the liver at E20 compared with E15, while the novel miRNA nc-miR-5 was significantly higher at E15. To further characterize two differentially expressed novel miRNAs (nc-miR-5 and nc-miR-33), their possible target genes were identified. These putative target genes included the metabolic genes l-arginine:glycine amidinotransferase (GTAM) and FAS, suggesting that these miRNAs are possibly involved in regulating avian metabolic pathways. These miRNAs also were predicted to target genes involved in the control of hepatogenesis and cell proliferation, such as CREB1 and BMP4.

The fetal liver possesses multiple functional roles in the developing embryo and so consists of a mixed cell population. Each cell type within this population is likely to express a specific set of miRNAs during development. Profiling miRNA expression in whole livers will likely identify specific sets of miRNAs derived from each of the various cell types that comprise the liver. In mammals, the embryonic liver is involved in various metabolic processes and serves as a major site of hematopoiesis. Studies in mice have revealed that the liver is colonized by hematopoietic stem cells (HSCs) early in embryonic development and this population of HSCs is maintained through birth (reviewed in Ref. 14). These HSCs migrate from the fetal liver to populate other hematopoietic sites such as the spleen and bone marrow. It is anticipated that mammalian embryonic liver miRNA profiles should include miRNAs expressed by hematopoietic cells, such as members of the miR-17–92 family. In contrast, it is thought that the avian liver has only a very minor role in hematopoiesis (reviewed in Ref. 34). In birds it appears that a site near the aorta, termed the para-aortic foci, serves a functional role similar to that of the mammalian fetal liver in hematopoiesis (reviewed in Ref. 11). Therefore it was not expected that miRNAs originating from hematopoietic cells would be identified in profiles of the embryonic avian liver. However, our profiles contain several miRNAs known to be associated with hematopoietic cells. The miRNAs miR-17–5p, miR-181a, miR-18a, miR-221, and miR-222 are all involved in regulating gene expression in mammalian immune cells and were also recently identified in miRNA profiles of embryonic chick immune organs (16). The miRNAs miR-17–5p and miR-18a are members of the miR-17–92 miRNA family, which is associated with the regulation of lymphocyte proliferation and function (41). The related miRNAs miR-221 and miR-222 are known regulators of kit ligand signaling during the recruitment and maintenance of precursor hematopoietic cells (13). There are a few plausible explanations for the expression of hematopoiesis-associated miRNAs in the embryonic chick liver. Even though the avian liver is not considered a major hematopoietic site, it is known to play minor roles in erythopoiesis and granulopoiesis (34). Thus hematopoietic cells are present in the developing avian liver and could be the source of the hematopoiesis-associated miRNAs identified in the present study. It is also possible that these miRNAs have currently unknown functions in nonhematopoietic cells present in chick liver. Further work determining miRNAs expressed by specific cell types present in the embryonic chick liver should enable us to distinguish among these possible reasons for the expression of hematopoiesis-associated miRNAs in the embryonic chick liver.

A recent study utilizing miRNA microarrays to profile miRNA expression during early liver development in humans (37) has extended our knowledge of liver miRNA expression in vertebrates. Comparison of the 20 most highly expressed miRNAs between the human embryonic liver and our present study reveals 6 overlapping miRNAs, miR-122, miR-17–5p, miR-125b, miR-16, miR-26a, and miR-130a, suggesting that these miRNAs might share conserved functions. Many of these miRNAs are known to function in the regulation of cell proliferation and apoptosis. Other miRNAs, such as miR-100, miR-130b, miR-24, and miR-193, appear to exhibit different expression patterns between developing human and avian livers. miR-100 and miR-130b are highly expressed in the chick liver (Supplemental Table S2), whereas miR-24 and miR-193 are highly expressed in the developing human liver, suggesting more specific regulatory roles in cellular differentiation or organism development for these miRNAs. In comparing our results with an adult human liver miRNA profiling study (23), 8 of the top 20 miRNAs identified from both studies overlapped. Of these, miR-122, miR-125b, and miR-16 are also highly expressed in the early human embryonic liver. miR-122 is involved in the regulation of cholesterol synthesis, while miR-125b and miR-16 are two of the more well-characterized miRNAs and are associated with the regulation of cell proliferation. Study of the fetal mouse liver (E12.5) has identified the expression of 55 miRNAs (43). Forty of these mouse miRNAs overlap with the miRNAs found in our present study, and the expression patterns appeared highly comparable in both studies, including miRNAs miR-21, miR-18a, and let-7a. However, quantitative differences in the expression patterns of some miRNAs (e.g., miR-140) were also observed. The larger degree of similarity in miRNA composition between human, mouse, and chick embryonic livers suggests that many miRNAs play conserved function(s) across vertebrates during hepatic development. The differences in expression patterns could be explained by species-specific differences in developmental timing or differential cellular process between these species.

Two key enzymes in lipogenesis are acetyl-CoA carboxylase (ACC) and FAS. ACC catalyzes the conversion of acetyl-CoA to malonyl-CoA, which is involved in regulating the oxidation and synthesis of fatty acids (38). FAS then utilizes malonyl-CoA to produce long-chain fatty acids. In the young chick, FAS expression increases as feed intake increases, indicating increased fatty acid synthesis in the posthatch chick. A microarray profile of metabolic gene expression between the embryonic chick and the posthatch chick revealed a large array of differentially expressed genes (9). The expression of a set of 21 genes was higher in the posthatch chicks compared with the embryos, and among these genes is FAS. Research has shown that FAS transcripts are present at very low levels in the E16 embryonic chick liver and then steadily increase through hatching (28). Our data demonstrate that nc-miR-33 can potentially regulate FAS expression. It has been well established that FAS activity is regulated by a variety of mechanisms; for example, glucagon prevents FAS mRNA accumulation (40). The interplay between these factors and miRNAs in fatty acid synthesis is unclear. Potential targeting of FAS by miRNA will add an additional layer to the complexity of lipid metabolic regulation.

Target prediction and validation of miRNAs expressed in the embryonic chick liver suggest that miRNAs are regulating other metabolic pathways in addition to lipid metabolism. The novel miRNA nc-miR-33 possibly targets GTAM mRNA. GTAM is an enzyme important in amino acid metabolism and functions by catalyzing the transfer of an amidino group from l-arginine to glycine (29). Research has shown that the expression of GTAM is higher in posthatch chicks than it is in embryos (9). These results suggest that hepatic-expressed chicken miRNAs are likely involved in the regulation of multiple metabolic pathways, such as those associated with lipid and amino acid metabolism.

BMP4-null mice revealed a role for this particular bone morphogenic gene in early mouse hepatogenesis. Signaling involving BMP4 is needed to induce liver formation from the ventral foregut endoderm and is also essential for the growth of the hepatic endoderm into a liver bud (31). Studies have also found that both BMP and FGF signaling induce hepatic gene expression in the endoderm to induce liver formation (44). Moreover, BMP4 is involved in the induction of bile duct differentiation (42). In addition to its roles in early liver formation, BMP4 enhances the promotion of cell proliferation by hepatocyte growth factor (HGF) in the adult liver (2, 27). These observations suggest that a robust expression of BMP4 is likely critical in the early formation of the liver, but once the liver is formed the expression of BMP4 needs to be tightly regulated in order to properly control hepatocyte proliferation. Therefore the targeting of BMP4 by an miRNA expressed in the late embryonic and posthatch chick liver, nc-miR-33, could serve to aid in the regulation of HGF-induced hepatocyte proliferation.

CREB is a transcription factor that, upon activation, binds to cAMP-responsive elements (CREs) in the promoters of target genes to direct their transcription. This transcription factor is an important regulator during development and has been found to regulate both cell proliferation and cell differentiation in many cell types (reviewed in Ref. 35). The 3′ UTR of CREB1 was predicted to contain multiple binding sites for the novel miRNA nc-miR-5 identified in small RNA profiles of the embryonic chick liver. For validation of these target sites two Renilla luciferase CREB constructs were generated (Fig. 6). The construct CREBa contains two putative nc-miR-5 binding sites, and the construct CREBb contains three putative nc-miR-5 binding sites. Both constructs displayed significantly reduced Renilla luciferase activity upon the overexpression of nc-miR-5 in DF1 cells, suggesting that these sites are authentic nc-miR-5 binding sites. Through the use of CREB mutants it was shown that one function of CREB in the mouse liver is in the control of hepatocyte growth and apoptosis (1). Therefore it is possible that nc-miR-5 may target CREB to regulate CREB-mediated hepatocyte growth.

A recent study discovered a signaling cascade involving the regulation of miRNA expression in hepatocellular proliferation (33). In that work the authors discovered that peroxisome proliferator-activated receptor α (PPARα), a steroid hormone receptor, could regulate miRNA expression in hepatocytes. In particular, inhibition of PPARα activity led to a reciprocal decrease in the miRNA let-7c expression, which ultimately led to increased cell growth. One of the let-7c targeted genes is c-myc, which binds to an upstream sequence of the miR-17 family of miRNAs (a member of which is miR-17–5p) and increases miR-17 family miRNA expression, which in turn leads to an increase in cell proliferation (30). The results of the present study suggest that miRNAs possibly regulate PPARα activity. Pparbp is a coactivator in PPARα signaling (17). The disruption of Pparbp in mouse embryos leads to impaired liver regeneration and disrupts hepatocyte proliferation (26). In our study, we identified a novel miRNA, nc-miR-5, that can target Pparbp and downregulate its expression (Supplemental Table S3; Fig. 6). Together these observations, along with the fact that PPARα signaling can regulate miRNA activity, suggest that a very complex check and balance system is in place to control hepatocyte proliferation via the PPARα signaling pathway, in which PPARα can regulate miRNA function and in turn miRNAs are also able to regulate PPARα function.

In conclusion, our studies have revealed that a large and diverse group of miRNAs are expressed in developing avian livers. These miRNAs likely have diverse functions. Because whole livers were utilized in this study, the miRNA profiles likely consist of a mixture of miRNAs expressed in different cell types associated with the embryonic liver. For example, it is possible that the miRNA miR-17–5p, which is known to be a hematopoietic miRNA in mammals, may be expressed in hematopoietic cells present in the liver, although the avian embryonic liver has only a very minor role in hematopoiesis, in contrast to the mammalian liver (34). Our work has also identified novel putative miRNA target genes that are critical components of metabolic pathways, such as those involved in lipogenesis, as well as genes involved in regulating liver growth.

GRANTS

This work was supported by the North Carolina Agricultural Research Service and by Agriculture and Food Research Initiative competitive grant (award number 2010-65205-20452) from the USDA National Institute of Food and Agriculture Animal Genome Program.

DISCLOSURES

No conflicts of interest are declared by the author(s).

ACKNOWLEDGMENTS

We greatly appreciate Drs. Jerry Dodgson and Mo Chen (Michigan State University) for providing pENTR3C-miR-30a entry vector and RCASBP(A)-YDV gateway vector.

Footnotes

  • 1 The online version of this article contains supplemental material.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
View Abstract