Introduction

Hibernation is a strategy that involves tightly regulated metabolic suppression and is instigated by several mammals to survive severe environmental conditions. Mammalian hibernators undergo sustained torpor periods at low body temperature (T b ~4 °C) and can experience a metabolic rate reduction of more than 20-fold, when compared to the euthermic state [1–3]. Other physiological changes observed in animals undergoing hibernation notably include reduced blood flow, breathing rate, and heart rate [4, 5]. In hibernating ground squirrels, this hypometabolic state translates to energy savings of almost 90 % that would otherwise be consumed if the animals were to remain active during the cold season [1]. Such changes are correlated with considerable regulation of ATP-consuming processes including transcription and translation [6, 7]. Large-scale approaches performed on hibernating animals have indeed highlighted significant modulation of global transcript and protein levels during torpor [8, 9]. Selected studies on translation factors such as eIF4E and eEF2 have further emphasized the importance of translational control in hibernation [10, 11].

MicroRNAs (miRNAs) are non-coding RNAs that regulate translation of selected transcript targets in animals, plants, and protozoa [12]. miRNAs are initially transcribed as primary miRNAs (pri-miRNAs) by the RNA polymerase II and subsequently cleaved by the enzyme Drosha to yield precursor miRNAs (pre-miRNAs). Pre-miRNAs are exported to the cytoplasm and subsequently cleaved to generate mature miRNAs. miRNAs associate with several proteins to form the miRNA-induced silencing complex (miRISC). This complex interacts with target mRNAs to repress translation or to promote transcript degradation [12]. Recent predictions suggested the likelihood that miRNAs can regulate 60 % of all protein-coding genes [13]. Thus, it is not surprising that these short RNAs, usually 19–25 nucleotides in length, influence a myriad of metabolic cascades and physiological processes. This includes miRNAs that can influence insulin signaling and glucose metabolism as well as miRNAs that can regulate the expression of target transcripts involved in lipid homeostasis [14, 15].

The present study was undertaken to assess the expression of six miRNAs in a small mammalian hibernator. The characterized miRNAs have been linked, in non-hibernating models, with processes relevant to hibernation including glucose metabolism (miR-29a and miR-223), lipid metabolism, (miR-195) and muscle atrophy (miR-378). miR-152 and miR-486, miRNAs previously shown to be differentially regulated in another hibernating ground squirrel, were also selected for this study. We report the expressions of these miRNAs in three tissues of the euthermic and hibernating thirteen-lined ground squirrels, Ictidomys tridecemlineatus, and notably highlight the differential expression of miR-195, a miRNA with potential implications in fatty acid homeostasis.

Materials and methods

Animals

Animal experiments on thirteen-lined ground squirrels, I. tridecemlineatus, were performed as previously described [16]. Squirrels weighing 150–300 g were captured by a licensed trapper (TLS Research, Bloomingdale, IL) and transported to the NIH facility (Bethesda, MD). Animals were kept in shoebox cages maintained at 21 °C and fed ad libitum until they accumulated sufficient lipid stores to enter torpor. A sensor chip was injected subcutaneously in anesthetized animals. To induce hibernation, squirrels were moved to a dark cold room maintained at 4–6 °C. Body temperature (Tb) and respiration rate were closely followed to identify the stage of torpor-arousal cycle. The animals constituting the hibernating group were sacrificed following a torpor period of at least 5 days (Tb between 5 and 8 °C), while the control animals had demonstrated stable Tb readings (between 34 and 37 °C) for at least 3 days. All squirrels were sacrificed by decapitation. Tissues were subsequently shipped to Université de Moncton in New Brunswick on dry ice and stored at −80 °C until use.

Total RNA isolation

Total RNA was isolated from the liver, heart, and skeletal muscle tissues of ground squirrels using TRIzol reagent (Invitrogen) according to manufacturer’s instructions. Briefly, 100 mg of tissue was homogenized in 1 ml of TRIzol using a Polytron homogenizer followed by the addition of 0.2 ml chloroform. Samples were subsequently centrifuged at 12,000 rpm for 15 min at 4 °C. Following upper layer transfer to a different tube, 0.5 ml of isopropanol was added and the mixture was left at room temperature for 10 min to precipitate the RNA. Samples were centrifuged at 12,000 rpm for 10 min at 4 °C. The resulting RNA pellet was washed with 75 % ethanol and centrifuged at 7,500×g for 5 min at 4 °C. Pellets were dried and resuspended in 100 μl of DEPC-treated water. The final products were analyzed using a NanoVue Plus Spectrophotometer to determine RNA concentration as well as RNA purity via the absorbance ratio at 260/280 nm. All samples were stored at −80 °C until use.

cDNA synthesis

Ictidomys tridecemlineatus miR-specific stem-loop primers were designed based on a consensus alignment sequence of known miR-29a, miR-107, miR-152, miR-195, miR-223, miR-378, and miR-486 sequences obtained from Homo sapiens, Mus musculus, and Rattus norvegicus miRNA species. Sequences for stem-loop primers used are listed in Table 1. miRNAs were subsequently amplified using a stem-loop based RT-PCR protocol [17] to confirm that the designed primers could specifically amplify the miRNAs of interest. First-strand synthesis was performed as described previously [18]. Briefly, 1 μg of total RNA was combined with 5 μl of 250 nM of miRNAs-specific stem-loop primer. This mixture was first incubated for 5 min at 95 °C, subsequently placed at 60 °C for 5 min and finally cooled on ice for 1 min. The remaining reagents—4 μl of 5× first-strand buffer, 2 μl of 0.1 M dithiothreitol, 1 μl of 10 mM dNTPs, 2 μl DEPC-treated water, and 1 μl M-MLV reverse transcriptase (Life Technologies)—were added to the reaction mixture. Samples were incubated at 16 °C for 30 min, 42 °C for 30 min, and 85 °C for 5 min. Serial dilutions of the cDNA were prepared (10−1–10−4) in water and were used for PCR amplification of the different miRNAs.

Table 1 Sequences and efficiencies of the primers used to amplify miRNAs of interest in this study by qRT-PCR
Full size table

qRT-PCR amplification of miRNAs

The qRT-PCR reaction was performed by mixing 5 μl of cDNA template (10−2) to 5.5 μl of DEPC-treated water, 1 μl of 25 μM miRNA-specific forward primer, 1 μl of 25 μM universal reverse primer, and 12.5 μl of EconoTaq PLUS GREEN 2X Master Mix (Lucigen). Sequences for forward and universal primers used are listed in Table 1. The amplification protocol consisted of an initial denaturing step at 95 °C for 10 min, followed by 35 cycles at 95 °C for 15 s and 60 °C for 60 s. PCR products were separated on 2 % agarose gels and bands were visualized on a UV box. Products were sequenced by the Université Laval sequencing platform (Quebec City, QC), and sequences were confirmed as encoding miR-29a, miR-107, miR-152, miR-195, miR-223, miR-378, and miR-486 by sequence comparison using BLAST. Initial qRT-PCR runs were performed to assess the amplification efficiencies for each primer set. Briefly, each miRNA was amplified in tenfold dilutions (10−1–10−4) of cDNA samples using the same qRT-PCR conditions as described above. Reactions showed efficiencies between 90 and 97 %, and miR-107 was selected as a housekeeping miRNA. This miRNA displayed considerable stability between the euthermic and hibernating ground squirrel tissue samples as assessed by the Excel 2003 add-in NormFinder version 0.953 [19].

Western blotting

The frozen ground squirrel liver, heart, or skeletal muscle tissue samples (~100 mg) were homogenized in 1 mL of buffer containing 100 mM MOPS, 25 mM HEPES, 25 mM β-glycerophosphate, 5 mM EDTA, 1 mM EGTA, and 250 μM Na3VO4, adjusted at pH 7.4, with 1 mM phenylmethylsulfonyl fluoride added immediately before homogenization. Homogenates were centrifuged at 10,000×g for 10 min at 4 °C, and supernatants were collected. Protein concentrations in lysates were determined using the Coomassie blue dye-binding method and the BioRad reagent (BioRad, Hercules, CA). SDS–polyacrylamide gel electrophoresis and blotting to polyvinylidene difluoride (PVDF) membranes were carried out as previously reported [20] with 8 % gels, 25 µg of protein lysates loaded per well, and electrophoresis at 110 V for 90 min. Proteins were subsequently transferred onto PVDF membranes for 70 min at 300 mA using a transfer buffer consisting of 25 mM Tris (pH 8.5), 192 mM glycine, and 10 % v/v methanol at 4 °C. Following transfer, membranes were blocked for 1 h in TBST (50 mM Tris–HCl pH 6.8, 150 mM NaCl, 0.05 % v/v Tween 20) with 5 % w/v powdered skim milk. This solution was removed and membranes were incubated overnight at 4 °C with primary antibodies solution. Antibodies specific for mammalian fatty acid synthase (FAS) were purchased from Abcam (#ab96863) and used at a 1:500 v:v dilution in TBST. The following day, membranes were incubated with HRP-linked anti-rabbit IgG secondary antibody (1:2,000 v:v dilution) in TBST for 1 h. Bands were visualized using the FluorChem imaging system and the FluorChem software (Alpha Innotech).

Quantification and statistics

For qRT-PCR results, quantification of the target miRNA products was performed using the Qbase + software version 2.5.1 (Biogazelle). Target miRNA levels were normalized against miR-107 amplified from the same cDNA sample. miRNA quantification was performed using the method described by Pfaffl, which accounts for differences in PCR efficiencies between the target and reference miRNAs [21]. For immunoblotting results, band intensity of protein of interest was normalized against band intensity of the housekeeping protein GAPDH in the same lane. Mean normalized band densities ± SEM were then calculated for aroused versus torpid samples, and significant differences between the groups were tested using the Student’s t test.

Results

Amplification and sequence of miRNAs in I. tridecemlineatus

Using RT-PCR and primers derived from consensus sequences of miRNAs obtained from other mammalian species, target miRNAs were amplified in ground squirrel tissues. The products were confirmed as encoding miR-29a, miR-107, miR-152, miR-195, miR-223, miR-378, and miR-486, and the sequences were submitted to GenBank with accession numbers KJ169551, KF856940, KF856942, KJ169552, KJ169553, KF856943, and KF856944, respectively. Figure 1 shows the nucleotide sequences of target miRNAs aligned with the corresponding miRNAs from three other mammalian sources sequenced to date. miRNAs are strongly conserved in mammals [22], and this was confirmed for the miRNAs amplified in this study.

Fig. 1
figure 1

MiR-29a, miR-152, miR-195, miR-223, miR-378, and miR-486 complete sequences in I. tridecemlineatus. MiR-29a (a), miR-152 (b), miR-195 (c), miR-223 (d), miR-378 (e), and miR-486 (f) sequences amplified from I. tridecemlineatus compared to sequences of H. sapiens, M. musculus, R. norvegicus, and/or E. caballus. Genbank accession numbers are AJ421751.1, AJ459720.1, NR_031836.1, and KJ169551 for miR-29a; NR_029687.1, AJ459764.1, NR_031894.1, and KF856942 for miR-152; NR_029712.1, AJ560737.1, NR_031912.1, and KJ169552 for miR-195; NR_029637.1, NR_029801.1, NR_031936.1, and KJ169553 for miR-223; NR_029870.1, NR_029879.1, NR_032139.1, and KF856943 for miR-378; and NR_030161.1, NR_030254.1, NR_033059.1, and KF856944 for miR-486, respectively. Dashes indicate identical nucleotides between sequences; spacer dots indicate missing nucleotides in one or more sequence

Full size image

miRNA expression in hibernating tissues of I. tridecemlineatus

miRNA levels of miR-29a, miR-152, miR-195, miR-223, miR-378, and miR-486 were assessed via qRT-PCR in three tissues of the euthermic and hibernating I. tridecemlineatus (Fig. 2). miR-29a, miR-152, miR-195, miR-223, and miR-486 content was significantly higher (P < 0.05) in the liver of hibernating I. tridecemlineatus; values were 1.33 ± 0.06-, 1.79 ± 0.26-, 1.68 ± 0.19-, 4.15 ± 0.58-, and 1.62 ± 0.17-fold higher than in tissues from the euthermic animals, respectively. In skeletal muscle tissue of hibernating I. tridecemlineatus, miR-378 levels decreased significantly to 66.2 ± 5.7 % of the euthermic value (P < 0.05), while the other miRNAs quantified remained unchanged between torpid and the euthermic samples. None of the miRNAs tested in this study were differentially expressed in the hibernating heart.

Fig. 2
figure 2

Effects of hibernation on the levels of six miRNA species in the liver, skeletal muscle, and heart of I. tridecemlineatus. Histogram shows the ratios of normalized miRNA levels, versus the housekeeping miR-107, in tissues from hibernating versus the euthermic animals. Data are mean ± SEM for n = 5–6 independent trials on tissue from different animals. Expression ratios and standard errors were calculated using the relative expression software tool (Qbase + software version 2.5.1 from Biogazelle). *Significantly different from control group (P < 0.05)

Full size image

FAS protein expression in the hibernating liver

FAS protein levels were measured by immunoblotting in three ground squirrel tissues. Western blot analysis of FAS in ground squirrel liver showed that the content of the 273-kDa FAS protein decreased significantly in this tissue during hibernation to values that are 55.9 ± 20.4 % of the ones observed in the euthermic samples (Fig. 3; P < 0.05).

Fig. 3
figure 3

FAS protein expression in the liver of I. tridecemlineatus. Representative Western blots showing relative expression of the FAS protein band at 273 kDa in the liver tissue from the euthermic versus hibernating squirrels

Full size image

Discussion

miRNAs can regulate translation of several target transcripts and concurrently impact a plethora of cellular processes including metabolism, cell cycle regulation, and apoptosis [12, 23, 24]. In recent years, a growing body of evidence has highlighted the potential importance of this family of nucleotides in modulating responses to different stresses including anoxia and cold temperatures [25]. In this study, we hypothesized that the inhibitory control of transcript translation by miRNAs could be a significant mechanism to impact gene and protein expression of targets involved in key metabolic cascades in a small mammalian hibernator. The differential expression of miR-29a, miR-195, miR-223, miR-378, and miR-486 in the hibernating liver and/or skeletal muscle ground squirrel tissues reported in the current work supports this statement.

Increased miR-29a levels are observed in hibernating ground squirrel liver tissues. This miRNA has been shown to target numerous molecules involved in glucose homeostasis in the non-hibernating models. miR-29a overexpressions in primary hepatocytes and mouse livers were associated with a down-regulation in protein levels of PGC-1α and G6Pase, characterized targets of this miRNA, leading to reduced hepatic glucose production [26]. Previous studies have reported elevated PGC-1α protein levels in the hibernating liver and skeletal muscle tissues of M. lucifugus as well as in the hibernating heart and skeletal muscle tissues of I. tridecemlineatus suggesting that other miRNAs could be involved in regulating this target in hibernators [27, 28]. Measurements of miR-27 and miR-696 levels, two miRNAs that have been proposed as potential PGC-1α regulators, would yield interesting information on this topic [29, 30]. In addition, miR-29a overexpression was also correlated with insulin resistance in mouse adipocytes [31]. Interestingly, a recent publication has highlighted the fact that several pathways differentially modulated in hibernating animals are also frequently found to be deregulated during insulin resistance in human type 2 diabetes, and thus making hibernators prospective models to study this condition [32]. miR-29a might yet be another commonality between these two models.

Studies in non-hibernating animal models have highlighted the involvement of miR-195 in both the pathogenesis of type II diabetes as well as its ability to target a key player in fatty acid synthesis, the enzyme fatty acid synthase (FAS) [33, 34]. These findings prompted us to assess miR-195 levels in the current study. Overexpression of miR-195 was observed in the hibernating liver tissues of I. tridecemlineatus as well as reduced FAS protein levels in the same tissue suggesting a role for miR-195 in regulating a key node of the fatty acid synthesis pathway. A recent proteomics-based approach performed in hibernating I. tridecemlineatus has reported a reduction in enzyme levels notably involved in the lipid biosynthetic pathways further emphasizing the importance of this cascade in the hibernation phenotype [35]. The current findings add to this knowledge and provide yet another means by which hibernators can regulate fatty acid homeostasis during the torpor bouts. It will be interesting to further investigate additional miRNAs that can impact, directly or indirectly, the expression of enzymes involved in fatty acid synthesis including FAS and acetyl-CoA carboxylase (ACC). A list of miRNAs that can target key players of the fatty acid synthesis pathway is presented in Table 2. These notably include miR-424 and miR-451, which have been reported to impact FAS expression in osteosarcomas and ACC expression in smooth muscle cells, respectively [36, 37].

Table 2 miRNAs with associated transcript targets involved in fatty acid synthesis reported in non-hibernating models
Full size table

Increased miR-223 levels in the hibernating liver tissue samples were also reported in this study. A potential impact of this differential expression might be the translational regulation of FOXO1, a transcription factor with diverse functions. miR-223 can regulate FOXO1 expression in colorectal cancer cells [38]. FOXO transcription factors can modulate the transcription of genes leading to diverse cellular effects including oxidative stress and glucose metabolism [39]. Interestingly, differential transcript and protein levels of selected FOXO proteins were associated with stress resistance in the anoxia-tolerant turtle Trachemys scripta elegans [40]. Another relevant target that miR-223 could potentially regulate is the hepatic scavenger receptor class B type I (SR-BI), a protein involved in high-density lipoprotein cholesterol (HDL-C) uptake. A recent study demonstrated that miR-223 could interact with the 3′ UTR of SR-BI and prevent its translation, thus suggesting a potential implication of this miR in cholesterol metabolism [41]. Additional work will be required to further characterize miR-223 function in the torpid mammals, yet the strong overexpression observed in this study suggests that it should not be overlooked.

Interestingly, hibernating mammals are perceived as excellent natural models to assess the mechanisms underlying skeletal muscle plasticity and metabolism during prolonged periods of inactivity [42]. Differential expressions of MEF2 and MyoD, transcription factors with functions in skeletal muscle myogenesis, were reported in hibernating I. tridecemlineatus and provided a glimpse of the molecular switches involved in skeletal muscle dynamics at low temperatures [43]. A set of differentially expressed miRNAs likely involved in preserving pectoral muscle mass of the hibernating little brown bat Myotis lucifugus was recently reported [44]. The current study shows down-regulation of miR-378 in hibernating ground squirrel tissues and contributes to the growing list of miRNAs with a suspected implication in muscle integrity maintenance during the prolonged periods of inactivity as experienced by the torpid mammals. Interestingly, it was shown that MyoD could up-regulate miR-378 during myogenic differentiation in mouse myoblasts [45]. While a miR-378 transcript target with a role in skeletal muscle dynamics remains to be characterized to fully measure the significance of this measured change, differential expression of the MyoD-miR-378 axis in hibernating skeletal muscle tissues nevertheless reinforces its potential importance during hibernation.

The current work reported up-regulation of miR-152 and miR-486 levels in the hibernating liver tissues of I. tridecemlineatus. Differential expressions were also reported for these miRNAs in a study that looked at expressions of more than 200 miRNAs in liver tissue of another hibernating species, the Arctic ground squirrel Spermophilus parryii [46]. The transcripts targeted by these miRNAs as well as the significance of these variations during hibernation remain to be clarified. It is worth mentioning that work in non-hibernating model has linked miR-486 to a potential role in tumor progression and cell growth [47]. In addition, a recent study reported that miR-486 can directly target key components of insulin growth factor (IGF) signaling including insulin-like growth factor 1 (IGF1) and IGF1 receptor (IGF1R) in lung cancer cells [48]. Interestingly, IGF axis has been demonstrated to be significantly down-regulated in hibernating golden-mantled ground squirrels, Spermophilus lateralis [49]. Further work will be required to determine if these effects on IGF signaling are mediated by miR-486.

In conclusion, results presented here evaluated the expression levels of six miRNAs with potential metabolic relevance in three tissues of a small mammalian hibernator. Five of them displayed differential expressions in hibernating tissue samples including miR-195 in liver, a miRNA that can target a key enzyme involved in fatty acid synthesis. Reduced protein levels of this enzyme were also detected in the same tissue. Overall, this work further reinforces the potential implications of miRNAs in torpid animals and paves the way for a more detailed analysis of the specific pathways and transcripts targeted by miRNAs during hibernation.