Life-long reduction in myomiR expression does not adversely affect skeletal muscle morphology

We generated an inducible, skeletal muscle-specific Dicer knockout mouse to deplete microRNAs in adult skeletal muscle. Following tamoxifen treatment, Dicer mRNA expression was significantly decreased by 87%. Wild-type (WT) and Dicer knockout (KO) mice were subjected to either synergist ablation or hind limb suspension for two weeks. There was no difference in muscle weight with hypertrophy or atrophy between WT and KO groups; however, even with the significant loss of Dicer expression, myomiR (miR-1, -133a and -206) expression was only reduced by 38% on average. We next aged WT and KO mice for ~22 months following Dicer inactivation to determine if myomiR expression would be further reduced over a prolonged timeframe and assess the effects of myomiR depletion on skeletal muscle phenotype. Skeletal muscle Dicer mRNA expression remained significantly decreased by 80% in old KO mice and sequencing of cloned Dicer mRNA revealed the complete absence of the floxed exons in KO skeletal muscle. Despite a further reduction of myomiR expression to ~50% of WT, no change was observed in muscle morphology between WT and KO groups. These results indicate the life-long reduction in myomiR levels did not adversely affect skeletal muscle phenotype and suggest the possibility that microRNA expression is uniquely regulated in skeletal muscle.

For Pax7 and laminin staining of plantaris muscle sections were fixed in 4% PFA, blocked in 1% Tyramide Signal Amplification kit (#T20935, Invitrogen, Carlsbad, CA) and incubated with anti-laminin antibody (1:100, L9393, Sigma) overnight at 4 °C. Following the overnight incubation, muscle sections were subject to epitope retrieval using sodium citrate (10 mM, pH 6.5) at 92 °C for 11 min. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide in PBS followed by an additional blocking step with Mouse-on-Mouse Blocking Reagent (Vector Laboratories, Burlingame, CA). Incubation with anti-Pax7 antibody (1:100, Developmental Studies Hybridoma Bank, Iowa City, IA) was followed by incubation with the biotin-conjugated secondary antibody and streptavidin-HRP included within the Tyramide Signal Amplification kit (#T20935, Invitrogen, Carlsbad, CA). Sections were mounted with Vectashield fluorescent mounting media with DAPI.
www.nature.com/scientificreports www.nature.com/scientificreports/ Immunohistochemistry quantification. All cross-sections images were acquired with an upright microscope (AxioImager M1; Zeiss, Göttingen, Germany) at 20x magnifications. To capture the entire Pl, Sol and Gastroc muscle cross-section, [15][16][17][18][19][20][9][10][11][12] and 110-120 images, respectively, were obtained. MyoVision analysis was used to quantify fiber-type composition, fiber-type specific cross-sectional area, and myonuclear number as described by us 22 . Fibers were classified as type 2A (green), type 2B (red) and type I (pink). Fibers that were counted as negative under all three channels were classified as type 2X. For the plantaris and soleus muscles, 400-800 fibers were counted per cross-section per animal whereas 2000-5000 fibers per cross-section were counted for each gastrocnemius muscle per. Satellite cell abundance was determined using the AxioVision Rel software (v4.8) with Pax7+/DAPI+ loci found within the laminin border scored as a satellite cell. Wheat germ agglutinin (WGA) staining was used to quantified the level of muscle fibrosis and employed a MATLAB script to calculate the total WGA area as previously described 23,24 . Single fiber analysis. Single fibers were isolated as previously described by Wada and coworkers 25 . Briefly, after careful removal of the gastrocnemius and soleus muscles, the plantaris muscle was left on the tibia and fixed in situ at resting length in 4% paraformaldehyde for 48 hr. Fixed muscle was divided into several bundles by pulling the tendon with tweezers and non-muscle tissue was removed. The muscle bundles were incubated in 40% NaOH for 3 hr at room temperature and then shaken to separate the bundle into single myofibers. Single myofibers were washed with PBS before being stained with DAPI for nuclear visualization. Suspended fibers were dispersed on a slide and mounted with Vectashield fluorescent mounting media. Images were captured using 20X magnification Z-stacks on an Axioimager MI upright microscope and all fiber and nuclear measurements were made using AxioVision Rel software (version 4.8). Twenty fibers from each animal were measured for fiber width (µm) and subsequently the nuclei from each fiber were counted to determine the number of myonuclei per defined fiber segment as described by Jackson and colleagues 18 . Gene expression. Total RNA was isolated from Gastroc muscle previously frozen in liquid nitrogen. Samples were homogenized in a tissue homogenizer (Bullet Blender, Next Advance Inc., Averill Park, NY) using TRIzol, according to the manufacturer's instructions. Following isolation, RNA samples were treated with TURBO DNase Quantification of mRNA and miRNA expression. qPCR reactions were performed in the ABI 7500 qPCR system as described by the manufacturer. qPCR efficiency was calculated by linear regression during the exponential phase using LinRegPCR software v11.1 26 . The comparison of mRNAs or miRNAs expression between WT and KO was determined following normalization with Rpl38 (Mm03015864) or Rnu6, respectively. Relative quantification of mRNA and miRNA expression were assessed by using REST software 2009 v2.0.13 27 . exon analysis. For exon deletion analysis, diluted cDNA was amplified by PCR with GoTaq ® G2 Hot Start Green Master Mix (Promega) using the following primers: exon 21 forward, 5′-TGGGAACGCTAAC ACATCTACCTCAGA-3′ and exon 26 reverse 5′-CTGGTTCCATCTCGAGCAATTCTCTCACTG-3′. PCR products were gel purified (Zymoclean Gel DNA recovery Kit, Zymo Research) and then sequenced (Eurofin Genomics, Louisville, KY).

Northern blot.
To determine miR-1 expression, Northern blot analysis was performed using total RNA isolated from Gastroc muscles as described above. Denatured RNA (95 °C for 5 min in gel loading buffer) was separated on 15% urea-polyacrylamide gel (National Diagnostics) in a Mini-Protean vertical electrophoresis cell (Bio-Rad). To confirm equal loading prior to transfer, gels were stained with ethidium bromide (4 µg/ ml) to determine tRNA abundance. Samples were transferred to positively charged, nylon membranes, using a Trans-Blot Semi-Dry Transfer Cell (Bio-Rad) for 75 min at 15 V in a cold room. Saturated 3MM Whatman chromatography paper in 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC, Sigma) reagent was used to cross-link the RNA to the membrane for 2 hr 60 °C. After thoroughly rinsing the membrane with distilled water, ULTRAhyb hybridization buffer (Ambion) was used to pre-hybridize the membrane at 37 °C for at least 30 min. Following this step, the membrane was hybridized in the same solution using 0.5 nM of DIG-labeled LNA probe (5′-ATACATACTTCTTTACATTCCA-3′) following the manufacturer's instructions (DIG Oligonucleotide Tailing Kit -2 nd Generation, Roche) overnight at 37 °C. After hybridization the membrane was washed at low stringency in 2xSSC with 0.1% SDS at 37 °C twice for 15 min, at high stringency in 0.1 SSC with 0.1% SDS at 37 °C twice for five minutes, followed by 10 min at washing buffer 1xSSC at 37 °C. The membrane was then incubated in blocking buffer (DIG wash and block buffer set, Roche) for 3 hr at room temperature followed by incubation for 30 min with DIG antibody (1:15000) in blocking buffer. The membrane was then washed in DIG washing buffer four times for 15 min. Probe detection was performed using the DIG luminescent detection kit (CSPD, Roche) according to manufacturer's protocol.
www.nature.com/scientificreports www.nature.com/scientificreports/ statistical analysis. All data was expressed as the mean ± SEM. Data were analyzed with GraphPad software (GraphPad Software, Inc., La Jolla, CA) using a two-tailed t-test for unpaired samples (aging experiments) or via a two-factor ANOVA (hypertrophy and atrophy experiments). If a significant interaction was detected, an appropriate post hoc analysis was employed to determine the source of the significance. Statistical significance was accepted at p ≤ 0.05.

Results
To determine the necessity of miRNAs in regulating skeletal muscle mass, we sought to deplete miRNAs by inactivating Dicer specifically in adult skeletal muscle. The myofiber-specific, inducible Cre mouse strain (HSA-MCM) was crossed to a floxed Dicer (Dicer f/f/ ) strain to generate HSA-MCM; Dicer f/f mice, designated HSA-Dicer. Adult (at least 16 weeks of age) HSA-Dicer mice were treated with either vehicle (15% in sunflower seed oil) or tamoxifen (2 mg) and then recombination assessed using genomic DNA isolated from liver, heart or skeletal muscle. As shown in Fig. 1, Cre-mediated deletion of floxed exon 23 and 24 only occurred upon tamoxifen treatment and only in skeletal muscle. Quantitative PCR for the floxed exons showed that Cre-mediated recombination was highly effective with full-length Dicer expression significantly lower by 87% in the KO group compared to the WT group (see Fig. 1C).
We next determined the effect of Dicer inactivation on skeletal muscle hypertrophy and atrophy. Following two weeks of synergist ablation (SA), plantaris muscle weight was significantly increased by ~2-fold with no difference between WT and KO groups (see Fig. 2A). Conversely, following two weeks of hind limb suspension, the muscle weight of the soleus, plantaris and gastrocnemius muscles were significantly decreased approximately 30%, 11% and 13%, respectively, with no significant difference between WT and KO groups ( Fig. 2B-D, respectively). Despite the almost 90% knockdown in Dicer mRNA, expression of the canonical myomiRs, miR-1, -133a and -206, were only decreased by 42%, 31% and 40%, respectively, following four weeks of inactivation (see Fig. 2E-G, respectively).
The relatively modest decrease in myomiR expression following Dicer inactivation suggested that myomiRs, and likely miRNAs in general, might be very stable in adult skeletal muscle, requiring a longer period of time to achieve the 80-90% depletion of miRNA levels observed in skeletal muscle and other tissues following Dicer inactivation 20,28-35 . To test this idea, we aged WT and KO mice for an additional ~18 months and then assessed skeletal muscle phenotype. As shown in Fig. 3A, Dicer mRNA expression remained significantly lower by 80% in 22-month-old KO muscle compared to WT. To confirm deletion of Dicer exon 23 and 24, we cloned the portion of the Dicer mRNA containing exons 22-25 from a 22-month-old skeletal muscle cDNA library generated from either WT or KO mice. As shown in Fig. 3B, no full-length Dicer mRNA was detected in KO muscle and sequencing of the clone Dicer mRNA KO fragment reveal deletion of exons 23 and 24.
We next performed qPCR to determine if myomiR expression was further reduced with aging in the Dicer KO. As shown in Fig. 3C, expression of miR-1, -133a and -206 were significantly lower by 38%, 53% and 47%, respectively, in KO skeletal muscle compared to WT; however, the average decrease in myomiR expression was modest with aging as it went from 38% in young KO skeletal muscle to 46% in aged KO skeletal muscle. We also observed a similar significant decrease of ~40% in the expression of non-myomiRs (miR-16, let-7b and let-7c) (see Fig. 3C). In agreement with the qPCR results, Northern blot analysis showed a marked reduction in miR-1 expression in the KO compared to WT (see Fig. 3D). Expression of primary miRNA-1A and -1B (pri-miRNA-1A/B) was significantly higher by 2.1-fold in Dicer KO skeletal muscle compared to WT, a finding consistent with the loss of Dicer activity (see Fig. 3E).
Although we were unable to deplete miRNAs to levels typically reported in other tissues following Dicer inactivation, we thought it would still be informative to determine the effect of a ~50% depletion of myomiRs on skeletal muscle phenotype. As shown in Fig. 4A-B, there was no difference in body weight or the weight of various www.nature.com/scientificreports www.nature.com/scientificreports/ skeletal muscles (plantaris, soleus, gastrocnemius, EDL or TA) or heart between WT and KO groups. In agreement with the muscle weight results, there was no difference in the fiber-type specific muscle fiber cross-sectional area of the soleus, plantaris or gastrocnemius muscles between WT and KO groups (see Fig. 5A-D). Moreover, as shown in Fig. 5E, there was no change in the fiber-type composition of the soleus, plantaris or gastrocnemius muscles of KO mice in comparison to WT mice; this analysis also demonstrated there was no difference in muscle fiber number for each of the different muscles under investigation. Myonuclear abundance, as assessed by cross-section or single fiber, was not different between WT and KO groups; accordingly, the calculated myonuclei/fiber length from the single fiber analysis was the same between WT and KO groups (see Fig. 6A-D). There also was no difference in the calculated myonuclear domain size between WT and KO groups (see Fig. 6E). As presented in Fig. 7, wheat germ agglutinin staining revealed no difference in the level of skeletal muscle fibrosis of the soleus, plantaris or gastrocnemius muscles between WT and KO groups. Finally, as shown in Fig. 8, Pax7 immunohistochemistry of plantaris muscle cross-sections showed no difference in satellite cell abundance normalized to fiber number between WT and KO groups.

Discussion
The initial purpose of the present study was to test the hypothesis that miRNAs are necessary to regulate skeletal muscle mass in response to either a hypertrophic or atrophic stimulus. To test this hypothesis we generated an inducible, skeletal muscle-specific Dicer knockout mouse designated HSA-Dicer. We chose to use the myofiber-specific inducible Cre (HSA-MCM) mouse because it has been reported to promote effective recombination in both slow-and fast-twitch fibers with minimal expression in the heart 16 . Myofiber-specific expression of inducible Cre is driven by the human skeletal muscle α-actin promoter (HSA); the HSA promoter contains 2,000 bp of human skeletal α-actin 5′-flanking sequence plus the first exon and 149 bp of the first intron and shown by Muscat and Kedes to promote robust, skeletal muscle-specific expression 36 . Following tamoxifen treatment, Dicer mRNA expression was 87% lower compared to WT mice, thus indicating effective inactivation of Dicer. The inactivation of Dicer resulted in deletion of exons 23 and 24 which encode the RNase III domain of the DICER protein; the RNase III domain is absolutely necessary to cleave the precursor miRNA stem loop to generate the mature miRNA [37][38][39][40][41] .
We hypothesized that the depletion of miRNAs following Dicer inactivation would affect the ability of adult skeletal muscle to adapt to a change in use. Based on our previous work, we thought the depletion of miR-1 in and -206) and non-myomiR (miR-16, let-7a and let-7b) expression remained significantly lower with aging in skeletal muscle of KO mice compared to WT mice. (D) Northern blot analysis showed miR-1 expression was significantly lower by ~50% in skeletal muscle of KO mice compared to WT mice, consistent with qPCR results presented in panel C. (E) qPCR analysis showed a two-fold increase in both miR-1 primary miRNA (pri-miR-1A or -1B) transcript levels in skeletal muscle of KO mice compared to WT mice indicating an inability to effectively process the pri-miRNA transcript. Values are presented as the mean ± SEM (n = 8-9) with expression normalized to WT. Asterisk denotes significant difference between WT and KO groups (p ≤ 0.05). www.nature.com/scientificreports www.nature.com/scientificreports/ particular would enhance skeletal muscle hypertrophy as well as protect against muscle atrophy 14,42 . This hypothesis is supported by evidence showing miR-1 targets pro-growth genes such as IGF-1 and cyclin D [43][44][45] . Despite the significant knockdown of Dicer mRNA expression, we observed no difference in the change in skeletal muscle mass between WT and KO in response to either a hypertrophic or atrophic stimulus. To confirm Dicer inactivation resulted in miRNA depletion, we assessed by qPCR myomiR expression in WT and KO skeletal muscle; however, despite robust Dicer inactivation, myomiR (miR-1, -133a and -206) expression was only lowered by ~40% in the KO compared to WT skeletal muscle. We also found that non-myomiR (miR-16, let-7a and Let-7b) expression was reduced by a similar magnitude indicating myofiber miRNA expression in general was depleted by approximately 40% following a lifetime of Dicer inactivation. This finding was completely unexpected given the previous success of this floxed Dicer strain to effectively deplete miRNAs in other tissues 20,[29][30][31][32][33][34][35] . Thus, the relatively modest depletion of myomiRs in KO muscle more than likely accounts for there being no difference in the hypertrophic or atrophic response between WT and KO skeletal muscle.
Our results are in stark contrast to O'Rourke and colleagues who reported a decrease in skeletal muscle mass as the result of hypoplasia following skeletal muscle-specific Dicer inactivation during embryonic development (E9.75-E12.5) 28 . This dramatic phenotype was associated with a ~80-90% decrease in myomiR expression after two days of Dicer inactivation; moreover, the significant reduction in myomiR expression occurred despite there still being an estimated 30% of wild-type Dicer transcript present 28 . We, on the other hand, were able to knockdown Dicer expression by almost 90% but myomiR expression by only roughly 40%. Further, as O'Rourke and coworkers suggested, the incomplete knockdown of Dicer expression is the result of Dicer expression in non-muscle tissue found within skeletal muscle such as fibroblasts, endothelial and neural tissue. Although we do not have a definitive answer, we speculate the discrepancy in the effectiveness of myomiR depletion between the two studies is likely related to how the loss of Dicer differentially effects miRNA turnover in mitotic versus post-mitotic cells -i.e., embryonic vs. adult skeletal muscle 46 . While the results of the hypertrophy and atrophy studies were inconclusive because of the relatively modest decrease in myomiR expression, they did suggest to us that myomiRs may be quite stable in adult skeletal muscle. As an initial effort to test this idea, we aged mice for 22 months following myofiber-specific Dicer inactivation and then measured myomiR expression as well as assessed skeletal muscle phenotype. Although Dicer expression remained significantly lower in aged skeletal muscle of KO compared to WT, myomiR expression only showed a further decrease to ~50% in aged KO skeletal muscle. Northern blot analysis confirmed that miR-1 expression was reduced by approximately 50% in the KO compared to WT in agreement with the qPCR results.
The finding that myomiR expression was only reduced on average 50% following life-long Dicer inactivation was perplexing given the reported effectiveness of Dicer inactivation to deplete miRNA levels in other tissues 20,[29][30][31][32][33][34][35] . To confirm our qPCR results showing effective recombination following tamoxifen treatment, we cloned from WT or KO skeletal muscle cDNA libraries a fragment of the Dicer mRNA encompassing exons 22-25. Sequencing of the cloned fragment confirmed that we had in fact successfully deleted exons 23 and 24 in KO skeletal muscle. Moreover, as shown in Fig. 3B, we were unable to detect any full-length Dicer fragment in KO skeletal muscle, although we were able to detect a small amount of the recombined fragment in WT skeletal muscle. Together with our qPCR result, this finding provided compelling evidence that we had successfully induced highly effective recombination following tamoxifen treatment which resulted in deletion of exons 23 and 24, the exons encoding the all-important RNase III domain of the Dicer protein.
One limitation of the study was our inability to reliably detect Dicer protein by Western blot analysis using WT and KO whole-muscle lysates. We used both commercial (ab167444, sc393328 and ABIN2253355) and researcher-derived (clone 1414 and 1416) Dicer antibodies but were unable to produce consistent results using the same samples 29 . So, while we were able to effectively knockdown Dicer mRNA levels by 80-90%, it remains possible translation of the remaining Dicer mRNA is able to generate a sufficient amount of Dicer protein to maintain myomiR levels at 50%. While we fully acknowledge this possibility, we think, as suggested above, that most of the remaining full-length Dicer mRNA is coming from non-muscle tissue found within skeletal muscle.
An alternative explanation is the possibility that myomiRs may be very stable in adult skeletal muscle. In trying to explain our results, we came to realize that it might be telling us something unique about the regulation of miRNAs in adult skeletal muscle -that miRNAs in adult skeletal muscle may be very stable. To the best of our knowledge, no studies have measured miRNA half-life in skeletal muscle, either in vitro or in vivo; however, a number of studies have determined miRNA stability in other tissues, reporting half-lives between 5-9 days,