Deletion of muscle Igf1 exacerbates disuse atrophy weakness in mice

Abstract

Muscle atrophy occurs as a result of prolonged periods of reduced mechanical stimulation associated with injury or disease. The growth hormone/insulin-like growth factor-1 (GH/IGF-1) axis and load sensing pathways can both aid in recovery from disuse through their shared downstream signaling, but their relative contributions to these processes are not fully understood. The goal of this study was to determine whether reduced muscle IGF-1 altered the response to disuse and reloading. Adult male mice with inducible muscle-specific IGF-1 deletion (MID) induced 1 wk before suspension and age-matched controls (CON) were subjected to hindlimb suspension and reloading. Analysis of muscle force, morphology, gene expression, signaling, and tissue weights was performed in nonsuspended (NS) mice, and those suspended for 7 days or reloaded following suspension for 3, 7, and 14 days. MID mice displayed diminished IGF-1 protein levels and muscle atrophy before suspension. Muscles from suspended CON mice displayed a similar extent of atrophy and depletion of IGF-1, yet combined loss of load and IGF-1 was not additive with respect to muscle mass. In contrast, soleus force generation capacity was diminished to the greatest extent when both suspension and IGF-1 deletion occurred. Recovery of mass, force, and gene expression patterns following suspension were similar in CON and MID mice, even though IGF-1 levels increased only in muscles from CON mice. Diminished strength in disuse atrophy is exacerbated with the loss of muscle IGF-1 production, whereas recovery of mass and strength upon reloading can occur even IGF-1 is low.

NEW & NOTEWORTHY A mouse model with skeletal muscle-specific inducible deletion of Igf1 was used to address the importance of this growth factor for the consequences of disuse atrophy. Rapid and equivalent loss of IGF-I and mass occurred with deletion or disuse. Decrements in strength were most severe with combined loss of load and IGF-1. Return of mass and strength upon reloading was independent of IGF-1.

INTRODUCTION

Prolonged bed rest and limb immobilization are unfortunate consequences of illness and recovery from injury or surgery (1, 2). Individuals experience atrophy of their skeletal muscles due to the reduced mechanical stimulation and load being placed on the body (3, 4). It is well documented that lack of mechanical stimulus leads to a loss of muscle mass, reduction in muscle fiber cross-sectional area, diminished force generating capacity, and a shift of slow-to-fast muscle fiber types (4, 5). It is encouraging, particularly in young healthy individuals, that the effects of disuse atrophy are reversible through the use of rehabilitatory interventions such as progressive reloading of the muscle groups and resistance training (6, 7). However, the time required to return to normal activity of daily living levels increases with age or disease (8–10). In the elderly population, the reduction of the growth hormone/insulin-like growth factor-1 (GH/IGF-1) axis can prolong rehabilitation and this is compounded by existent diminished muscle mass and strength associated with sarcopenia, as well as a lack of regular activity (11–13). The more aggressive muscle wasting in cachexia has also been associated with reduced IGF-1 at the mRNA and protein levels (14) where concurrent impairments in IGF-1 ligand bioactivity and its signaling cascade has also been observed (14–16). Diminished IGF-1 is one of the multitude of factors that contribute to sarcopenia and cachexia, as increased myostatin levels and other transforming growth factor (TGF)-β superfamily members (17, 18), inflammation, disruption of autophagic flux, and heightened protein degradation (19) are also evident. Although we and others have implemented strategies to increase IGF-1 activity in muscle and have demonstrated improvements in and/or maintenance of muscle mass and function in aging and recovery following disuse atrophy (20, 21), the consequence that a specific reduction of IGF-1 has in the context of physiological alterations in these additional pathways has not been addressed.

A common model of disuse atrophy in animal studies is hindlimb suspension, which mimics the loss of load and still retains the ability for movement. Rodents subjected to hindlimb suspension exhibit significant loss of muscle mass most evident in postural muscles, such as the soleus, as well as the gastrocnemius, plantaris, and adductor longus (22). Furthermore, upon reloading, the time course of recovery displays early phases of transient muscle injury, boosts in protein turnover, and ultimately a return to normal levels of mass and force generating capacity (23). The role of IGF-1 in the prevention of muscle loss and its subsequent recovery has been thoroughly examined in terms of normal to supraphysiological levels of IGF-1. IGF-1, a known potent stimulus for coordinating muscle growth and repair, acts via the IGF-1 receptor (IGFR) to lead to a cascade of protein signaling that ultimately results in the upregulation of mammalian target of rapamycin (mTOR) activity and increased protein synthesis and a prevention of protein degradation through phosphorylation of the transcription factor forkhead box O (FoxO) (24–26). A mechanical stimulus, via progressive overload resistance training, electrical stimulation, or reloading of previously unloaded muscle, has also been shown to cause increases of growth hormone in circulation, upregulation of IGF-1 activity, and direct effects on the downstream targets of IGF-1 (27). Previous studies have shown that increases in mass as a result of functional overload can occur in the absence of IGF-1 signaling through IGFR (28). Furthermore, although IGF-1 can boost anabolic drive in the muscle, it cannot prevent atrophy during hindlimb suspension (29, 30). Benefits of IGF-1 are more apparent during recovery from disuse, where heightened local production of pro- and mature-IGF-1 can reduce the recovery time after a bout of hindlimb suspension (21). These previous findings demonstrate that increased IGF-1 can promote recovery and muscle hypertrophy but it may not be necessary. Consistent with this assertion, in a mouse model affording inducible deletion of skeletal muscle Igf1 [muscle-specific IGF-1 deletion (MID)] (31), ablation of muscle Igf1 in young adult male mice did not alter skeletal muscle mass 1 mo following deletion, and only transient growth deficits occurred in male MID mice when deletion was induced at birth. This points to the fact that muscle mass is regulated by many factors in addition to IGF-1. However, in the face of disuse, the extent to which loss of IGF-1 is a primary factor in regulating mass and function has not been clarified. Thus, we took advantage of the MID mouse model to determine whether the lack of muscle IGF-1 would exacerbate disuse atrophy following hindlimb suspension or impair the muscle recovery upon reloading, ultimately asking the question: Is muscle-derived IGF-1 necessary for the maintenance of functional muscle mass?

METHODS

Animal Studies

Animal studies were performed within the guidelines of the Institutional Animal Care and Use Committee and approved by the University of Florida. All mice used in this study were male aged 21.5 ± 5 wk and housed in the animal facility with a 12-h:12-h light/dark cycle. The mice had ad libitum access to food and water throughout the experimental procedures.

Muscle IGF-1-deficient mice (MID) were generated from crossing mice harboring the floxed exon 4 of Igf1 gene (Igf1^fl/fl) (12663; all strains from The Jackson Laboratory, Bar Harbor, ME) (32) with mice also harboring the doxycycline (DOX)-inducible Cre recombinase, driven by the human skeletal actin (HSA) promoter (HSA-rtTA/Dox-Cre) (012433). The MID mouse line was further crossed with Gt(ROSA)26Sor_tm1Sor targeted mutant mice (ROSA) (003474) carrying a loxP flanked neo cassette upstream of a β-galactosidase (lacZ) sequence that is responsive to Cre recombinase expression and backcrossed more than five times onto the C57Bl/6J strain (000664) to eliminate strain-dependent variability and was characterized previously by our laboratory (31). DOX was dissolved in the animals’ drinking water (2 mg/mL) with 5% sucrose added to offset the bitter taste. The DOX solution was held in light-proof water bottles and changed every 3–4 days. Mice had free access to the treated water 7 days before and up until termination of the experiments. Degree of deletion of exon 4 of the Igf1 gene was validated through use of end point PCR on cDNA generated from the gastrocnemius muscle (Fig. 1, A and B), with a lower second band indicating removal of exon 4. Densitometry of the bands was performed using ImageJ. Primers used for deletion validation are listed in Supplemental Table S1 (all Supplemental material is available at https://doi.org/10.6084/m9.figshare.13764433.v1), as Igf1 exon 3–6. Control mice (CON) lacking floxed Igf1 but containing HSA-rtTA/DOX-Cre with and without the ROSA reporter also received the same treatment as the MID mice. In contrast to our previous study, all animals received DOX throughout the experiment to control for a drug effect and to assure maintenance of muscle fiber deletion of Igf1 in the event of satellite cell recruitment.

Figure 1.

Validation of muscle-specific IGF-1 deletion (MID) model and experimental design. A: visual representation of the Igf1 gene including the lox-P sites flanking either side of exon 4 (L). Black arrows indicate the location of the primers used to determine deletion of exon 4. Also included are the regions of analysis for extent of insulin-like growth factor- 1 (IGF-1) deletion (Deletion PCR Analysis) and Igf1 expression (qPCR Expression Analysis) to show how deletion of exon 4 did not overlap with the qPCR primers. B: endpoint RT-PCR to demonstrate exon 4 deletion from muscles of control (CON) and MID mice. Upper band indicates a normal Igf1 gene and the lower band indicates loss of exon 4. C: image of mouse undergoing hindlimb suspension. Diagram of the time course of deletion and hindlimb suspension for the study. D: soleus muscle cross sections stained with X-Gal and counterstained with Eosin-Y, all obtained from MID mice undergoing doxycycline (DOX) treatment. Blue fibers arise from activation of the ROSA reporter by Cre-recombinase, which remained active throughout all conditions. Scale bar represents 200 µm. NS, nonsuspended; SUS, suspended.

Hindlimb Suspension and Reloading

We induced disuse atrophy in the hindlimb by using the suspension model described previously (21). Seven days after the beginning of DOX treatment, mice were suspended under light anesthesia (isoflurane). Their tails were attached to a short metal chain via foam tape (Skin Trac; Zimmer, Warsaw, IN). The suspension cages had a crossbar along the center of the cage on which a second small bar could move to allow greater movement ability for the animals. The metal chain was attached to the small bar and the length was adjusted so that only the hindlimbs of the mice were suspended. The mice were able to move via their forelimbs by using the metal grid on the cage floor and were able to maintain normal activity (Fig. 1C). Mice were monitored and hand cleaned at least once a day during the suspension period. After 7 days of suspension, the mice were either euthanized for analysis [suspended for 7 days (SUS)] or returned to a standard cage to allow reloading of their hindlimb muscles for 3, 7, or 14 days (Re3, Re7, and Re14, respectively). Food monitoring during suspension and reloading displayed no apparent differences with respect to suspension condition or genotype. Nonsuspended (NS) CON and MID mice were included for comparison and were euthanized for analysis after 7 or 21 of DOX treatment to control for the effects of time. The experimental design is shown in Fig. 1C. For each strain and timepoint, n = 6–12 mice were used. Multiple cohorts of mice were examined over time; however, each cohort retained a set of nonsuspended mice to control for any drift over time in data acquisition and to bracket all experimental animals with individual controls.

Tissue Collection and Muscle Function Testing

At each experimental timepoint, mice were anesthetized using a combination of xylazine (80 mg/kg) and ketamine (10 mg/kg) to allow removal of the soleus muscles for functional testing and histological analysis, as described in Muscle Fiber Morphology. Mice were then euthanized and additional skeletal muscles (including extensor digitorum longus (EDL), tibialis anterior (TA), gastrocnemius, quadriceps, and a strip of diaphragm), adipose tissue, and other internal organs were harvested and processed for histological and biochemical analysis. The liver, serum, epididymal fat [epididymal adipose tissue (eWAT)], and retroperitoneal fat [retroperitoneal adipose tissue (rWAT)] and one of each muscle were blotted, weighed, and frozen in liquid nitrogen for biochemical analysis, and the contralateral muscles pinned at a fixed length, surrounded in optimal cutting temperature compound (Sakura, Torrance, CA), and frozen in liquid nitrogen-cooled isopentane for histological analysis.

For each mouse, one soleus muscle was used for isolated force measurements before freezing for morphological analysis. The contralateral soleus was processed, as described above, for biochemical analysis. The muscles for mechanics were placed in a Sylgard-based petri dish and held at approximately resting length by minutien pins and bathed in Ringer’s solution [(in mM) 120 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 KH₂PO₄, 1.2 MgSO₄, 25 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), and 5.5 glucose] gas equilibrated with 95% O₂ and 5% CO₂. Loops on the proximal and distal muscle tendons were tied with nonabsorbable braided silk sutures (Fine Science Tools, Foster City, CA) and the muscle was transferred into an in vitro apparatus (Model 800 A, Dual Mode Muscle Lever 300 C; Aurora Scientific, Aurora, ON, Canada) filled with gas equilibrated Ringer’s solution at 22°C. The distal tendon was attached to a rigid post, and the proximal tendon was attached to a force transducer at resting length. The optimal muscle length (L_o) was established by adjusting the muscle length in isometric twitch conditions, until maximum force was obtained. At L_o, soleus muscles were stimulated for a period of 500 ms, with a series of 0.5-ms pulses at supramaximal stimulation and at 100-Hz frequency to determine maximum isometric tetanic force. Measurements were performed three times with resting periods of 5 min between tetanic stimulations. After the mechanical procedures, the muscles were blotted, weighed, and frozen for subsequent histological analysis, as described below.

Circulating and Tissue IGF-1 Content

Serum and gastrocnemius muscles were used for determining IGF-1 levels in response to loading conditions and to deletion of Igf1 exon 4. Mouse IGF-1 levels were measured with Rat/Mouse IGF-1 ELISA Kit (MG100; R&D Systems), according to the manufacturer’s recommendations. This kit detects total rodent IGF-1 and there is no cross-reactivity with insulin-like growth factor-2 (IGF-2) or IGF binding protein (IGFBPs). Data were obtained in duplicate with a SpectraMax M5 Plate Reader (Molecular Devices, Sunnyvale, CA) at 450 nm, and the results were averaged. Data are expressed in nanograms per milliliter, for serum, or picogram per microgram of total protein, for muscle.

Immunoblot Analysis

Snap-frozen, nonstimulated soleus muscles were powdered using a mortar and pestle in dry ice and homogenized in radioimmunoprecipitation assay (RIPA) buffer (9806; Cell Signaling) supplemented with phenylmethylsulfonyl fluoride (PMSF) (36978; Thermo Fisher Scientific), protease (P8340; MilliporeSigma), and phosphatase (P5726; Millipore) inhibitors. Homogenates were incubated in ice for 60 min and centrifuged at 15,000 g for 15 min. The supernatant was used for subsequent measurements, and protein quantification was determined by the Bradford assay (1863028; Thermo Fisher Scientific). Protein (40 µg) was loaded in 4%–12% Bis-Tris Midi Protein Gels (WG1402A, Thermo Fisher Scientific) for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), followed by transfer to Amersham Protran 0.45 NC nitrocellulose (10600003, GE HealthCare). The membranes were blocked for 90 min at room temperature with 5% BSA (BP9703100, Fisher BioReagents) in Tris-buffered saline (TBS). The following antibodies were used for overnight incubation at 4°C: Phospho-Akt (Ser473) (D9E) XP (1:2,000; 4060S, Cell Signaling Technology); Phospho-Akt (Thr308) (D25E6) XP (1:1,000; 13038, Cell Signaling Technology); Akt (pan) (40D4) (1:2,000; 9220S, Cell Signaling Technology); phospho-ERK1/2 (1:2,000, 9101, Cell Signaling Technology); and total-ERK1/2 (1:2,000; 9107, Cell Signaling Technology). Blots were incubated for 90 min at room temperature with the corresponding secondary antibody (1:15,000; Li-Cor Biosciences). After incubation with the secondary antibody, the blots were washed and then scanned with the Odyssey CLx Imaging System (Li-Cor Biosciences). The band intensities were automatically determined by the accompanying software Image Studio v.5.2 (Li-Cor Biosciences). Ponceau staining was used as a loading control for each blot, with the intensity of all bands in a lane used to normalize the intensity of a given band of interest for analysis.

Gene Expression Analysis

Total RNA was isolated from gastrocnemius muscles with TRIzol Reagent (15596018, Life Technologies), according to the manufacturer protocol. The extracted RNA was quantified with NanoDrop 2000 Spectrophotometer (ND-2000, Thermo Fisher Scientific) and RNA integrity was confirmed by gel electrophoresis. One microgram was treated with recombinant RNase-free DNAse I (04716728001; Roche, Basel, Switzerland). cDNA was generated from 500 ng of purified mRNA with the High-Capacity cDNA Reverse Transcription Kit (436881, Applied Biosystems). Duplicates of cDNA samples (10 ng) were amplified on the Step One Plus Real-Time PCR System or QuantStudio 3, using Power SYBR Green PCR MasterMix (4367659, Applied Biosystems). Genes of interested were referenced to the average of two separate housekeeping genes [18s and TATA-binding protein (Tbp)] to eliminate dependence of gene expression changes on loading condition or Igf1 deletion (Supplemental Fig. S1). All data were normalized to NS-CON for the respective gene and plotted as log₂ fold changes. The oligonucleotide primers are shown in Supplemental Table S1.

Muscle Fiber Morphology

Fiber size and type were determined on soleus cryosections (Supplemental Fig. S2) stained with laminin (rabbit Ab-1; Neomarkers, Fremont, CA) to outline each muscle fiber, SC-71 (36), BA-F8 (36), and BF-F3 (36) (Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA). Soleus muscles were stained for MHC I/β and IIa, and IIb (unstained fibers assumed to be IIx). Secondary antibodies used were, AlexaFluor 488 goat anti-mouse IgG (H + L) (A-11029), Alexa Fluor 488 goat anti-mouse IgM (µ-chain specific), Alexa Fluor 350 goat anti-mouse IgG (H + L), and Alexa Fluor 568 goat anti-rabbit IgG (H + L) (A-11011; all from Thermo Fisher Scientific). Negative control slides, incubated with secondary antibodies only, were also included. Whole soleus muscle cross sections from the midbelly from at least three mice per experimental conditions were used, where all fibers in the cross sections were measured.

Muscle sections were stained overnight with β-galactosidase activity, as previously described (31), to assess the DOX activation of the Cre-recombinase via the ROSA reporter. After staining, sections were counterstained with Eosin-Y to show fiber morphology.

Samples were visualized with a epifluorescence microscope (DMR, Leica Microsystems, Buffalo Grove, IL), equipped with a DFC7000T camera (Leica Microsystems, Buffalo Grove, IL). Images were acquired and processed with the Leica Application Suite and Microscope Imaging software (Leica Microsystems). Fiber typing and fiber quantification were achieved with semiautomatic muscle analysis using segmentation of histology (known as SMASH) software (33).

Statistical Analysis

Two-way ANOVA was used to assess main effects of Igf1 deletion, loading condition, or the interaction of the two, for all assays performed. Post hoc analysis using Dunnett and Sidak multiple comparisons were performed for potential differences between specified conditions (NS vs. each loading condition) or if deletion had an effect on loading conditions, respectively. A P value < 0.05 was considered statistically significant.

RESULTS

Muscle IGF-1 Content Is Independently Affected by Igf1 Deletion and Suspension

The first goal of this study was to determine whether Igf1 deletion and hindlimb suspension were independent modulators of muscle mass and function. DOX treatment was effective in deleting the floxed exon in Igf1 (Fig. 1A), as end-point RT-PCR used to assess the extent of deletion, showed that at least 50% of Igf1 transcripts lacked exon 4 in muscles from MID mice from all conditions (Fig. 1B). Additional confirmation was performed using β-galactosidase activity for evidence of the ROSA reporter activation by Cre recombinase in MID animals (Fig. 1D). Fibers that were positive for Cre recombinase ranged from 60% to 100% in the soleus from MID mice. There was no correlation between positive fibers and fiber type or loading condition (data not shown).

One week after DOX treatment, the effects of Igf1 deletion on IGF-1 levels in skeletal muscle and in the circulation were measured by ELISA measurements on muscle lysates from the gastrocnemius, as a representative posterior hindlimb muscle, and collected serum. Both DOX treatment and suspension independently caused a reduction in muscle IGF-1 levels. IGF-1 content decreased by 50% in muscles from NS MID mice when compared with those from NS CON mice (Fig. 2A). Furthermore, suspension for 7 days resulted in a similar 50% drop in IGF-1 content in muscles from CON mice, whereas muscles from MID mice did not lose additional IGF-1 during suspension. Circulating IGF-1 diminished only in the combined condition of suspension of MID mice, where it was 31% lower than that in suspended CON mice (Fig. 2B). To determine whether the reduction in IGF-1 levels was due to altered transcription, Igf1 expression was monitored with oligonucleotides recognizing exon 6 and 3′UTR (untranslated region) to avoid the region of deletion (Fig. 1B). Expression in nonsuspended MID mice was 40% lower than in CON mice, but did not achieve significance (Fig. 2C). Likewise, muscles from suspended CON mice displayed a similar nonsignificant reduction. There was high variability of Igf1 expression within each condition, which prevented our ability to ascribe decreases in Igf1 transcriptional changes to account for the loss of IGF-1 protein, but also suggests that additional contributors lead to reduced IGF-1.

Figure 2.

Effect of hindlimb suspension and/or Igf1 deletion on insulin-like growth factor- 1 (IGF-1) content and expression of IGF-1 regulators and IGF-1 signaling. A: muscle IGF-1 content was 50% lower in nonsuspended (NS) muscle-specific IGF-1 deletion (NS-MID) mice compared with NS-control (CON) mice and did not decrease further in suspension. IGF-1 content in suspended (SUS)-CON was decreased by 45%, reaching similar levels to those of SUS MID. B: circulating IGF-1 levels in MID mice, not significantly different from CON mice in the NS condition, was 31% lower relative to CON mice in the SUS condition. C–J: expression of Igf1 and regulators of IGF-1 activity. K–M: phosphorylation of Akt and ERK1/2, revealed load dependence of Thr308 Akt phosphorylation, with a significant reduction following suspension in muscles from CON mice. No changes in signaling of extracellular signal-regulated kinase (ERK) were observed as a result of deletion or suspension. Data shown are displayed as means ± SE for samples obtained from n = 3–7 mice (across all experiments represented). Number of symbols indicates statistical significance (e.g., */**/***/****: P < 0.05, 0.01, 0.001, and 0.0001); ns, not significant. Two-way ANOVA: ^αmain effect of deletion; ^βmain effect of unloading; ^γmain effect of interaction. Multiple comparisons by Dunnett and Sidak post hoc tests: *NS vs. loading condition within same deletion condition; †CON vs. MID within same loading condition.

To determine whether modulators of IGF-1 actions were also affected by suspension or deletion, expression patterns of regulators of IGF-1 bioavailability (34) were evaluated (Fig. 2, D and J). Expression of IGF binding protein 3 gene (Igfbp3) exhibited main effects of loading and a 3.3-fold increase in muscles from MID mice following suspension (Fig. 2D). IGF binding protein 4 gene (Igfbp4) decreased 2.6-fold in muscles from CON mice following suspension (Fig. 2E). IGF binding protein 4 gene (Igfbp5) expression (Fig. 2F) was not dependent upon deletion or suspension. Expression of additional upstream regulators were also measured with only load-dependent changes observed for levels of pregnancy-associated plasma protein A (Pappa; Fig. 2G), which can cleave IGFBPs (35), whereas Pappa2 (Fig. 2H) and stanniocalcin 2 gene (Stc2; Fig. 2I), which inhibits pregnancy associated plasma protein A (PAPP-A) protease activity (36) exhibited no apparent dependence on IGF-1 levels or on suspension. Expression of Igf1r and the two isoforms of the insulin receptor (Insr) was measured to determine whether receptor abundance changed in any experimental condition (Supplemental Fig. S3). The combined reduction of IGF-1 with suspension led to an increase in Igf1r, but no other changes were observed during suspension. As IGF-1 activity can be countered via inhibition of Akt phosphorylation by myostatin (Mstn), a negative regulator of muscle mass and protein synthesis (37), expression was assessed, with a 4.2-fold increase found in muscles from suspended MID mice (Fig. 2J). Finally, to determine whether any of the above changes resulted in alterations in IGF-1-dependent signaling, phosphorylation of Akt and extracellular signal-regulated kinase (ERK) were measured in soleus muscle lysates (Fig. 2, K–M). Phosphorylation of Thr308 on Akt, which is a substrate of PtdIns (20, 23, 38) P₃-dependent protein kinase-1 (PDK1) (38), was dependent upon loading, and there was a significant decrease in phosphorylation of this site in response to suspension in CON mice (Fig. 2L). Neither Akt Ser 473 nor ERK phosphorylation exhibited changes associated with deletion of suspension (Fig. 2, K and M).

Taken together, MID mice exhibit diminished IGF-1 both in muscle tissue and in the circulation, as previously shown (31). Importantly, loss of load or induction of deletion cause alterations in muscle IGF-1 levels within 1 wk, but these are not additive nor synergistic. However, it is primarily with the combined conditions that changes in gene expression are evident, including Igfbp3, Mstn, and Igf1r.

Igf1 Deletion Exacerbates Disuse-Associated Weakness and Atrophy

Previous studies have shown that increased IGF-1 levels in muscle can boost muscle mass when load is present, but upon unloading, a proportional loss of muscle occurs regardless of whether there is normal or high initial mass and IGF-1 (13, 29). By extension, low IGF-1 levels may also alter the initial set point, but not alter the proportional changes in mass associated with disuse or may result in more extensive atrophy. To address this question, soleus muscles from CON and MID mice were carefully harvested and assessed for force generation capacity and morphology (Fig. 3). The soleus muscles were chosen to represent muscles with well-recognized responses to load.

Figure 3.

Soleus muscle properties in response to Igf1 deletion and suspension. A: absolute soleus mass of muscle-specific IGF-1 deletion (MID) mice was 15% lower when compared with control (CON) in the nonsuspended (NS) groups while both CON and MID soleus mass decreased by at least 20% in the suspended (SUS) groups. B: only CON mice displayed reduction in normalized soleus mass (18%) as a result of suspension compared with NS. C: maximal tetanic force generation of soleus muscles was reduced following suspension in both CON and MID groups by 31% and 35%, respectively. MID soleus force was also impaired relative to CON by 16% in the NS and 22% in the SUS conditions. D: only SUS-MID muscle exhibited a reduction in specific force (26%) compared with NS-MID. E: both CON and MID soleus muscle exhibited reductions in muscle fiber cross-sectional area of 33% and 31%, respectively, in response to suspension. F: no changes to fiber type distribution were observed in any experimental group. Data shown are displayed as means ± SE for samples obtained from n = 5–11 mice (across all experiments represented). Number of symbols indicates statistical significance (e.g., */**/***/****: P < 0.05, 0.01, 0.001, and 0.0001). Two-way ANOVA: ^αmain effect of deletion; ^βmain effect of unloading. Multiple comparisons using Sidak post hoc tests: *NS vs. loading condition within same deletion condition; †CON vs. MID within same loading condition.

Absolute soleus mass was reduced 15% by deletion of IGF-1 (Fig. 3A), which was not apparent when normalized to body weight (Fig. 3B). With suspension, mass of soleus muscles from CON mice decreased by 25% and those of MID mice lost 18%. Thus, the mass of soleus muscles from suspended mice reached similar values regardless of genotype (Fig. 3A). For CON mice, the soleus displayed the greatest proportional loss of mass when compared with other muscles (Fig. 3B and Supplemental Table S2). In contrast, although MID mice also exhibited soleus muscle atrophy, the loss of body weight (16%, Supplemental Table S2) may have occurred across many tissues, negating the proportional mass changes specific to the soleus (Fig. 3B). Absolute maximal tetanic force of the soleus followed similar trends as absolute mass, with 16% lower force output in muscles from nonsuspended MID mice compared with CON mice (Fig. 3C). Upon suspension, both CON and MID mouse muscle tetanic forces were 30% lower. The consequences of the losses in absolute force became apparent when values were normalized to muscle cross-sectional area (CSA; Specific Force, Fig. 3D). The soleus muscles from MID mice lost 26% of their strength when subjected to suspension, whereas no other differences were observed. Taken together, the dual loss of load and muscle IGF-1 compounds the functional consequences of disuse.

To determine the dependence of mass and functional changes on muscle fiber properties, muscle cross sections were stained via immunofluorescence with laminin and myosin heavy chain (MHC) isoforms (Type I and IIa) (Supplemental Fig. S2) to determine CSA and any potential changes in fiber type, as a result of deletion and/or suspension (Fig. 3, E and F). Both the CON and MID suspended groups displayed a 33% and 31% reduction in median CSA from their NS counterparts but no differences between CON and MID were observed (Fig. 3E). Furthermore, the relative fiber type distributions in the soleus did not depend on Igf1 deletion or suspension (Fig. 3F).

Reloading Transiently Enhances Igf1 Expression Production and Activity

To complement the analysis of disuse atrophy in a diminished IGF-1 environment, we next addressed whether reduced IGF-1 impaired muscle recovery upon reloading. In gastrocnemius muscles from CON mice, IGF-1 levels were depressed at 3 days of reloading, but by 7 days, IGF-1 content returned to nonsuspended levels (Fig. 4A). On the other hand, muscle IGF-1 in MID mice remained at ∼50% of CON levels throughout the reloading period, never deviating from the initial NS-MID levels. Diminished circulating IGF-1 was apparent only in MID mice at 7 days of reloading (Fig. 4B). Transcriptional increases of Igf1 were observed in muscles from both CON and MID mice (Fig. 4C), where there was a twofold increase at Re3, preceding the changes in IGF-1 protein in CON mice.

Figure 4.

Effect of hindlimb reloading and/or Igf1 deletion on insulin-like growth factor (IGF-1) content, and expression of IGF-1 regulators and IGF-1 signaling. A: gastrocnemius IGF-1 content in control (CON) groups returned to nonsuspended (NS) levels following 7 days of reloading, whereas muscle-specific IGF-1 deletion (MID) muscle displayed no changes in IGF-1 content. B: serum levels of circulating IGF-1 did not differ from NS levels in either CON or MID mice following reloading of the hindlimb, although Re7-MID serum displayed 31% lower circulating IGF-1 compared with CON. C–J: expression of Igf1 and most regulators of IGF-1 activity are primarily dependent on loading. K: increased activation of Akt Ser473 phosphorylation in the soleus was observed in CON muscle 7 days following reloading. L: no change in Akt Thr308 phosphorylation was observed during reloading. M: soleus extracellular signal-regulated kinase (ERK) phosphorylation was increased at Re7 in CON but not MID muscle. Data shown are displayed as means ± SE for samples obtained from n = 3–7 mice (across all experiments represented). Number of symbols indicates statistical significance (e.g., */**/***/****P < 0.05, 0.01, 0.001, and 0.0001); ns, not significant. Two-way ANOVA: ^αmain effect of deletion; ^βmain effect of unloading; ^γmain effect of interaction. Multiple comparisons by Dunnett and Sidak post hoc tests: *NS vs. loading condition within same deletion condition; †CON vs. MID within same loading condition.

Similar evaluation of IGF-1 regulator gene expression was performed. Expression of Igfbp3-5, Pappa, and Stc2 displayed dependence upon loading, whereas Pappa2 and Mstn had no dependence upon loading or Igf1 deletion (Fig. 4, D–J). Only Igfbp5 exhibited a main effect of Igf1 deletion. Expression of receptor populations also shared the dependence upon loading for Igf1r and Insrb, and Insrb, the receptor isoform accounting for ∼60% of insulin receptor expression in the muscle (39), with continued increase of Igf1r in muscles from Re3 MID mice, and 3.8-fold increase of Insrb at Re7 in MID mice (Supplemental Fig. S3). The effects of deletion on the expression pattern of these receptors for nonsuspended MID muscle were consistent with our initial assessments of the receptor populations through 16 wk of age (31).

Upon reloading, soleus muscles from CON mice had twofold elevated P-AKT Ser 473 and P-ERK 7 days after reloading compared to nonsuspended conditions, with no significant change in Akt Thr308 phosphorylation (Fig. 4, K–M). In contrast, phosphorylation of AKT in MID muscle lysates remained unchanged, and P-ERK was approximately twofold lower than in lysates from nonsuspended MID mice. These results suggest that the changes in phosphorylation were dependent upon the tissue levels of IGF-1, given the similarity in patterns.

Recovery from Disuse Atrophy Does Not Require IGF-1

We previously demonstrated that recovery of soleus muscle mass and function from disuse is aided by viral-mediated upregulation of IGF-1, supporting that boosting IGF-1 within muscle is sufficient to enhance this process (21). However, whether reduced IGF-1 negatively impacts recovery has not been addressed. To determine whether MID muscles were deficient in the ability to regain mass or function, mice were examined at 3, 7, and 14 days of reloading. Soleus muscle mass displayed a main effect of Igf1 deletion, where the absolute masses of MID muscles were at most 10% lower than those from CON mice at all reloading timepoints (Fig. 5A). This difference disappeared when muscle masses were normalized to body weight (Fig. 5B).

Figure 5.

Loss of muscle insulin-like growth factor- 1 (IGF-1) does not impair functional recovery of soleus muscle upon hindlimb reloading. Return of load to the hindlimb resulted in improvements of absolute (A) and normalized (B) soleus mass to comparable levels of nonsuspended (NS) groups for both control (CON) and muscle-specific IGF-1 deletion (MID). C: CON soleus force production did not return to NS levels until the Re14 timepoint. However, MID soleus returned to NS levels by Re3. D: specific force of the soleus did not differ from NS values for either CON or MID upon the return of load. E: soleus median fiber cross-sectional area (CSA), initially still depressed at Re3, was no longer significantly less than NS CSA by Re7 for either CON or MID. F: no changes in fiber type distributions were observed as a result of deletion or load. Data shown are displayed as means ± SE for samples obtained from n = 5–11 mice (across all experiments represented). Number of symbols indicates statistical significance (e.g., */**/***/****P < 0.05, 0.01, 0.001, and 0.0001); ns, not significant. Two-way ANOVA: ^αmain effect of deletion; ^βmain effect of unloading. Multiple comparisons by Dunnett and Sidak post hoc tests: *NS vs. loading condition within same deletion condition; †CON vs. MID within same loading condition.

Muscle reloading is commonly associated with delays in recovery of force (21), and 21%–24% deficits in absolute tetanic force were observed in soleus muscles from CON mice when compared with those from nonsuspended CON mice; however, MID mice had no differences in absolute force output (Fig. 5C). More striking was the higher specific force measured at Re7 in MID mice when compared with CON mice (Fig. 5D), although by Re14, both mouse strains had regained all force generation capacity. Fiber sizes were predominantly affected by loading, as median fiber cross-sectional areas returned to nonsuspended values in a similar timeframe for both CON and MID mice. As with suspension, fiber type distributions were unaffected by reloading or deletion (Fig. 5F). These results suggest that subphysiological levels of IGF-1 do not impair the restoration of muscle mass and strength upon reloading. Only the nadir of functional deficits with disuse is deeper with the combined loss of muscle IGF-1 and load.

Although muscle accounts for ∼40% of body weight, other organs also contribute to mass and may be responsive to load or muscle IGF-1 content. In addition to muscles, normalized liver mass, in particular, was significantly reduced in both CON and MID mice following suspension (Supplemental Table S2). Normalized epididymal fat (eWAT) mass decreased by 39% in CON and 49% in MID during the reloading period. Whether the changes in nonmuscle tissue mass altered IGF-1 levels or other circulating factors were not pursued.

Atrogene Expression Is Primarily Dependent upon Load, Not IGF-1

Atrophy-related genes, or “atrogenes,” are well-known negative regulators of muscle mass, and are generally suppressed by high IGF-1 levels (40, 41). We examined expression of common atrogenes across the experimental conditions (Fig. 6, A–G). Load-dependent gene expression was most evident, with cell division control protein 48 gene (Cdc48), forkhead-box only protein 21 gene (Fbxo21) [specific of muscle atrophy and regulated by transcription (SMART)], forkhead-box only protein 30 gene (Fbxo30) [muscle ubiquitin ligase of the SCF complex in atrophy-1 (MUSA1)], forkhead-box only protein 32 gene (Fbxo32) (Atrogin 1/MAFbx), and Trim63 [muscle RING-finger protein-1 (MuRF-1)] exhibiting main effects of loading (Fig. 6, A–D, G). Fbxo32 also displayed dependence upon Igf1 deletion (Fig. 6D). Fbxo40 and tripartite motif containing 32 gene (Trim32) expression exhibited no dependence on any experimental conditions (Fig. 6, E and F). A general pattern for the altered atrogenes was exemplified by Fbxo32, with a 4.0-fold increase after 7 days of suspension in muscles of CON mice and a return to NS levels by Re3 (Fig. 6D). The increase in expression with suspension and subsequent downregulation during the early stages of reloading is consistent with the catabolic drive associated with disuse (19) and the anabolic drive during reloading. The absence of Igf1-dependent effects on atrogene expression suggests that normal IGF-1 levels are not sufficient to override the load-dependent changes associated with disuse and reloading.

Figure 6.

Change in load is the main determinate of atrogene expression. A: Cdc48 expression in muscle-specific IGF-1 deletion (MID) gastrocnemius muscle was 7.0-fold lower than control (CON) after 14 days of reloading. B and C: although both Fbxo21 and Fbxo30 displayed a main effect of loading on expression patterns, their expression did not significantly differ from nonsuspended (NS) or between CON and MID gastrocnemius muscle. D: Fbxo32 expression increased by 3.9-fold during suspension in CON gastrocnemius muscle but returned to comparable levels of NS upon reloading. MID expression of Fbxo32, already slightly elevated in the NS group, responded in a similar manner as CON until Re14. For Re14-MID, Fbxo32 expression was 7.7- and 8.2-fold lower than NS-MID and Re14-CON gastrocnemius muscle, respectively. E–F: neither Fbxo40 nor Trim32 displayed any significant changes to expression as a result of deletion or hindlimb suspension. G: expression of Trim63 was highly dependent on loading condition and followed a similar pattern of expression for both CON and MID gastrocnemius muscle as a response. Data shown are displayed as means ± SE for samples obtained from from n = 3–5 mice (across all experiments represented). Number of symbols indicates statistical significance (e.g., */**/***/****P < 0.05, 0.01, 0.001, and 0.0001); ns, not significant. Two-way ANOVA: ^βmain effect of unloading; ^γmain effect of interaction. Multiple comparisons: *NS vs. loading condition within same deletion condition; †CON vs. MID within same loading condition. SUS, suspended.

DISCUSSION

The interplay between IGF-1 and muscle load converges upon a common pathway that promotes anabolic actions and prevents catabolism (42–44). The regulation of muscle mass through their respective actions can be synergistic to enhance hypertrophy, as demonstrated in several animal models with increased IGF-1 levels combined with resistance training (45). In contrast, the combined loss of both IGF-1 and load can exacerbate atrophy, a situation commonly found associated with aging, where there is diminished drive through the GH/IGF-1 axis and a tendency for reduced physical activity (11–13). In more extreme cases of muscle wasting or disuse, catabolic processes are associated with reduced Igf1 expression (14) and inhibition of the IGF-1 signaling pathway, thereby preventing its activity even in the presence of the ligand (16). In light of this, many studies, including our own, have introduced high levels of IGF-1 in models of disuse to counter atrophy. These past results demonstrate that a new initial set point of increased mass can be achieved, but the removal of load causes a similar extent of atrophy (21, 29, 30). However, the return of load in the presence of enhanced IGF-1 levels can accelerate muscle recovery (21, 46). Because IGF-1 and load are so entwined, it has been an ongoing challenge to uncouple their specific roles. The goal of our current study was to employ acute disruption of muscle IGF-1 production in adult animals to examine the effects of disuse independently. This was enabled by using the MID mouse model recently developed in our laboratory (31).

We have previously evaluated the MID mouse model at monthly intervals either with induction of deletion at birth or in young adults. An impairment of growth was observed in young MID mice particularly in muscles, which was anticipated because of the well-known actions of IGF-1 in this period. This was not the case in 12- to 16-wk-old male mice, where after 1 mo of IGF-1 deletion, there was little change in normalized muscle mass and instead a metabolic phenotype emerged. Hence, it was unclear that whether the anabolic actions of IGF-1 would be evident more immediately after induction of deletion in 5-mo-old animals, which had not been evaluated in the former study. The diminished IGF-1 content of 50% only 1 wk after deletion points to the rapid turnover of the local IGF-1 pool in muscle (Fig. 2A). In addition, the loss of local IGF-1 was associated with a significant loss of Thr308 phosphorylation of Akt, downstream of IGF-1 receptor activation (38), and muscle mass within the same timeframe (Figs. 2L and 3A). Thus, removing the progrowth actions of this IGF-1 was sufficient to cause acute muscle loss, and the reduced P-Akt at the Thr308 residue was consistent with the reduction in IGF-1. We initially interpreted this as a new set point, mirroring that which occurs with high IGF-1 levels. Furthermore, in keeping with this model, we anticipated that suspension of both MID and CON mice would result in a similar extent of atrophy, maintaining the proportional differences in mass between the two strains. Instead, we found that muscles in MID mice did not lose mass with unloading, and the absolute muscle masses in MID and CON mice were similar following suspension. The loss of muscle mass in CON mice was accompanied by a 50% decrease in IGF-1 protein without a robust reduction in Igf1 gene expression, ultimately reaching the same levels of IGF-1 that had already occurred in MID mice before suspension. Taken together, this suggests that a significant factor leading to disuse atrophy is the active targeted degradation of the leading anabolic driver, namely, IGF-1.

Where the consequences of combined loss of load and IGF-1 are more apparent in the assessment of function. In the soleus, the loss of mass and force go hand in hand in CON mice, such that specific force is preserved across all conditions. However, tetanic forces in soleus muscles from MID mice decreased to a greater extent than mass during suspension, leading to reduced specific forces (Fig. 3, C and D). This implies that IGF-1 and load have additive actions on force generation capacity, unlike their actions on mass. This distinction may only be apparent with shorter durations of disuse, such as those in the current study, but we were unable to extend the suspension time without impairing animal welfare.

The return of load following disuse is accompanied by an early and transient phase of damage that can slow the process of rehabilitation. This was most evident in the assessment of soleus fiber sizes, which did not return to normal values until 7 days after reloading (Fig. 5E), but what was striking was that lack of IGF-1 did not cause a greater delay in fiber size recovery. In fact, functional recovery was just as robust in MID mice as CON mice (Fig. 5). Thus, for reloading, the recovery process appears to be governed primarily by load, even though there is clearly increased Igf1 expression in both strains, resulting in higher IGF-1 levels in CON mice over the same time period. Related studies addressing the necessity of the IGF-1 pathway for muscle hypertrophy have been performed. Using the MKR strain of mice, a model harboring dominant negative variant of the IGF-1 receptor in striated muscle (47, 48), the response of muscle to functional overload in the plantaris muscle was examined. The group found that muscle’s ability to grow was similar degree in control and MKR mice (28). This previous study, like ours, suggests that load is sufficient to regulate muscle mass during recovery or overload. However, this does not eliminate the benefits of IGF-1 actions. Boosting IGF-1 levels accelerates recovery of mass and force production in models of disuse (21, 46). Thus, for recovery from disuse, IGF-1 and load are synergistic both in the extent and speed for regaining functional muscle mass.

The deletion strategy in the MID mouse model ablates exon 4 in muscle fibers, and Cre activity is evident throughout muscle cross sections following DOX treatment. Even so, the extent of deletion, however, around 50%, both by analysis of exon 4 excision, and protein measurements by ELISA. Even with extended treatment durations, we have not observed total IGF-1 in skeletal muscle lysates to fall below 50% of endogenous levels. These observation point to the facts that muscle can be supplied with IGF-I through the circulation and is a sink for IGF-1. Furthermore, several additional cell types within the muscle tissue express and produce IGF-1 (49). In all cases, the IGF-1 reservoir is stabilized in the extracellular matrix directly or indirectly through association with its family of binding proteins. Hence, this pool of IGF-1 may remain unaltered, even when muscle fiber production is blocked.

To capture the major findings of this study, we find that there are independent, additive, and synergistic actions of muscle load and IGF-1 on the maintenance of muscle mass and strength. As shown in Fig. 7, loss of muscle mass and force production capabilities with disuse is associated with an inherent loss of IGF-1, compounding the apparent atrophy. Removing IGF-1 but retaining load points to the proportion of mass and strength that is IGF-1 dependent only. Returning load to muscle does the reverse, in that Igf1 expression/production is increased. However, load-dependent changes predominate, suggesting that endogenous IGF-1 is not sufficient to alter the recovery trajectory. Only with supraphysiological levels it is possible to accelerate rehabilitation.

Figure 7.

Interplay of load and insulin-like growth factor-1 (IGF-1) on functional response to disuse atrophy and recovery. Deletion of Igf1 expression causes acute atrophy, minimizing the effect of removing load on mass. Control muscles lose proportionally similar mass with disuse, which is accompanied by a significant reduction in IGF-1 content. Consequences of combined loss of load and IGF-1 were most apparent with the decrement in muscle strength. CON, control; MID, muscle-specific IGF-1 deletion.

Gene expression analyses provided insight into pathway activity and these tended to shift in the anticipated direction. Cdc48, also known as the p97/VCP ATPase complex, binds ubiquitinated proteins and E3 ligases to aid in degradation by using ATP, but had only been previously assessed in denervation or fasting-induced atrophy models (50). Confirming that Cdc48 expression is also load dependent (Fig. 6A) gives credence for the potential classification of Cdc48 as an atrogene.

The current study focused on male mice, in part because of males having a more pronounced phenotype for the metabolic consequences when compared with female mice with deletion of Igf1 induced in young adults (31). Although the sex differences warrant a deeper investigation into the underlying mechanisms associated with muscle IGF-1 loss, we did not want additional confounding factors to interfere with interpretations. Our primary outcome measures were muscle mass, force, and IGF-1 content, and the study design was powered to enable detection of differences in these measures because of loading or Igf1 deletion conditions. As a consequence, secondary measures with more biological variability, such as gene expression, were not sufficiently powered to provide robust post hoc comparisons between conditions. Even so, conclusions as to whether loading or IGF-1 were significant factors by ANOVA were possible, and set the stage for future studies to delve more deeply into these aspects. Given initial characterization of MID mice revealed differences in body composition, we were concerned that interpretation of the current study would be complicated by changes in fat and potentially other organs, particularly increased body fat associated with sustained deletion of IGF-1 (50). However, similar to previous studies (51), we observed a main effect of loading for the normalized masses both fat pads analyzed (Supplemental Table S2). Furthermore, deletion of IGF-1 did not alter normalized fat mass in any condition. A more intriguing observation was the loss of liver mass in suspension, although again unaffected by Igf1 deletion. While the reasons underlying the mass changes will require further investigation, this points to the fact that disuse atrophy certainly has effects on more than just skeletal muscle, and these broader responses should be borne in mind. The relevance of the changes to nonmuscle tissues that occur with hindlimb suspension is unclear in a clinical context. Indeed, with aging, disuse, and cachexia, there are a variety of changes that occur, including increased adiposity with inactivity and sarcopenia, as well as a loss of fat mass associated with cachexia (52).

The loss of muscle mass in this animal model is highly relevant for these clinical situations, particularly when there is combined loss of IGF-1 production with reduced activity or load. With aging, skeletal muscle mass decreases and IGF-1 levels are lowered in both the circulation and within the muscle (11–13). Patients with sarcopenia and cachexia exhibit a similar muscle wasting phenotype that can be exacerbated with disuse that is common among this population (14–16, 53). Targeting treatments that increase IGF-1 to the recovery phase is likely to provide benefit to patients as the process of regaining mass and strength is similar at both normal and diminished IGF-1, whereas disuse atrophy appears to prevent IGF-1 actions not only through interference in the signaling pathway, but also by directly targeting the ligand.

AVAILABILITY OF DATA

All data generated or analyzed during this study are included in this published article and its supplementary information files.

SUPPLEMENTAL DATA

Supplemental Tables S1 and S2 and Supplemental Figs. S1–S3: https://doi.org/10.6084/m9.figshare.13764433.v1.

GRANTS

This work was supported in part through the National Institutes of Health, Paul D. Wellstone Muscular Dystrophy Cooperative Research Center Grant U54 AR052646.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.A.S., G.V., and E.R.B. conceived and designed research; R.A.S., G.V., M.K.M., N.C.J., J.L.G., H.L., and E.R.B. performed experiments; R.A.S., G.V., M.K.M., N.C.J., H.L., and E.R.B. analyzed data; R.A.S., G.V., M.K.M., H.L., and E.R.B. interpreted results of experiments; R.A.S. and G.V. prepared figures; R.A.S. drafted manuscript; R.A.S., G.V., M.K.M., and E.R.B. edited and revised manuscript; R.A.S., G.V., M.K.M., N.C.J., H.L., and E.R.B. approved final version of manuscript.