Endocrinology and Metabolism

Overexpression of IGF-I in skeletal muscle of transgenic mice does not prevent unloading-induced atrophy

David S. Criswell, Frank W. Booth, Franco DeMayo, Robert J. Schwartz, Scott E. Gordon, Marta L. Fiorotto


This study examined the association between local insulin-like growth factor I (IGF-I) overexpression and atrophy in skeletal muscle. We hypothesized that endogenous skeletal muscle IGF-I mRNA expression would decrease with hindlimb unloading (HU) in mice, and that transgenic mice overexpressing human IGF-I (hIGF-I) specifically in skeletal muscle would exhibit less atrophy after HU. Male transgenic mice and nontransgenic mice from the parent strain (FVB) were divided into four groups (n = 10/group):1) transgenic, weight-bearing (IGF-I/WB); 2) transgenic, hindlimb unloaded (IGF-I/HU); 3) nontransgenic, weight-bearing (FVB/WB); and4) nontransgenic, hindlimb unloaded (FVB/HU). HU groups were hindlimb unloaded for 14 days. Body mass was reduced (P < 0.05) after HU in both IGF-I (−9%) and FVB mice (−13%). Contrary to our hypothesis, we found that the relative abundance of mRNA for the endogenous rodent IGF-I (rIGF-I) was unaltered by HU in the gastrocnemius (GAST) muscle of wild-type FVB mice. High-level expression of hIGF-I peptide and mRNA was confirmed in the GAST and tibialis anterior (TA) muscles of the transgenic mice. Nevertheless, masses of the GAST and TA muscles were reduced (P < 0.05) in both FVB/HU and IGF-I/HU groups compared with FVB/WB and IGF-I/WB groups, respectively, and the percent atrophy in mass of these muscles did not differ between FVB and IGF-I mice. Therefore, skeletal muscle atrophy may not be associated with a reduction of endogenous rIGF-I mRNA level in 14-day HU mice. We conclude that high local expression of hIGF-I mRNA and peptide in skeletal muscle alone cannot attenuate unloading-induced atrophy of fast-twitch muscle in mice.

  • insulin-like growth factor I

mechanical unloading of skeletal muscles normally involved in postural maintenance results in a rapid loss of muscle mass and protein content. This is due initially to a decrease in protein synthesis, followed by a transient increase in the rate of protein degradation (24). This phenomenon results in the loss of strength and mobility in humans exposed to a wide variety of situations, including spaceflight, limb casting, long-term bed rest, and/or aging. It is therefore of great importance that effective countermeasures be devised to maintain skeletal muscle mass and function during periods of mechanical unloading. In this study, we have used rodent hindlimb unloading, which results in similar changes in skeletal muscle size and phenotype as spaceflight (3, 10, 24) and is a popular Earth-based model of the physiological effects of microgravity.

All three members of the insulin-like growth factor (IGF) family of growth factors (IGF-I, IGF-II, and insulin) have been shown to induce anabolic processes in skeletal muscle (11). Much of this work has been done in cell culture (9, 11); however, recent evidence indicates that IGF-I may be important to the maintenance and growth of adult skeletal muscle mass in vivo. Linderman et al. (20) found that chronic injection of growth hormone (GH) in hypophysectomized rats together with a resistance exercise protocol attenuated the unloading-induced atrophy in rat muscle. These results may be attributable to the fact that both GH administration and functional overload have been shown independently to increase IGF-I production in the muscle (8, 11). Furthermore, transgenic mice that overexpress IGF-I in skeletal muscle exhibit a 15–20% hypertrophy of the muscle fibers (5). These findings indicate a close association between local IGF-I levels in adult skeletal muscle and growth or maintenance of muscle mass. On the basis of these and other data, Vernikos (25) hypothesized that IGF-I mRNA expression in unloaded, atrophying skeletal muscles would be reduced. We have examined this hypothesis in nontransgenic mice after 14 days of hindlimb unloading.

Because intermittent loading of skeletal muscle may not be possible or practical in all situations, the potential for complete amelioration of unloading-induced atrophy by growth factor administration should be explored. It seems logical, given the known autocrine/paracrine actions of IGF-I, that full efficacy in the prevention of atrophy may require the hormone to be elevated within the muscle tissue, rather than systemically. We have examined the efficacy of locally produced IGF-I in maintaining muscle mass during hindlimb unloading by studying a transgenic mouse line that overexpresses human IGF-I specifically in fully differentiated skeletal muscle.



Twenty male transgenic mice (6–9 mo old) harboring the human IGF-I (hIGF-I) gene driven by regulatory regions from the chicken skeletal α-actin gene [the promoter sequence extending −424 bp upstream of the transcription initiation site, the first intron, and the 3′-untranslated region (SK733 IGF-I 3′-SK)] (5) were used to assess the potential of locally produced hIGF-I to prevent unloading-induced atrophy. These mice have been shown to express hIGF-I protein and mRNA at high levels specifically in skeletal muscles, without increases in the level of circulating hIGF-I. Furthermore, this transgenic mouse line exhibits a 10–20% hypertrophy of skeletal muscle (5). Twenty male mice (∼6 mo old) from the parent strain to this transgenic line (FVB) were used as controls. We used fully grown adult mice (≥6 mo old) in both transgenic and nontransgenic groups; however, we were unable to match the groups for age and body mass. The mice were divided into four groups:1) transgenic, weight-bearing (IGF-I/WB); 2) transgenic, hindlimb unloaded (IGF-I/HU); 3) nontransgenic, weight-bearing (FVB/WB); and4) nontransgenic, hindlimb unloaded (FVB/HU). To estimate endogenous rodent IGF-I (rIGF-I) peptide concentration in FVB mice, an additional 14 mice (8 WB and 6 HU) were subjected to the identical protocol.

Hindlimb unloading.

Hindlimb unloading (HU) was performed using a modification of the protocol used for tail suspension of rats by Babij and Booth (2). Each mouse was weighed and anesthetized with an intraperitoneal injection of a cocktail containing ketamine (73.9 mg/ml), xylazine (3.7 mg/ml), and acepromazine (0.7 mg/ml) at a dose of 1.8 ml/kg body mass. Two strips of Elastoplast elastic adhesive bandage (Beiersdorf, Norwalk, CT) were cut, ∼15 cm × 0.5 cm each. The bandages were wrapped around the tail in a helical pattern starting at the base of the tail and extending 2 cm past its tip. After the mouse had recovered from the anesthetic, a swivel hook was placed through the bandage just distal to the tip of the tail. The swivel hook was then raised such that the hindlimbs were elevated just off the cage floor (this produces an ∼30° head-down tilt). Forelimbs remained in contact with the cage floor, allowing the mouse to move through a 360° circle around the tail suspension apparatus. A similar procedure has been described by McCarthy et al. (21). Mice had ad libitum access to chow and water throughout the HU protocol.

Tissue sampling.

At 14 days of HU, mice were lowered and immediately anesthetized, as described previously. Post-HU body mass was measured, and the gastrocnemius (GAST), tibialis anterior (TA), soleus, and extensor digitorum longus (EDL) muscles were removed, weighed on an analytical balance, and quick-frozen in liquid nitrogen. Muscle masses were normalized to tibia length in an attempt to account for any age and/or body weight differences between experimental groups.

RNA isolation.

Contralateral pairs of GAST, TA, and soleus muscles from each mouse were powdered under liquid nitrogen with a mortar and pestle. Total RNA was extracted from ∼100 mg of powdered muscle by use of the guanidine thiocyanate method of Chomczynski and Sacchi (4) with Trisolve (Biotecx Laboratories, Houston, TX). The extracted RNA was dissolved in diethylpyrocarbonate-treated water, and the RNA concentration was determined spectrophotometrically at a wavelength of 260 nm.

Northern blot analyses.

Northern blot analysis was used to assess the relative abundance of IGF-I and skeletal α-actin mRNAs and ribosomal 18S RNA in the GAST and TA. Extracted total RNA (15 μg) for each muscle was loaded onto a denaturing 1% agarose-formaldehyde gel [1 × 3-(N-morpholino)propanesulfonic acid and 6.7% formaldehyde] and electrophoresed at 4 V/cm for 3 h. The RNA was then transferred to a nylon membrane (Hybond-N+, Amersham, Arlington Heights, IL) by capillary action and ultraviolet cross-linked to the membrane.

Radiolabeled cDNA probes were prepared by random priming with [32P]dCTP. A 422-bp sequence of rat IGF-I cDNA, generously supplied by Dr. Charles T. Roberts (23), and a 630-bp sequence of the rat skeletal α-actin cDNA (17) were used as templates for random priming. The 18S probe was made by random priming an 80-bp highly conserved region of the human ribosomal RNA gene, obtained commercially (Ambion, Austin, TX).

The membranes containing the RNA were prehybridized with 12 ml of hybridization buffer (QuickHyb, Stratagene, La Jolla, CA) for 40 min at 68°C. One probe (IGF-I or actin) with total counts ∼1 × 106 cpm/ml and specific activity >1 × 109 cpm/μg DNA was then mixed with the hybridization buffer and incubated for 3 h at 68°C. The membrane was washed twice with 2× sodium chloride-sodium citrate (SSC) with 0.1% SDS at 22°C for 15 min each and once with 0.1× SSC-0.1% SDS at 55°C for 30 min. The membranes were then visualized by autoradiography, and the bands corresponding to IGF-I and actin mRNA were quantified by scanning densitometry (BioImage, Millipore, Ann Arbor, MI). After hybridization and autoradiography with probes for either IGF-I or actin, membranes were stripped by incubation in boiling 0.5% SDS for 5 min. Stripped membranes were then hybridized to radiolabeled probe for 18S RNA as just described. Integrated optical densities (IOD) for IGF-I and actin bands were divided by the IOD of the 18S band from the same lane to correct for variations in loading and/or transfer. A pilot experiment in which various quantities of total RNA were transferred to a membrane and hybridized to the 18S probe confirmed a linear relationship between quantity of RNA loaded and the IOD of the corresponding autoradiographic band (data not shown).

RNase protection assay.

The level of endogenous rIGF-I mRNA was measured in the GAST of the nontransgenic mice using an RNase protection assay. Radiolabeled riboprobes for rIGF-I and 18S ribosomal RNA were prepared by in vitro transcription with [32P]CTP (Amersham, Arlington Heights, IL) and in vitro transcription kits (Riboprobe Gemini System II for IGF-I and RiboMax Large Scale system for 18S; Promega, Madison, WI). RNase protection assays were performed as described earlier (14, 15). Briefly, total RNA (15 μg) from mouse muscle samples was coincubated with rIGF-I and 18S riboprobes (1–3 × 105 cpm/sample) for 16–18 h at 45°C, followed by 1 h of digestion with RNase T2 (Life Technologies, Gaithersberg, MD). Protected fragments were resolved by electrophoresis through a 6% acrylamide, 8 M urea gel. The bands were visualized by autoradiography and quantified by scanning densitometry (BioImage, Millipore, Ann Arbor, MI). The 18S probe produced two closely migrating protected bands at ∼80 nt, as previously reported for this probe (14). Therefore, the bands were scanned together to determine one 18S IOD value for each lane. The riboprobe for rIGF-I mRNA detection was obtained by subcloning the portion of the rat IGF-I cDNA described above for Northern blot probes into the pBluescript SK vector (Stratagene, LaJolla, CA). The vector was linearized with BamH I restriction enzyme (Promega), which produced a 494-nt riboprobe from in vitro transcription with T7 RNA polymerase (Promega). Part (340 nt) of this probe is complementary to the rIGF-I mRNA sequence in mouse skeletal muscle. To prepare the 18S probe, the pT7 RNA 18S template (Ambion, Austin, TX) was used. This vector yields a 109-nt transcript with T7 RNA polymerase, 80 nt of which are a conserved 18S sequence. Yeast tRNA was hybridized to the probes, digested, and separated by electrophoresis along with the muscle RNA samples to confirm specificity of the protected bands. Furthermore, incubation of the probes with increasing amounts of RNA (4–24 μg) yielded protected bands with linearly increasing IOD values (data not shown).

IGF-I peptide measurement.

The sample size requirement was such that the measurements of IGF-I peptide abundance were performed in the GAST only. The concentration of hIGF-I peptide was determined using a two-site immunoradiometric assay (IRMA) commercial kit specific for hIGF-I (Diagnostic Systems Laboratories, Webster, TX). Briefly, muscle tissue was powdered under liquid nitrogen and homogenized on ice in seven volumes of PBS (pH 7.4) containing protease inhibitors (in mM: 0.06 leupeptin, 0.01 aprotinin, 0.04 antipain, and 2.0 phenylmethylsulfonyl fluoride). Homogenates were then extracted in three volumes of acid-alcohol extraction solution for 30 min at room temperature and neutralized with four volumes of neutralization solution. The neutralized samples were frozen overnight at −80°C. The samples were then thawed and microfuged at 13,500 rpm for 10 min, and the supernatant was used for the IRMA. The IRMA was performed in duplicate on hIGF-I standards, blanks, and the extracted/neutralized samples, according to the manufacturer’s instructions. Muscle samples from nontransgenic mice were included as negative controls to verify the absence of cross-reactivity between the hIGF-I and endogenous peptides in this assay. All samples were analyzed on the same day with the same reagents. The intra-assay coefficient of variation for this assay was ∼10%.

The endogenous rIGF-I peptide concentration in muscles of FVB mice was determined using a dual-antibody radioimmunoassay specific for rodent IGF-I by use of reagents provided in kit form (Diagnostic Systems Laboratories) (18). Samples were extracted and prepared as described for the hIGF-I measurements. Additionally, porcine muscle samples were included as negative controls to verify the absence of cross-reactivity with the hIGF-I (human and porcine IGF-I have identical amino acid sequences), and rodent kidney samples were used as a positive control. All samples were assayed on the same day with the same reagents. The intra-assay coefficient of variation was 14%; the sensitivity of this assay under the conditions described permitted detection of rIGF-I peptide concentrations of ≥160 ng/g.


ANOVA with 2 × 2 factorial design was used to determine main effects and interactions for mouse lines (transgenic vs. FVB) and treatment groups (WB vs. HU) for all dependent measures except body mass. Body mass was analyzed by 2 × 2 × 2 ANOVA with repeated measure (pre- vs. posttreatment). A priori contrasts were performed to determine where individual group means differed from the FVB/WB and IGF-I/WB groups. Statistical significance was established atP < 0.05. All values are reported as means ± SE.


Body mass and tibia length.

The pre-HU body mass of the transgenic mice was ∼20% greater than that for the nontransgenic FVB mice (FVB: 33.3 ± 0.6 g; IGF: 40.4 ± 1.2 g). Tibia lengths in the transgenic mice were significantly greater than in the FVB mice; however, the mean difference was only 2.7% (mean tibia length: FVB/WB = 18.0 ± 0.05 mm; FVB/HU = 18.0 ± 0.15 mm; IGF-I/WB = 18.5 ± 0.13 mm; IGF-I/HU = 18.5 ± 0.11 mm). Body mass did not change during the 14-day treatment period in the WB groups of either mouse line. However, posttreatment body mass (29.0 ± 0.7 g) was 13.0% lower than pretreatment body mass in the FVB/HU group and 9.3% lower (36.5 ± 0.9 g) in the IGF-I/HU group.

Muscle mass.

Absolute muscle mass was greater in IGF-I/WB mice than in FVB/WB mice for the GAST (+21.5%), TA (+15.0%), soleus (+20.5%), and EDL (+11.6%). This difference remained when muscle mass was expressed per millimeter tibia length (see Table 1). Fourteen days of HU resulted in a significant loss of mass from the GAST, TA, and soleus muscles. The percentage loss for these muscle masses between WB and HU groups within mouse lines was very similar (Table 1, Fig. 1). HU GAST and TA muscle masses of IGF-I mice became similar to masses of the FVB/WB mice, but soleus muscles of IGF-I/HU mice weighed less than those in FVB/WB mice. In agreement with earlier observations in the rat (15), the mass of the EDL was unaltered in both mouse strains after HU. This suggests that the atrophy observed in the GAST and TA muscles results from unloading these muscles, rather than from a general muscle wasting accompanying the HU-induced loss of body mass.

View this table:
Table 1.

Muscle mass data

Fig. 1.

Percent atrophy of gastrocnemius (GAST) and tibialis anterior (TA) muscles after 14 days of hindlimb unloading, relative to average mass of corresponding muscles in weight-bearing controls. FVB, nontransgenic mice from the FVB strain. IGF-I transgenic, transgenic mice that overexpress human insulin-like growth factor I in skeletal muscle. * Significantly different (P < 0.05) from weight-bearing control.

mRNA measurements.

The very low level of expression of endogenous rIGF-I mRNA precluded its detection by Northern blot analysis in the skeletal muscle of nontransgenic FVB mice. However, the RNase protection assay, which is ∼100 times more sensitive than Northern blot analysis, allowed assessment of rIGF-I mRNA expression in the GAST of the FVB mice. No difference in rIGF-I mRNA was observed between WB and HU groups of FVB mice (Fig. 2).

Fig. 2.

Data from RNase protection assay for insulin-like growth factor I (IGF-I) mRNA in gastrocnemius of weight-bearing (WB) and hindlimb-unloaded (HU) nontransgenic FVB mice.A: autoradiograph showing detection of a 340-nt protected fragment for IGF-I mRNA and protected fragments for ribosomal 18S RNA (2 closely migrating bands at ∼80 nt) hybridized simultaneously to 15 μg of total RNA (seemethods).B: quantification of IGF-I mRNA by densitometry, normalized to density of combined 18S bands. Values are means ± SE; n = 5/group.

Overexpression of hIGF-I mRNA was clearly evident by Northern blot analysis in the GAST and TA of the transgenic mice (Fig.3). However, expression of the transgene could not be detected in the soleus muscle of transgenic mice analyzed on the same Northern blots as GAST and TA samples (data not shown). Muscle hIGF-I mRNA abundance was significantly decreased after HU in the GAST (−56%) and TA (−60%) of transgenic mice. Nevertheless, hIGF-I mRNA expression was still clearly evident by Northern blot in the GAST and TA of transgenic HU mice (Fig. 3). Skeletal α-actin mRNA level did not differ between transgenic and FVB mice or between WB and HU groups (Fig. 4).

Fig. 3.

Insulin-like growth factor I (IGF-I) mRNA detected by Northern blot of 15 μg of total RNA from GAST and TA of WB and HU mice.A: autoradiograph showing expression of IGF-I mRNA in muscles of IGF-I transgenic mice and illustrating downregulation of this transcript after HU.B: quantification of IGF-I mRNA by densitometry in skeletal muscle of IGF-I transgenic mice, normalized to density of 18S band. Values are means ± SE;n = 5/group. * Significantly different (P < 0.05) from weight-bearing control.

Fig. 4.

Skeletal α-actin mRNA detected by Northern blot of 15 μg total RNA from GAST and TA of WB and HU mice. A: autoradiograph showing expression of actin mRNA in these muscles from FVB and IGF-I transgenic mice. B: quantification of actin mRNA by densitometry, normalized to density of 18S band. Values are means ± SE; n = 5/group.

IGF-I peptide measurements.

The transgenically expressed hIGF-I peptide was clearly detectable at high levels in the GAST muscles of the transgenic mice. In a manner similar to muscle hIGF-I mRNA, the hIGF-I peptide concentration decreased from 1,818 ± 163 ng/g of muscle for the WB group to 882 ± 305 ng/g for the HU group. Although the transgenic animals had higher muscle masses and hIGF-I peptide levels (870–2,900 ng/g) than the FVB wild-type animals, we found no correlation between hIGF-I peptide concentration and GAST mass within the transgenic WB group.

Endogenous rIGF-I peptide was above the minimum detectable limit in approximately one-half of the muscle samples from the FVB mice and precluded the ability to assess the effect of hindlimb suspension on this measurement in nontransgenic animals. There was no difference between the nontransgenic WB and HU groups for either the number of animals in which peptide levels were above the detectable limit (3/8 for WB and 3/6 for HU) or in the average value for those mice in which peptide levels were measurable (317 ± 124 ng/g for WB and 264 ± 86 ng/g for HU). Therefore, although the GAST hIGF-I peptide concentration was significantly lower in the HU group compared with the WB group for the transgenic animals, these levels still exceeded the endogenous rIGF-I level of the FVB mice by ≥3.5- to 7-fold.


In view of the abundant literature that demonstrates an association between muscle hypertrophy and an upregulation of IGF-I expression (5,8, 11), as well as the recent suggestion that reduced IGF-I expression may be responsible for the atrophy of human muscle during spaceflight (25), we hypothesized that the presence of high levels of locally expressed IGF-I could prevent the muscle atrophy that results from mechanical unloading. We accomplished our objective by studying the response to unloading of transgenic mice in which high levels of IGF-I expression are sustained in skeletal muscle through the use of a hybrid gene construct in which IGF-I expression is driven by the skeletal α-actin promoter (5). Our use of this model was supported by recent studies in which the administration of GH was more effective in mitigating the loss of muscle mass in the hindlimb of tail-suspended hypophysectomized rats when combined with a resistance exercise program (12, 20). These results may be attributable to the fact that both GH administration and functional overload have been shown to increase IGF-I production in the muscle (8, 11). Moreover, Coleman et al. (5) found that myoblasts transfected with an IGF-I- overexpressing vector produced higher levels of muscle-specific mRNAs than did control myoblasts treated with exogenous IGF-I added to the culture medium. This suggests that the autocrine/paracrine action of endogenous IGF-I is necessary to obtain the maximum myogenic effects.

In the present study, we observed that the local overexpression of IGF-I was associated with a higher absolute muscle mass in the IGF-I/WB mice (GAST: +21%; TA: +15%) compared with the FVB nontransgenic mice, even though tibial lengths differed by only 2.7%. After 14 days of suspension, however, the percent loss of mass in the GAST and TA of the transgenic mice (∼20%) did not differ from the HU-induced atrophy in the FVB wild-type mice (Fig. 1), resulting in an absolute mass of these two muscles in the HU transgenic mice equivalent to that of the FVB/WB mice. Therefore, our hypothesis that locally produced IGF-I would be more effective in preventing unloading-induced atrophy of skeletal muscle was not supported.

Although our findings indicate that locally overexpressed IGF-I does not have a protective effect on unloading-induced muscle atrophy, this conclusion presupposes that the observed reduction in hIGF-I expression in the IGF-I/HU group was not biologically significant. This supposition is indeed supported by the absence of a correlation between GAST mass and its hIGF-I concentration in the IGF-I/WB group over a range of hIGF-I values between 870 and 2,900 ng/g. The lack of such a relationship could be interpreted as indicating the presence of a threshold hIGF-I value above which there is no further enhancement of IGF-I-induced muscle growth. Because the mean hIGF-I concentration of the IGF-I/HU animals (882 ng/g) falls within this range, it is unlikely that this unloading-induced decrease in IGF-I expression would be sufficientto lessen any potential protective effects. Regardless, the level of IGF-I concentration in the GAST of the HU transgenic animals still far exceeded that of nontransgenic mice.

The inability of locally overexpressed hIGF-I to attenuate the unloading-induced skeletal muscle atrophy in our transgenic animals may be due to other uncharacterized alterations in the IGF-I autocrine/paracrine system resulting from high tissue IGF-I expression. For instance, a decrease in receptor density could occur with IGF-I overexpression. However, receptor number generally adapts to a new homeostatic IGF-I level by increasing or decreasing in an inverse fashion (19). Thus the reduction in IGF-I overexpression with HU could likewise be accommodated by an increase in receptor number, thereby mitigating the decrease in local IGF-I expression. Changes in local binding protein production are another possibility. Although IGF-binding protein-2, -4, -5, and -6 are all potentially expressed in skeletal muscle, their exact function is currently unclear (11). More research is warranted concerning their role in modulating IGF-I activity in WB and HU animals when IGF-I is overexpressed in skeletal muscle. Finally, postreceptor events or the interactive effects of other hormones may also be involved with the lack of local IGF-I overexpression in protecting against the skeletal muscle atrophy.

Although hIGF-I expression in the GAST of the transgenic animals decreased with HU, a parallel response was not evident for endogenous rIGF-I mRNA in the GAST of the FVB animals. We were surprised not to observe a decrease in endogenous IGF-I mRNA in atrophying skeletal muscle, because of the abundant literature demonstrating an association between muscle hypertrophy and an upregulation of IGF-I expression (5,8, 11, 25). However, in a follow-up investigation in a second species, the rat, we have observed the same phenomenon in both the GAST and the soleus muscles after 2 wk of HU, despite significant atrophy of these muscles (−29% and −38%, respectively). Northern blot analysis for IGF-I mRNA in these rats produced mean IOD values of 0.46 ± 0.05 (WB) vs. 0.50 ± 0.12 (HU) in the GAST and 1.58 ± 0.63 (WB) vs. 1.46 ± 0.07 (HU) in the soleus. The absence of a downregulation in IGF-I mRNA expression with unloading-induced skeletal muscle atrophy complements our previous report indicating that skeletal muscle IGF-I mRNA expression is unaltered in the age-related atrophy of fast-twitch muscles of rats (14).

The study of this transgenic mouse line additionally revealed unique regulatory features of the skeletal α-actin/hIGF-I transgene. In these transgenic mice, expression of the hIGF-I gene is driven by regulatory regions of the chicken skeletal α-actin gene (i.e., −424 promoter, first intron, and the 3′-untranslated region). These regulatory elements are very effective in conferring high-level, skeletal muscle-specific expression to the transgene (5). However, we found that the transgene is not expressed entirely as expected on the basis of regulation of the endogenous skeletal α-actin gene. Overexpression of hIGF-I mRNA was clearly evident by Northern blot analysis in the GAST and TA of the transgenic mice (Fig.3). However, hIGF-I mRNA expression was not detected in the soleus muscle of these mice. The soleus of the mouse is composed of ∼60% type I and ∼40% type IIa fibers (7), whereas the GAST and TA are mixed fiber type muscles containing primarily fiber types IIa, IIx, and IIb (22). This suggests that transgene expression, unlike endogenous skeletal α-actin, is likely restricted to the type IIx and/or IIb fibers. For this reason, the effect of hIGF-I overexpression on HU-induced atrophy in the slow soleus muscle cannot be ascertained by this study.

In addition to the fiber type-specific expression of the transgene, we also found that 14 days of HU caused a significant reduction in the hIGF-I mRNA per microgram of total RNA in the GAST and TA of the transgenic mice, despite the observation that mRNA concentrations for the endogenous rIGF-I and skeletal α-actin mRNAs were unaltered by the HU protocol. This reduction in hIGF-I transgene mRNA abundance (−56%) was mirrored by a significant decrease in hIGF-I peptide concentration (−52%) in the IGF-I/HU GAST, implying that something unique about the actin/hIGF-I chimeric transgene causes its expression in skeletal muscle to be sensitive to mechanical loading. Along with the fiber type-specific expression discussed above, this finding may be explained by some unknown regulatory element(s) that confers constitutive activity and is present in the endogenous skeletal α-actin gene but not the transgene. Conversely, these results could be due to the contextual location of the transgene in the genome of the transgenic mice. However, other mouse lines harboring this transgene also exhibit fiber type-specific expression (data not shown). Presently, we can only speculate as to the cause of the unique regulation of this transgene. Nevertheless, a minimum of an ∼3.5- to 7-fold increase in IGF-I peptide concentration did not inhibit atrophy of fast-twitch muscle during HU.

The unaltered skeletal α-actin mRNA per microgram of total RNA in the mouse GAST and TA after HU was unexpected. Previous observations in the hindlimb-suspended rat indicated a decreased skeletal actin mRNA per microgram of total RNA (2). This is not the first account of varying skeletal muscle responses to unloading between the mouse and rat. For example, Haida et al. (13) reported that the proportion of type II fibers decreased in the soleus of mice after 14 days of HU. This is in direct opposition to reports of increased type II fiber percentages after unloading of the soleus muscle in the rat (6). Therefore, the differences between species should be carefully considered when skeletal muscle data from mice and rats are interpreted and compared.

In conclusion, we found that the skeletal muscle atrophy caused by 14 days of HU was not attenuated by the local overexpression of hIGF-I in mice. Mechanical unloading caused a similar percent loss of mass in both the GAST and TA muscles, regardless of the level of hIGF-I expression in these muscles. We also found that endogenous rIGF-I mRNA level was unaltered in atrophying GAST muscle of nontransgenic mice, suggesting that a downregulation of IGF-I may not be involved in unloading-induced skeletal muscle atrophy, as we had hypothesized. Conversely, we report that the expression of the skeletal α-actin/hIGF-I transgene is downregulated after 14 days of HU and that this expression also occurs in a fiber type-specific manner. However, despite this downregulation, the GAST and TA of the transgenic mice still expressed hIGF-I mRNA and peptide levels that were magnitudes higher than in the corresponding muscles in nontransgenic FVB mice. The hypothesis that locally produced IGF-I would be more effective than bolus injections of growth factors in maintaining muscle mass during unloading was not supported in this study. However, because of the lack of overexpression of hIGF-I mRNA in the soleus of our transgenic mice, future experiments will be necessary to extend this finding to slow-twitch skeletal muscles. Therefore, we conclude that elevated IGF-I expression alone is ineffective in preventing unloading-induced muscle atrophy within fast-twitch skeletal muscle.


Technical assistance was provided by Peter Zhang and Ramey Benfield.


  • Address for reprint requests: F. W. Booth, Dept. of Integrative Biology, Univ. of Texas Medical School, 6431 Fannin St., Houston, TX 77030.

  • This project has been funded by National Aeronautics and Space Administration (NASA) Grant NAGW-3908 and with federal funds from the US Department of Agriculture, Agricultural Research Service, under cooperative agreement number 58-7MN1-6-100, and the National Space Biomedical Research Institute. The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. Dr. Scott Gordon is the recipient of a NASA Postdoctoral Research Associate Award in Space Biology.


View Abstract