The mechanical regulation of the forkhead box O (FOXO) subclass of transcription factors in the respiratory pump and its implication in aging are completely unknown. We investigated the effects of diaphragm stretch on three FOXO isoforms, Foxo1, Foxo3a, and Foxo4, in normal mice at different ages. We tested the hypotheses that 1) FOXO activities are regulated in response to diaphragm stretch and 2) mechanical properties of aging diaphragm are altered, leading to altered regulation of FOXO with aging. Our results showed that stretch downregulated FOXO DNA-binding activity by a mechanism that required Akt and IKK activation in young mice but that these pathways lost their mechanosensitivity with age. This aberrant regulation of FOXO with aging was associated with altered viscoelasticity, compliance, and extensibility of the aged diaphragm. Curiously, the dramatic decrease of the nuclear content of Foxo1 and Foxo3a, the two isoforms associated with muscle atrophy, with aging correlated with higher basal activation of Akt and IKK signaling in diaphragms of old mice. In contrast, the stability of Foxo4 in the nucleus became dependent on JNK, which is strongly activated in aged diaphragm. This finding suggests that Foxo4 was responsible for the FOXO-dependent transcriptional activity in aging diaphragm. Our data support the hypothesis that aging alters the mechanical properties of the respiratory pump, leading to altered mechanical regulation of the stretch-induced signaling pathways controlling FOXO activities. Our study supports a mechanosensitive signaling mechanism that is responsible for regulation of the FOXO transcription factors by aging.
- respiratory muscle
- mechanical stretch
mechanical stimuli are known to affect several mechanosensitive signaling pathways in different types of cells. Particularly in muscle, modifications in the magnitude and duration of mechanical stimuli provoke changes in gene expression that affect metabolism and produce adaptations in muscle mass and muscle function (50). The mechanisms by which muscle cells are able to translate physical mechanical force into chemical signals are just beginning to be explored. The respiratory pump, i.e., the diaphragm, is one of the few skeletal muscles that is constantly subjected to mechanical loading. Therefore, the diaphragm may be among the skeletal muscles most affected by mechanical forces.
The phophatidylinositol-3-OH kinase (PI3K)-Akt signaling pathway is one of the most recognized mechanosensitive signaling pathways in skeletal muscles. Several reports have shown rather convincingly the activation of Akt by increased skeletal muscle contraction or passive muscle stretch (23, 40, 42, 43, 51). Akt signaling contributes to muscle mass growth and maintenance by two independent mechanisms: one involves stimulation of protein synthesis, and the other includes inhibition of protein degradation. The first mechanism occurs by activation of the rapamycin-mammalian target of rapamycin pathway, which stimulates protein synthesis by activation of the p70/S6 kinase and phosphorylation of translation factors (36, 38, 41). Recently, the age-related loss of induced mass growth in fast-twitch muscles has been correlated with an increase in AMP-activated protein kinase signaling, which downregulates the rapamycin-mammalian target of rapamycin pathway and, consequently, protein synthesis (48). The second mechanism acts through downregulation of the family of forkhead homeobox type O (FOXO) transcription factors, which are responsible for expression of the muscle atrophy-induced ubiquitin ligases muscle ring finger 1 (MuRF1) and muscle atrophy F box (MAFbx), also called atrogin (44, 46).
Regulation of FOXO factor activity is mainly controlled by a shuttling system that modulates its cellular localization by phosphorylation sites located in the COOH-terminal domain. Phosphorylation of these sites by Akt provokes their nuclear export (52). Four members of the FOXO subfamily of forkhead transcription factors have been identified in the mouse. Three of these, Foxo1, Foxo3a, and Foxo4, have the same tissue distribution pattern as their human counterparts: Foxo1 is highly expressed in ovaries, Foxo4 is highly expressed in muscle, and Foxo3a is widely expressed in a variety of tissues. The fourth member found in mice, Foxo6, appeared restricted to the brain (4, 46) and is regulated in a different manner by Akt, which controls its activity but does not affect its nuclear localization (53). Foxo1 has been reported to contribute to myotube fusion during myoblast differentiation by an Akt-independent mechanism (6, 24), and its overexpression in mice has a detrimental effect on muscle mass and upregulates type I slow-twitch fibers (27). In C2C12 myotubes, the expression of constitutively active Foxo3a is enough to induce atrophy, with a concomitant activation of MAFbx expression (44).
Changes in RNA levels of Foxo1, Foxo3a, and Foxo4 have been reported in aged rats that are affected by calorie restriction (18). However, the regulation of the Akt signaling pathway and its effect on FOXO transcriptional activity in aging skeletal muscles have not been explored, despite a clear relevance of FOXO transcription factors to skeletal muscle formation. We postulated that FOXO transcriptional activity is sensitive to mechanical stretch and that stretch-induced regulation of FOXO in the aging respiratory pump is distinct from that in the young respiratory pump. Our data support the novel hypothesis that aging of the respiratory pump is associated with aberrant regulation of the mechanical stretch-induced signaling pathways that control FOXO activity.
MATERIALS AND METHODS
[32P]dCTP and the Ready-to-Use dCTP labeling kit were obtained from Amersham; FKHRL1 (Foxo3a) antibody from Upstate Biotechnology; Akt inhibitor II, IKK-2 inhibitor IV (SC-514), and JNK inhibitor II (SP-60025) from Calbiochem; aprotinin, benzamidine, leupeptin, wortmannin, and anti-tubulin antibody from Sigma-Aldrich; Akt and phosphospecific (Ser473) Akt, phosphospecific (Ser176/180) IKKα and IKKβ, and IKKα and IKKβ antibodies from Cell Signaling; JNK1, JNK2, phosphospecific (Thr183 and Tyr185) JNK1/2, FKHR (Foxo1), and AFX (Foxo4) antibodies and secondary antibodies from Santa Cruz Biotechnology; and oligonucleotides from Invitrogen. The sequences of the oligonucleotides used in EMSAs were as follows: CTATAAGTAAACAACTGTGACTAGT for total FOXO activity, GGGATAAATACTGTGCTCGGGCAG for Foxo1, and CTAGCAAGCAAACAAACTTATTTTGAACACGGG for Foxo3a. A mutated version of the Foxo1 oligonucleotide (GGGATCACTACTGTGCTCGGGCAG) was used as cold noncompetitor.
Mice and tissue preparation.
Pathogen- and tumor-free mice from the C57BL6 strain (National Institute on Aging) were maintained in the animal facility at Baylor College of Medicine. The animals were housed and fed in stainless steel cages under a 12:12-h light-dark schedule. All animal studies were approved by the Baylor College of Medicine Institutional Animal Care and Use Committee and conformed to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Three groups of mice were used for mechanics and biochemical experiments: the first group (n = 42) consisted of young (2-mo-old) mice (63.33 ± 20.14 days old, 22.47 ± 3.35 g body wt), the second group (n = 17) was ∼1 yr old (382.71 ± 21.28 days old, 30.53 ± 3.22 g body wt), and the third group (n = 22) was ∼2 yr old (692.09 ± 79.63 days old, 29.87 ± 3.55 g body wt). The mice were deeply anesthetized with an injection (0.5–0.7 ml/kg ip) of rodent III combination anesthetic [ketamine (37.6 mg/ml), xylazine (1.92 mg/ml), and acepromazine (0.38 mg/ml)]. The entire diaphragm muscle and its rib attachments were quickly excised and immersed in a muscle bath containing a modified Krebs-Ringer solution [in mM: 137 NaCl, 5 KCl, 1 NaH2PO4, 24 NaHCO3, 2 CaCl2, and 1 MgSO4 (pH 7.4)] bubbled with 95% O2-5% CO2. The solution was maintained at 25°C throughout the preparation and experimental phases of the study.
Muscle mechanical testing.
After excision, the hemidiaphragms were quickly attached to force carriages by alligator clamps in an in vitro biaxial muscle-testing apparatus, as previously described (33). This apparatus has two perpendicular axes, each driven by a micrometer. Two small identical alligator clamps (0.9 cm for the y-axis and 1 mm for the x-axis) were used to stretch the muscle. To secure the diaphragm muscle sheet, one clamp was fixed on the central tendon and the opposing clamp on the myotendinous junction at the chest wall insertion, with the rib cage intact. A force transducer attached to the force carriage (model LQB, Cooper Instruments; 150 ± 50 g, differential bridge type) was used to measure applied mechanical loading. Forces were then amplified, conditioned with a low-pass-frequency filter (CD19A carrier demodulator, Validyne, Northridge, CA), and recorded with a data acquisition board (PCI-6229 DAQm series, National Instruments).
Passive length-tension relationships.
The length-tension relationship was obtained by placement of four silk suture markers (7-0 or 8-0 Surgilene) on the abdominal membrane surface of the left costal hemidiaphragm before the muscle was clamped in the biaxial apparatus. The muscle was mounted so that the position of markers during passive stretch could be viewed with a black-and-white CCTV-type camera and recorded on a VHS tape. All four markers were placed in the central region of the muscle to minimize the boundary effects. The markers were placed in a ∼1-mm2 configuration on adjacent myofibers. Lengthening and shortening stretch cycles were initiated from an unstressed state to assess length-tension relationships. Left costal diaphragms were stretched to peak tensions of 1.11, 2.22, and 4.44 g/cm passive tension in directions along and transverse to muscle fibers. Taped images of the markers during stretch cycles were converted digitally with a video capture card. A custom-made marker-tracking program was used to select marker positions on an x-y Cartesian plane, with the assumption that all markers lie within the same plane. Peak strains were than aligned, and a custom-made Matlab script was used to match peak strains with recorded force data.
Computation of two-dimensional strains.
Mechanical strains were calculated on the basis of methods established in our previous work. The strains in the plane of the muscle were computed by the following procedure. The marked region is divided into triangles, with markers forming the apices. The coordinates of these points in that unstressed plane of the diaphragm are denoted xi and yi (i = 1, 2). The displacement, ui, from the unstressed state to the deformed state is assumed to be a linear function of position and is computed as follows (1) which, with known values of the marker displacements and the position of three markers substituted for ui, xi, and yi, provides a set of three equations for the coefficients a1, a2, and a3. The marker data provided the information required to determine the coefficients a4, a5, and a6 in the following equation (2) The values of the coefficients (e.g., a1, a2, a3) were used to find the partial derivatives, which were substituted into the following equations (3) (4) where ∂u and ∂v denote the marker's displacement in the directions of the muscle fibers (x), as well as the direction transverse to the fibers (y). These displacements were used to compute mechanical strains. The strains, εx and εy, were computed for each triangle. εxy is defined as shear strain and is essentially negligible. Therefore, εx is aligned with the muscle fiber direction, whereas εy is in aligned in the transverse direction to the fibers (5) where λ is defined as the extension ratio and ε as mechanical strain.
Measurements and modeling of viscoelastic properties.
We measured stress-relaxation curves for the diaphragm muscles from the same mice, as previously described (33). Diaphragms were stretched to peak tensions of 1.11 g/cm. The muscle was then kept at a constant length and allowed to relax asymptotically until it reached a plateau at 300 s. Acquired force data were filtered utilizing the Matlab smooth tool and normalized to peak force. A Matlab nonlinear least-squares routine was used to fit the normalized data to the standard linear solid model of viscoelasticity (7). This simple model of viscoelasticity describes the muscle as a parallel combination of a dashpot with coefficient of viscosity η1 and a linear spring of spring constant μ1 with a second linear spring of spring constant μ0 that is in parallel with the first spring and the dashpot. The relaxation function based on this model has the form (6) where F is the relaxation force, t is time, ER is the relaxed elastic modulus, τε is relaxation time for constant strain, and τσ is the relaxation time for constant stress. After the data were fit to the standard viscoelastic model, ER was obtained and averaged across each age group. ER can be interpreted graphically as the steady-state fraction of peak force achieved after 300 s.
Ex vivo mechanical stretch protocol for biochemical experiments.
This experimental protocol was performed essentially as described in our previous work (31). The left hemidiaphragms were quickly excised and placed into the biaxial apparatus with oxygenated Krebs-Ringer solution. A single stretch to 2.22 g/cm was applied to the left hemidiaphragm and held under constant mechanical stretch for 0, 15, 30, or 60 min, depending on the experiment. The right hemidiaphragm from the same mouse was used as a nonstretched control and remained in the same bath for the same amount of time as the stretched left hemidiaphragm.
Pharmacological inhibitors were added to Krebs-Ringer solution, and diaphragms were allowed to incubate for 30 min before mechanical stretching. The inhibitors were as follows: Akt inhibitor II (20 μM; Calbiochem); SC-514, a specific inhibitor for IKKβ (2 μM) (28); SP-60025, a specific inhibitor for JNK1 and JNK2 (10 μM) (29); and wortmannin (1 μM) (37).
Nuclear and whole cell extract preparation.
Nuclear extracts were prepared at the end of each stretch protocol as follows. Muscles were immediately suspended in 300–400 μl (15 μl/g body wt) of low-salt lysis buffer [10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 2.0 μg/ml leupeptin, 2.0 μg/ml aprotinin, and 0.5 μg/ml benzamidine] and subjected to mechanical grinding. The disrupted tissue was allowed to swell on ice for 5 min, subjected to two cycles of freeze-thaw lysis, and vortexed vigorously for 10 s. The lysates were centrifuged at 14,000 rpm for 30 s at 4°C, and the supernatants (cytoplasmic extracts) were kept at −70°C. Nuclear pellets were washed in low-salt buffer, resuspended in ∼100 μl (5 μl/g body wt) of high-salt buffer [20 mM HEPES (pH 7.9), 420 mM NaCl, 1 mM EDTA, 1 mM EGTA, 25% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 2.0 μg/ml leupeptin, 2.0 μg/ml aprotinin, and 0.5 μg/ml benzamidine], and incubated on ice for 30 min with intermittent vortexing. The nuclear lysates were then centrifuged at 14,000 rpm for 5 min at 4°C, and the supernatant (nuclear extracts) was assayed immediately or kept at −70°C for further analysis. Whole cell extracts were prepared as follows. After the stretch protocol, muscles were washed in PBS containing 0.5 mM phenylmethylsulfonyl fluoride and homogenized by mechanical grinding in 20 mM HEPES (pH 7.5), 2 mM EDTA, 250 mM NaCl, 0.3% Nonidet P-40, 1× protease inhibitor cocktail (Pierce), and 1× phosphatase inhibitor cocktail (Pierce). Samples were kept on ice for 15 min and centrifuged at 14,000 rpm for 2 min. Appropriate aliquots were used to determine the protein concentration of the samples by measurement of its absorbance at 595 nm in the presence of Bio-Rad protein assay reagent, and BSA was used as protein standard.
Hemidiaphragms were stretched in the direction of the muscle fibers for 15 min or as indicated, and nuclear extracts were prepared as previously described. 32P-labeled FOXO or Foxo1- and Foxo3a-specific consensus oligonucleotides were incubated with 10–15 μg of nuclear extract for 20 min at room temperature in binding buffer [25 mM HEPES (pH 7.9), 0.5 mM EDTA, 0.5 mM dithiothreitol, 1% Nonidet P-40, 5% glycerol, and 50 mM NaCl]. DNA-protein complexes were separated in a 5% native polyacrylamide gel that had been prerun for 30 min. Radioactive bands were visualized and quantitated after overnight exposure in a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) using Image Quant software. The specificity of the DNA binding was assessed by completion of the reaction with a fivefold excess of the corresponding specific oligonucleotide and the same excess of a mutated version lacking the sequence recognized by FOXO. DNA-binding activity was calculated by establishing that 1 arbitrary unit was equal to the measurement obtained for the specific 32P-labeled band corresponding to the young nonstretched diaphragm.
Protein samples were subjected to 10% SDS-PAGE and transferred to nitrocellulose membranes with a semidry transfer cell (Bio-Rad) for 30 min at 10 V. Membranes were stained with Ponceau S (Sigma) to assess equal loading, blocked in 5% nonfat dry milk (blotting-grade blocker, Bio-Rad) in 1% Tween in PBS, and exposed to proper dilutions of primary antibody in 5% milk-1% Tween in PBS for 3 h at room temperature or overnight at 4°C. Immunoreactive bands were detected with horseradish peroxidase-conjugated secondary antibodies and a bioluminescent assay for peroxidase (ECL, Amersham).
All experiments were repeated three times unless otherwise indicated. Values are means ± SD. For statistical analysis, Student's t-test or ANOVA was used to compare quantitative data populations with normal distribution and equal variance. P < 0.05 was considered statistically significant. Linear regression analysis was used for JNK activation with age.
Diaphragm muscle extensibility and compliance decrease with age.
Older diaphragms exhibited a decreased extensibility as demonstrated by a leftward shift in the passive length-tension curves of diaphragms from 2- and 1-yr-old mice relative to the 2-mo-old animals. Figure 1A shows representative length-tension relationships along the muscle fibers of the costal diaphragm from 2-mo-old, 1-yr-old, and 2-yr-old mice. Muscle length is expressed as extension ratio, i.e., the ratio of fractional increase in length to length at which no passive force was applied. The representative data demonstrate a leftward shift in length-tension relationships as age increases. All length-tension curves demonstrated strong nonlinearity. Mean extension ratios were also computed from three stretch cycles for each age group at 0.56, 1.11, 2.22, and 4.44 g/cm passive tension (Fig. 1B). These data indicate that mean extension ratios are statistically smaller for diaphragms from 2-yr-old (n = 4) and 1-yr-old (n = 6) than 2-mo-old (n = 4) mice. The difference between diaphragms from old and young animals increased with increasing tension: at 0.56 g/cm, the difference in mean extension ratios between diaphragms from the 2-yr-old and 2-mo-old mice was 0.407 (t-value = 6.37, P < 0.05); at 4.44 g/cm, the difference increased to 0.63 (t-value = 6.146, P < 0.05). Next, we used the mean extension ratios in Fig. 1B to calculate stiffness and compliance ratios. The change in mean extension ratio from 2.22 to 4.44 g/cm is greatest in diaphragms from 2-yr-old diaphragms and smallest in diaphragms from 2-mo-old animals. This region can be approximated by linear slopes, which then define stiffness. The stiffness (g·cm−1·extension ratio−1) for diaphragms from the 2- and 1-yr-old mice was 6.21 and 4.06, respectively. The inverse ratio of stiffness of the 2-yr-old diaphragm to that of the 1-yr-old diaphragm defines the compliance ratio relative to the 1-yr-old diaphragm and was 0.65, indicating a much less compliant muscle at 2 than at 1 yr of age. Similarly, the compliance ratio of the 2-yr-old to the 2-mo-old diaphragm was 0.45. Over the range of applied mechanical strains, the hysteresis, or difference between passive shortening curve and passive lengthening curve, at the same tension did not appear to be sensitive to aging. Taken together, these data show decreased muscle extensibility and decreased compliance of the aging mouse diaphragm.
Aging contributes to altered muscle viscoelasticity.
With regard to viscoelastic behavior, the force needed to maintain a particular constant stretch once that stretch has been produced decreases over time. Using the standard linear viscoelastic model, we started by extending the series spring. As time passed, the springs-and-dashpots system relaxes, manifesting as a force decay with time. This reduction of mechanical stress at constant applied mechanical strain is known as stress relaxation, and the ratio of stress to strain is known as the ER, which decreases over time. Using the standard linear solid model, we determined ER for diaphragm muscles from young and old mice. In this analogy, the spring, i.e., the diaphragm, was stretched to 1.11 g/cm passive tension and allowed to relax for 5 min under constant length. The ensuing force decay profile was normalized for each experiment to the peak tension and fit to the standard linear solid model according to Eq. 6. Fitted stress-relaxation data are shown in Fig. 2A for diaphragms from 2-mo-old, 1-yr-old, and 2-yr-old mice. The raw data are fitted to the stress-relaxation model and plotted as fraction of peak stress vs. time. The diaphragm from the 2-mo-old animal relaxed to ∼52% of its initial peak compared with the diaphragm from the 2-yr-old animal, which relaxed to ∼63% of its peak tension. The mean steady-state ER is shown in Fig. 2B. Although ER of the diaphragms from 2-mo-old mice was smaller than that from 1- or 2-yr-old animals, the differences were not significant: 0.52705 ± 0.0649 for 2-mo-old (n = 2), 0.620 ± 0.034 for 1-yr-old (n = 3), and 0.639 ± 0.0175 for 2-yr-old (n = 4) mice (P = 0.19). These data are consistent with data in Fig. 1A showing that muscle passive stiffness is higher in 2-yr-old than 2-mo-old mice.
Effect of mechanical stretch on the DNA-binding activity of FOXO in the mouse diaphragm.
Hemidiaphragm muscle sheets from 2-mo-old mice were excised and subjected to constant mechanical stretch for 5 min to 1 h. The DNA-binding activity of FOXO was detected by EMSA with a probe that binds to the three isoforms in nuclear extracts from stretched and nonstretched samples. The samples from stretched hemidiaphragms showed lower FOXO-specific DNA-binding activity (Fig. 3A). The ratio of stretch-stimulated to basal activity was calculated for each time (Fig. 3B); a maximal decrease of 60% was obtained at 15 min of stretch, and lower but significant decreases were observed at 30 and 60 min. This result indicates that the DNA-binding activity of FOXO is downregulated by mechanical stretch in a time-dependent manner.
PI3K/Akt and IKK are responsible for the downregulation of FOXO transcriptional activity in response to mechanical stimuli in the young diaphragm. FOXO transcription factors can be regulated by the protein kinases Akt and IKK, which respond to mechanical stimuli in the diaphragm (23, 31). It has been shown that Akt phosphorylates Foxo1 (3, 22), Foxo3a (8), and Foxo4 (30, 47). Akt phosphorylation localizes them to the cytoplasm and elicits their degradation. This mechanism depends on the activation of Akt by PI3K (1). IKKβ can also modulate Foxo3a by impeding its nuclear localization and triggering its proteolysis (25). To evaluate the involvement of Akt and IKKβ in the inhibition of the FOXO DNA-binding activity caused by mechanical stretch of the diaphragm, the diaphragms were preincubated for 30 min and during the stretch protocol with wortmannin, an inhibitor of PI3K; a benzimidazole compound that inactivates Akt but not its upstream regulator PI3K or SC-514, a specific inhibitor for IKKβ were added to the bath at final concentrations of 1, 20, and 2 μM, respectively. DMSO was added to the bath of control specimens during the same preincubation period. Nuclear proteins of the treated stretched and nonstretched hemidiaphragms were evaluated for FOXO DNA-binding activity by EMSA, and the stretched-to-control activity ratios were determined (Fig. 4, A and B). Wortmannin and SC-514 suppressed the inhibitory effect of mechanical stretch on the total FOXO-specific DNA-binding activity of the diaphragm, whereas Akt inhibition provoked a slight nonsignificant rise in the activity of the stretched compared with the control samples. These results indicate that Akt and IKK are involved in the mechanical stretch-induced downregulation of FOXO. We were unable to detect the phosphorylated species of FOXO; using antibodies specific for an Akt-phosphorylated form of Foxo3a, we detected a ∼40-kDa band in cytoplasmic fractions that peaks at 30 min of stretch (not shown), which suggests that the phosphorylated forms of FOXO are promptly degraded. The effect of 15 min of stretch on Akt and IKK activation was assessed by determination of the content of its phosphorylated forms. The content of phosphorylated Akt and IKKs changes dramatically in response to mechanical stretch, but the content of total Akt and IKKs does not (Fig. 4C).
Effect of mechanical loading on Akt activation and FOXO transcriptional activity in diaphragms from old mice.
We investigated the effect of mechanical stretch on FOXO activity in diaphragms of mice at different ages. No significant change in basal activity was observed between different ages (Fig. 5, A and B). The downregulatory effect of mechanical stretch in diaphragms from 2-mo-old mice was not observed in the diaphragms from older animals. Mechanical stretch of the diaphragm at different ages caused a strong age-dependent increase in Akt and IKK activation; however, the stretch-dependent activation of these two protein kinases in the diaphragms from 2-mo-old mice was not observed in the diaphragms from older animals (Fig. 5, C and D). The similar basal DNA-binding activities of FOXO among the different age groups did not correlate with the higher activation of Akt and IKK in the older animals under basal conditions. To discriminate among the different FOXO isoforms, we used specific probes for Foxo1 and Foxo3a to evaluate the mechanosensitivity of Foxo1 and Foxo3a DNA-binding activities by EMSA at different ages. The EMSA results (Fig. 6, A and B) showed that Foxo1 and Foxo3a DNA-binding activities were significantly lower in stretched diaphragms of 2-mo-old mice, whereas this inhibitory effect did not occur in diaphragms of 2-yr-old animals. However, basal DNA-binding activity diminished in the old compared with young diaphragms, in contrast to our measurements of total FOXO DNA-binding activity. To clarify this discrepancy, the nuclear contents of Foxo1, Foxo3a, and Foxo4 were determined (Fig. 6, C and D). The levels of Foxo1 and Foxo3a in the nuclei diminished in response to aging to the same extent as in response to mechanical stretch in young animals, consistent with the age-dependent activation of Akt and IKK. Interestingly, Foxo4 nuclear content was not modified in response to stretch or with age.
JNK protects Foxo4 from degradation in diaphragms from old mice.
Only Foxo4 has been reported to be phosphorylated by JNK at two sites, Thr447 and Thr452, which are absent in the other FOXO isoforms. Foxo4 has been shown to be upregulated by JNK-dependent phosphorylation in response to oxidative stress and TNF-α (16). To evaluate the effect of JNK activity on FOXO, we explored the impact of JNK inhibition on total FOXO activity and the nuclear content of Foxo4 in diaphragms from young and old mice and in response to mechanical stretch. Inhibition of JNK activity had no effect on FOXO activity in young mice but altered the DNA-binding activity of FOXO in diaphragms of 2-yr-old animals independently of stretch (Fig. 7, A and B). It also affects the nuclear content of Foxo4 in diaphragms from old mice (Fig. 7D); the detection of bands of smaller molecules by the Foxo4 antibody suggests that Foxo4 is proteolyzed in the nuclei when JNK is inhibited. To assess JNK activation in response to mechanical stress in diaphragms from mice at different ages, we measured the content of the activated phosphorylated forms of JNK1 (p46) and JNK2 (p54). As shown in Fig. 7, C, E, and F, age and JNK1 activity are strongly correlated. In response to stretch, significant changes in JNK activation were observed only for JNK2 in the diaphragms of younger mice.
Our study provides novel data on the effect of age on the mechanical regulation of FOXO transcription factors in the respiratory pump. Our data support the notion that mechanical stretch downregulates FOXO activity in diaphragm muscles of young mice by a mechanism that requires Akt and IKK activation. We provided experimental evidence that muscle compliance and stress relaxation are reduced with aging of the mouse diaphragm. The decreased muscle compliance observed in diaphragms from 1- and 2-yr-old animals correlates with a failure to respond to mechanical stimuli of Akt and IKK at the same ages and, consequently, a lack of effect of mechanical stretch on FOXO activity. We also observed an age-dependent activation of Akt and IKK that correlates with a reduction in Foxo1 and Foxo3a levels in the nuclei. In contrast, Foxo4 nuclear content remains almost unaffected in the explored range of ages. Inhibition of JNK elicited a strong reduction of Foxo4 in the nuclei of muscle cells and of total FOXO DNA-binding activity in diaphragms from 2-yr-old animals. This finding demonstrates that JNK is responsible for the stability of Foxo4 in the nuclei of the old diaphragms. A strong increase in JNK1 activity occurs with age and appears to positively regulate FOXO DNA-binding activity, counteracting the inhibitory effect of the age-dependent increase in Akt and IKK signaling pathways.
The FOXO transcription factors are well-known targets of Akt signaling (for review see Refs. 5, 9, 10). Other investigators have established that the PI3K-Akt pathway is a mechanosensitive signaling pathway in several cell types, including the diaphragm muscle (13, 19, 23, 49), but the response of FOXO to mechanical stimuli has been barely explored and only reported in cardiomyocytes (45) and gingival fibroblasts (14). We previously reported that IKKs are activated in response to mechanical stimuli in the diaphragm muscle (32). Data from the present study reveal that mechanically induced IKK also contributes to the downregulation of FOXO by stretch. Interestingly, the mechanosensitivity of these two kinases is affected in a similar way with aging. Akt and IKKs lose the ability to be activated in response to mechanical stretch. More precisely, our data show a clear stretch-dependent activation in the diaphragm from young mice that is essentially lost in muscles from 1-yr-old mice. Akt can be localized to the nuclei, and IKK is basically cytoplasmic; these findings imply that Akt-phosphorylated FOXO is translocated to the cytoplasm, where it is phosphorylated by IKK. In this case, the stretch-dependent translocation of FOXO from the nuclei could be easily reversed in the absence of IKK activity, as inhibition of IKK is sufficient to completely restore FOXO basal activity. Although we were unable to detect the phosphorylated species of FOXO in the cytoplasm, phosphorylation-specific antibodies recognized lower-molecular-weight bands (not shown), suggesting that they are promptly degraded in this compartment. Whether Akt and IKK are acting sequentially, our results clearly established that Akt and IKK are necessary for the stretch-dependent inactivation of the FOXO factors, as the inhibition of Akt or IKK resulted in a lack of regulation of the FOXO transcription factors in response to mechanical stretch.
Altered mechanical properties of the aging diaphragm have been reported by others in rats and correlated with changes in the extracellular matrix components with increased cross-linking and higher content of intramuscular collagen (12, 20). Our results demonstrated clearly that aging of the mouse diaphragm caused loss of muscle compliance and reduction in stress relaxation in the direction of the muscle fibers. These observations correlated with changes in mechanoresponsiveness of the Akt and IKK signaling pathways with aging. Therefore, we speculate that the increase in muscle stiffness associated with aging could be responsible for the lack of response of the Akt and IKK pathways to mechanical stimuli. In addition, the increase in basal activation of Akt and IKKs and, consequently, lower nuclear content of Foxo1 and Foxo3a observed in diaphragms from the old animals may be consistent with continued in vivo mechanical loading of the diaphragm. The observation that Foxo1 and Foxo3a nuclear levels were not altered by aging in a less active muscle, such as the soleus (unpublished data), supports this idea. However, it is important to recognize that other physiological changes occurring with skeletal muscle aging may play a role in modulating these signaling pathways.
Another interesting observation is the distinct behavior of the FOXO isoforms in response to mechanical stretch and aging. In both conditions, activation of Akt is responsible for Foxo1 and Foxo3a exclusion from the nuclei, whereas Foxo4 nuclear content remains unaffected. Foxo4 has been reported to be modulated by Akt in other tissues by cytoplasmic translocation of the phosphorylated species (30, 47). In contrast, phosphorylation by JNK determines Foxo4 nuclear localization and activation (39). We were unable to detect the Akt-phosphorylated form of Foxo4 (data not shown); therefore, it is probable that Foxo4 is not controlled by Akt in the mouse diaphragm. Another possible explanation is that activation of JNK2 by mechanical stretch in young diaphragms opposes the effect of Akt activation on Foxo4 localization. However, inhibition of JNK caused a nonsignificant decrease in the nuclear content of Foxo4 in response to mechanical stretch in young mice.
As summarized in Fig. 8, in the young mouse, Akt activation by mechanical stress promotes Foxo1 and Foxo3a exclusion from the nuclei, which correlates with the observed downregulation of the FOXO transcriptional activity. In the cytoplasm, Foxo1 and Foxo3a are phosphorylated by IKK and further degraded. Foxo4 nuclear content remains unchanged, possibly as the result of the counteractive effect of JNK activation by mechanical stimuli. On the other hand, in the older animal, the mechanosensitivity of all the pathways controlling FOXO is lost at an early age. Compared with young diaphragms, there is an increase in basal activities of Akt and IKK, which results in lower nuclear content and activity of Foxo1 and Foxo3a. However, FOXO transcriptional activity remained at levels comparable to those of the young mice because of activation of JNK activity, which maintained Foxo4 nuclear content by protecting it from degradation.
FOXO transcription factors played important roles in the aging process (34, 35), particularly because of their ability to suppress the generation of reactive oxygen species. Activation of the FOXO homologs is associated with longevity in worms and flies (21). However, the protective role of the FOXO transcription factors in muscle aging appears somewhat controversial, because Foxo1 and Foxo3a promote the expression of two atrophy-related ubiquitin ligases, MAFbx and MURF1 (32, 44, 46). This effect has been explored in three recent reports (11, 15, 54), which support no change, downregulation, or upregulation of the expression of MAFbx and MURF1 in aged muscles. These different observations could be possibly due to the different types of muscles used in these three recent studies. FOXO transcription factors could also play a detrimental role in the aging muscles by controlling proapoptotic genes. Further investigation of the targets of Foxo4, the major isoform present in old diaphragms, and their possible favorable or detrimental effect on the aging of the respiratory pump is needed.
In summary, our study supports the hypothesis that aging alters the mechanical properties of the respiratory pump. This hypothesis is associated with altered regulation of the signaling pathways controlling FOXO activities. Our study provides the first experimental evidence that two distinct signaling mechanisms are responsible for the mechanical regulation of the FOXO transcription factors in the young and old respiratory pump, with different effects on the three members of the FOXO family.
This work was supported in part by grants from the Muscular Dystrophy Association and the National Science Foundation and National Heart, Lung, and Blood Institute Grants HL-63134 and HL-07839.
We thank Drs. Fred Pereira and Nick Timchenko for helpful discussion of some of the proposed techniques and Dr. Deshen Zhu for invaluable technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society