Compromised Exercise Capacity and Mitochondrial Dysfunction in the Osteogenesis Imperfecta Murine (oim) Mouse Model

Abstract

Osteogenesis imperfecta (OI) is a heritable connective tissue disorder that most often arises from type I collagen, COL1A1 and COL1A2, gene defects leading to skeletal fragility, short stature, blue-gray sclera, and muscle weakness. Relative to the skeletal fragility, muscle weakness is much less understood. Recent investigations into OI muscle weakness in both patients and mouse models have revealed the presence of an inherent muscle pathology. Understanding the mechanisms responsible for OI muscle weakness is critical, particularly in light of the extensive cross-talk between muscle and bone via mechanostransduction and biochemical signaling. In the following study we initially subjected wild-type (WT) and oim/oim mice, modeling severe human OI type III, to either weight –bearing (voluntary wheel running) or non-weight bearing (swimming) exercise regimens as a modality to improve muscle strength and ultimately bone strength. The oim/oim mice ran only 35–42% of the distance ran by age and sex-matched WT mice and exhibited little improvement with either exercise regimen. Upon further investigation, we determined that oim/oim gastrocnemius muscle exhibited severe mitochondrial dysfunction as characterized by 52–65% decrease in mitochondrial respiration rates, alterations in markers of mitochondrial biogenesis, mitophagy, and the electron transport chain components, as well as decreased mitochondrial citrate synthase activity, relative to age and sex-matched WT gastrocnemius muscle. Thus, mitochondrial dysfunction in the oim/oim mouse likely contributes to the compromised muscle function and reduced physical activity levels.

Introduction

Osteogenesis imperfecta (OI) is a heritable connective tissue disorder characterized most often by bone fragility, short stature, blue-gray sclera, and muscle weakness ^(1,2). OI is a heterogeneous disease, both in phenotype and genotype that ranges in severity and can be classified into four main types: I-IV, and at least nine subtypes ^(1,3). Type II is perinatal lethal; type III is the most severe viable form with patients often non-ambulatory, while patients with Types I and IV experience more mild and moderate manifestations, respectively^(1,3). Roughly 85% of OI cases are the result of autosomal dominant mutations in the genes that encode the alpha chains of type I collagen (COL1A1 and COL1A2), while the remaining cases are due to rare mutations of genes primarily involved in type I collagen post-translational modifications and folding, bone mineralization, and osteoblast differentiation^(1,2).

In addition to the skeletal fragility, patients with OI often experience muscle weakness and exhibit reduced cardiopulmonary fitness ^(4–6). Both muscle weakness and reduced cardiopulmonary fitness in OI are less characterized compared to other manifestations, and whether they reflect an inherent pathology or are the result of inactivity is only beginning to be elucidated ^(7,8). Recent investigations have begun to unveil the nature of the OI muscle weakness in both patients and mouse models, including recognition that approximately 80% of patients with type I OI due to mutations in the type I collagen genes exhibit muscle force deficits ^(9,10).

In humans, peak bone mass, the amount of bone accrued at the end of skeletal maturity, is achieved at or around the age of 30 with dramatic accumulation of bone during the pre-pubertal/pubertal growth period and then gradual decline initiating around 50 years of age⁽¹¹⁾. Bone mass and size are largely attributed to heritable factors, as well as environmental factors, such as activity levels and hormonal and nutritional status. Currently, there is no cure for OI, and treatment options for OI patients are limited primarily to anti-resorptive bisphosphonate therapy and surgical rodding ^(2,12). Although the contribution of environmental factors to bone strength are not as great as genetic factors, they are postnatally modifiable. Specifically, physical exercise has been suggested as one of the safest therapeutic options to enhance bone strength due to the mechanosensitive characteristics of bone. Several studies have demonstrated increases in bone mineral density (BMD) and bone mineral content (BMC) in more physically active children or children who underwent an exercise regimen ^(13,14). Increased BMD has been attributed partly to increases in muscle mass and strength where muscle and bone mass are positively associated. Van Brussel et al. demonstrated improved aerobic capacity and muscle force in OI patients following a physical training regimen which suggests that exercise may be a viable treatment option for patients with mild and moderate OI⁽⁸⁾.

In young and mature rodents treadmill/voluntary wheel running exercise improved BMD, bone microarchitecture and biomechanical strength^(15–20), and was protective against estrogen deficiency induced bone loss ^(19,20). Non-weight bearing swimming exercise also prevented bone loss associated with ovariectomy or hindlimb suspension ^(21–23). In the following study we investigated the effect of weight bearing (voluntary wheel running) and non-weight bearing (swimming) physical activity on musculoskeletal mass and strength in wildtype (WT) and homozygous osteogenesis imperfecta murine (oim/oim) mice. Gentry et al. was the first to identify an intrinsic muscle pathology in OI by demonstrating that oim/oim mice have reduced specific muscle contractile function (described as the peak tetanic force normalized to myofiber cross sectional area) as compared to WT littermates ⁽²⁴⁾. The oim/oim mice model moderately severe human OI type III due to a nucleotide deletion in the Col1a2 gene leading to a frameshift resulting in non-functional α2(I) chains and the formation of homotrimeric type I collagen, α1(I)_3, rather than the normal heterotrimeric type I collagen, [α1(I)₂ α2(I)]⁽²⁵⁾. We hypothesized that wildtype and oim/oim mice that underwent 1) voluntary wheel running would exhibit increased skeletal mass and strength (ultimate force) as previously seen in C57BL mice⁽¹⁶⁾, and 2) a swimming regimen would exhibit increased skeletal stiffness and sheer modulus of elasticity, with less of an impact on ultimate force ^(26,27). We demonstrate that the present exercise regimens were unable to improve either muscle or bone strength in the oim/oim mice compared to their in-cage control oim/oim mice, even though WT mice demonstrated a positive correlation between distances ran and bone strength.

In addition to the severe bone phenotype, oim/oim mice exhibit decreased activity levels, muscle weights, reduced contractile force and elevated skeletal muscle citrate synthase activity compared to age and sex matched WT littermates ^(24,28) The presence of altered citrate synthase activity and the recent report of altered energy metabolism in another moderately-severe OI mouse model, Col1a1^Jrt/+, which in younger mice exhibited reduced physical activity levels with increased oxygen consumption and CO₂ production concomitant with increased whole body energy expenditure, and altered energy homeostasis ⁽²⁹⁾, led us to investigate skeletal muscle mitochondrial bioenergetics in the oim/oim mouse. In the second half of this study, we demonstrate that oim/oim mouse skeletal muscle has poor mitochondrial respiration, decreased mitochondrial citrate synthase activity, and decreased mitochondrial DNA content in addition to alterations in mitochondrial electron transport chain (ETC) complex IV, markers of mitochondrial biogenesis, and markers of autophagy and mitophagy, suggesting a potential and critical role of mitochondria in the muscle weakness associated with OI.

Methods

Animals

Col1a2^oim mice (stock #001815) are publicly available from Jackson Laboratory (Bar Harbor, ME, USA). All mice were maintained on the C57BL/6J background and genotyped as previously described ^(25,30). Mice were housed in an AAALAC-accredited facility at the University of Missouri, and all experimental manipulations were performed under an approved University of Missouri Animal Care and Use protocol.

Exercise regimen

Four week old male and female WT and oim/oim mice were randomly divided into swimming exercise, volunteer wheel running, or control (non-exercise in-cage activity) groups. At six weeks of age, mice in the swimming exercise group were acclimated to the swimming exercise protocol in the first 3 weeks as described in Supplementary Table 1, and then full 60 minutes (3 sessions of 20 minutes of swimming exercise with 3 minutes of rest in between each session) of swimming exercise was maintained. Four to five week old mice in the volunteer wheel running group were housed individually in cages equipped with a running wheel [Kaytee Silent Spinner, 4.5inch diameter,136g (wheel weight alone; 75g), Kaytee Products, Inc, Walnut Creek, CA] assembled with cyclocomputer (SpeedZone® Comp, Specialized Bicycle Components Inc., Morgan Hill, CA) for recording max speed (km/h), average speed (km/h), distance (km), running time (hr), and odometer reading (km) daily. Mice were monitored daily for evidence of adverse effects of exercise [injury, lameness, weight loss, poor grooming, and change in behavior (major change in time on the wheel)]. At 4 months of age, mice were euthanized and hindlimb muscles and femurs evaluated.

Contractile Properties

Contractile properties of the soleus (sol), plantaris (plant), gastrocnemius (gast), and tibialis anterior (TA) muscles in male and female WT and oim/oim mice were evaluated as previously described ⁽²⁴⁾. Briefly, mice were anesthetized and the left sol, plant, gast, and TA muscle surgically exposed at their distal insertions. The distal tendon of each muscle was attached to the Grass force transducer and sequentially tested, sol➔plant➔gast➔TA. The distal tendon was adjusted in length so that the passive tension was zero grams. The sciatic nerve was isolated and placed on a stimulating electrode and a twitch was obtained as previously described ⁽²⁴⁾. At optimal length, a peak tetanic contraction (P_o) was elicited by pulses delivered at 150Hz, 300-ms duration, and an intensity of 6V for each type muscle⁽³¹⁾. Previous studies of force curves generated at 15, 50, 75, 100 and 125Hz (6V, 300ms) demonstrated that all the muscles were maximally recruited by the time 100Hz was reached ⁽²⁴⁾. All data were collected using PowerLab® (ADinstruments, Colorado Springs, CO).

Femoral Geometry and Torsional Loading to Failure

Right femora were evaluated by µCT scan analysis [Siemens Inveon µCT equipped with Siemens Inveon Acquisition Workplace Software Version 1.5 (Siemens Preclinical Solutions, Knoxville) with an X-ray peak of 80 kVp and an exposure time of 140ms] prior to ex vivo torsional loading to failure analyses as previously described⁽³⁰⁾. µCT image slices were analyzed using the Amira 5.3.3 software package (Mercury Computer Systems/TGS, Chelmsford, MA) to give a cubic voxel dimension of 0.083 mm³. The mid-shaft slice located and modeled as a hollow elliptical cross-section. The marrow cavity diameter (MCD) reported as the average of the two endosteal diameters, the cortical bone width (CBW) reported as the average of the four thicknesses on the major and minor axes of the cortical mid-slice, and the polar moment of area (K: mm⁴) determined as previously described. Following µCT analyses, the femora were potted into individualized cylindrical holders and the torsional loading to failure evaluated using the TA-HDi testing machine (Stable Micro Systems, Surrey, UK)⁽³⁰⁾. Applied torque T (Nmm) was calculated and plotted as a function of relative angular displacement θ (degrees). The whole bone parameters of strength [torsional ultimate strength (T_max, Nmm) and strain energy to failure (U, Nmm)] and stiffness [torsional stiffness (Ks, Nmm/rad), which take into account bone geometry and material properties, were determined as previously described⁽³⁰⁾. The bone material properties, tensile strength (Su, N/mm²) and stiffness [shear modulus of elasticity, (N/mm²)] were also determined according to Carleton et al⁽³⁰⁾.

Western Blot Analysis

Western blots were performed using a modified version of a previously described protocol⁽³²⁾. Briefly, right gastrocnemius muscle, previously stored at −80°C, was homogenized in lysis buffer. Protein was mixed with 6x Laemmli buffer, boiled, and loaded on a Criterion TGX Pre-cast gel (Bio-Rad, Hercules, CA). After size separation via gel electrophoresis, the proteins were transferred to PDVF membrane. 5% milk in tris-buffered saline with Tween 20 (TBST) was used for blocking. The blots were labeled with primary antibodies against the total OXPHOS Antibody Cocktail (Abcam, Cambridge, MA, ab10413) containing 5 mouse monoclonal antibodies [CI subunit NDUFB8 (ab110242), CII-30kDa (ab14714), CIII-Core protein 2 (ab14745) CIV subunit I (ab14705) and CV alpha subunit (ab14748)], TFAM (Abcam, Cambridge, MA, ab89818), PGC-1α (MilliporeSigma, Burlington, MA, 516557), BNIP3 (Cell Signaling, Danvers, MA, 3769), PINK1 (Cell Signaling, Danvers, MA, 6946), Parkin (Cell Signaling, Danvers, MA, 4211S), P62 (Cell Signaling, Danvers, MA, 5114), and LC3 A/B (Cell Signaling, Danvers, MA, 12741). Primary antibodies were diluted to 1:1,000 in 5% BSA in TBST. Secondary horseradish peroxidase-conjugated antibodies were diluted to 1:5,000 in 5% milk in TBST. Blots were imaged and analyzed using Image Lab™ Software (Bio-Rad, Hercules, CA); protein content was normalized using amido black total protein stain (Supplemental Figure 4).

Mitochondria Isolation and Respiration

Mitochondrial isolation and high resolution respirometry were carried out using modified versions of previously described methods ^(33–35). For mitochondrial isolation, both right and left gastrocnemius were removed, cleaned of extraneous fat and tissue, and placed in mitochondrial isolation buffer (100 mM KCl, 50 mM MOPS, 5 mM MgSO₄, 1 mM EGTA, and 1 mM ATP), on ice. Gastrocnemius muscles were homogenized using a rotor type homogenizer, the homogenate was centrifuged (800xg, 4°C, 10 min), and the supernatant filtered through gauze. The pellet was re-suspended in 5 mL mitochondrial isolation buffer, homogenized using a Potter-Elvehjenm homogenizer (Teflon on glass), and centrifuged (800xg, 4°C, 10 min). The supernatant was filtered, added to the previously filtered supernatant, and centrifuged (12,000xg, 4°C, 10 min). The pellet was re-suspended in 3 mL isolation buffer with glass on glass homogenization, centrifuged (8,000xg, 4°C, 10 min), and the process repeated with 2 mL isolation buffer. 0.2% Bovine Serum Albumin (BSA) was added to 2 mL of the isolation buffer and used to re-suspend the pellet a third time via glass on glass homogenization followed by centrifugation (8,000xg, 4°C, 10 min). Finally, the pellet was re-suspended in 200 µL MiPO₃ buffer (0.5 mM EGTA, 3 mM MgCl2·6H20, 60 mM K-lactobionate, 20 mM Taurine, 10 mM KH2P04, 20 mM HEPES, 110 mM Sucrose, 1g/l BSA, 20 mM Histidine, 20 μM vitamin E succinate, 3 mM glutathione, 1 μM leupeptine, 2 mM glutamate, 2 mM malate, 2 mM Mg-ATP)⁽³³⁾ and allowed to equilibrate on ice for 30 min.

High resolution respirometry was used to measure mitochondrial respiration by the isolated mitochondria using the Orosboros Oxygraph-2k (Orosboros Instruments; Innsbruck, Austria). 30–45 µl of isolated mitochondria were added to the respiration chambers to measure basal respiration. Steady-state oxygen flux was measured by the addition of 2 mM malate and 5 mM glutamate. Oxygen flux through complex I was measured by the titration of 125–375 µM of ADP; through complex I+II by the titration of 1–5 mM succinate, and maximal uncoupled respiration was measured by the addition of 0.125 µM FCCP. Finally, 10 µM cytochrome C was added to assess the quality of mitochondrial isolation. All measurements were normalized to the mitochondrial protein content.

Transmission Electron Microscopy and Analysis

Soleus muscle was excised, cleaned of any extraneous tissue or fat and placed in glutaraldehyde fixative at room temperature for 1 hour before being transferred to 4°C. Tissue was processed and prepared by the University of Missouri Electron Microscopy Core. Samples were imaged using a JEOL JEM-1400 120kV TEM. Images were captured at a magnification of 3,000X. Analysis was done using classical stereological methods as previously described to calculate the following, intermyofibrillar mitochondria: mitochondrial volume density, cross sectional area, and numerical density using the low magnification images ^(36,37).

Mitochondrial DNA Content

Mitochondrial DNA content was measured as a ratio of mitochondrial DNA to nuclear DNA using quantitative PCR (qPCR) methods and primers described by Malik et al ⁽³⁸⁾. Briefly, DNA was isolated from gastrocnemius muscle using a DNeasy Blood and Tissue Kit (Qiagen, Germantown, MD). qPCR was carried out in 10 µl reaction volume containing 5 µl PowerUp SYBR green master mix, 0.5 µl forward primer, 0.5 µl reverse primer, and 4 µl of diluted DNA. The mitochondrial DNA forward and reverse primers used were mMitoF1 (5’-CTAGAAACCCCGAAACCAAA-3’) and mMitoR1 (5’-CCAGCTATCACCAAGCTCGT-3’), respectively, and the nuclear DNA forward and reverse primers used were mB2MF1 (5’-ATGGGAAGCCGAACATACTG-3’) and mB2MR1 (5’-CAGTCTCAGTGGGGGTGAATA-3’), respectively. A standard curve with copy numbers ranging from 10² – 10⁹ was amplified simultaneously with the isolated DNA used to quantify the ratio of mtDNA to nuclear DNA.

Citrate synthase activity

Citrate synthase activity was measured in mitochondria isolated from mixed gastrocnemius muscle using methods by Sere et al.⁽³⁹⁾, as previously described⁽³²⁾.

Glycogen Content

Snap frozen gastrocnemius muscle was powdered using a mortar and pestle on liquid nitrogen and approximately 15–30 mg of powdered muscle was added to 0.5 mL of 1 N HCL. The muscle was placed in a boiling water bath for 2.5 hours. After boiling, tubes were vortexed, and 1.5 mL of 0.67 M NaOH was added. The acid hydrolysis converts glycogen to free glucose which was then measured using Infinity Glucose Hexokinase Liquid Stable Reagent (Thermo Fischer Scientific, Waltham, MA) according to the manufacturer protocol using a sample to reagent ratio of 22.5:1⁽⁴⁰⁾.

Statistical Analysis

Statistical analyses were performed using SAS (SAS Institute Inc., Cary, NC). Swimming and voluntary wheel running exercise regimens were analyzed separately as a 2X2X2 factorial [2 genotypes (WT and oim/oim) X 2 sexes X 2 treatments (control and exercise) and mitochondrial analyses were analyzed as 2X2 factorials [2 genotypes (WT and oim/oim) X 2 sexes] using Fisher’s Protected Least Significant Difference^(41,42). If heterogeneous variations were present, a log transformation was used to stabilize the variation. If the log transformation failed to stabilize the variation, a nonparametric ranked analysis was performed according to Conover et al⁽⁴³⁾. Differences were considered significant at p≤0.05.

Results

Effect of volunteer wheel running and swimming exercise on muscle weight and function

As was demonstrated in our previous study ⁽²⁴⁾, oim/oim mice exhibited reduced hindlimb muscle weights and absolute contractile forces as compared to WT littermates (Figure 1), with minimal impact on relative contractile forces. Previously the decreased absolute contractile forces in male oim/oim hindlimb muscles was directly associated with decreased specific contractile force in male oim/oim plantaris (plant), gastrocnemius (gast), and tibialis anterior (TA) muscles, even though relative contractile forces did not always reach significance^(24,28). Volunteer wheel running and swimming exercised male oim/oim mice exhibited decreases in plant, gast, and TA muscle wet weight when compared to in-cage control counterparts (Figure 1 B-D). Peak tetanic force of plant, gast, and TA muscle in volunteer wheel running exercised and plant and TA muscle in swimming exercised male oim/oim mice were reduced as compared to their in-cage control counterparts (Figure 1F-H). These decreases were still evident in the relative contractile force [peak tetanic force (P_o;g)/muscle weight (mg)] of volunteer wheel running exercised male oim/oim plant and gast muscles, and swimming exercised male oim/oim plant muscle (Figure 1J&K). However, interestingly, female mice did not exhibit any changes in muscle weight or function with both type of exercises regardless of genotypes (Supplemental Figure 1). Both male and female oim/oim mice ran shorter total distances, 42% and 35% of the total distances ran by WT littermates, respectively (Supplemental Figure 2). The average speed (km/hr) ran per week was also reduced in male and female oim/oim mice being 57% and 63% of WT speeds, respectively. In addition to reduced distance and average speed, the average time on the wheel (minutes) per week spent by oim/oim was 37% and 51% less that WT mice for both males (p=0.122) and females, respectively (Supplemental Figure 2).

Figure 1:

Volunteer wheel running and swimming exercised male oim/oim mice had reduced hindlimb skeletal muscle weight and peak tetanic force as compared to in-cage control counterparts. (A) Soleus, (B), Plantaris, (C) Gastrocnemius, and (D) Tibialis Anterior muscle wet weights. (E) Soleus peak tetanic force (P_o), (F) Plantaris Peak P_o, (G) Gastrocnemius Peak P_o, (H) Tibialis Anterior Peak P_o, (I) Soleus relative P_o, (J) Plantaris relative P_o, (K) Gastrocnemius relative P_o, and (L) Tibialis Anterior relative P_o of 4-month-old male WT and oim/oim in-cage control, volunteer wheel running, and swimming exercised mice. Graphs are presented as MEAN±STDEV. *p≤0.05 vs. WT; †p≤0.05 vs. in-cage control. (n=6–13 per group). Red and blue symbols represent WT and oim/oim, respectively. Circles, squares, and triangles represent control, wheel running, and swimming, respectively.

Effect of volunteer wheel running and swimming exercise on skeletal properties

Similar to previous investigations, oim/oim mice exhibit compromised bone microarchitecture and biomechanical properties as compared to their WT littermates regardless of sex and exercise regimen (Supplemental Tables 2&3, Figure 2 A-D). Male WT mice that underwent volunteer wheel running exercise demonstrated decreases in several geometric and biomechanical parameters, including cortical bone width (CBW), polar moment of area (K), torsional ultimate strength (T_max), torsional stiffness (Ks), and energy to failure (U), while shear modulus of elasticity increased as compared to in-cage control counterparts (Figure 2 A & C and Supplemental Table 2). Swimming exercised male WT mice also showed decreases in CBW and K and increase in marrow cavity diameter (Supplementary Table 2). Male oim/oim and female WT and oim/oim mice did not exhibit any changes in bone geometric or biomechanical properties with either of exercise regimen; except that female WT mice had decreases in T_max and tensile strength (Su) and female oim/oim mice had decreased in T_max with swimming exercise (Figure 2 B and D and Supplementary Table 3). Pearson’s correlation coefficients were evaluated between total distances run and bone biomechanical parameters in WT and oim/oim mice. The total distances run by male and female WT mice exhibited positive correlations with T_max (+0.508; p=0.022) and U (+0.500; p=0.025), while oim/oim mice did not (Figure 2E-H). The lack of positive correlations in oim/oim mice suggest that oim/oim mice are not able to respond as WT mice to weight bearing exercise, leading us to further question the musculoskeletal function of oim/oim.

Figure 2:

Oim/oim mice exhibited compromised femoral biomechanical properties as compared to WT littermates. Volunteer wheel running exercised WT mice showed positive correlation between biomechanical properties and the total distance ran (km), but oim/oim mice did not. Torsional ultimate strength (T_max) of (A) male and (B) female mice; energy to failure (U) of (C) male and (D) female mice. Red and blue symbols represent WT and oim/oim, respectively. Circles, squares, and triangles represent control, wheel running, and swimming, respectively.

Linear regression analysis of the correlation between T_max and total distance run by (E) WT and (F) oim/oim mice, linear regression analysis of the correlation between U and total distance run of (G) WT and (H) oim/oim mice. Graphs are presented as MEAN±STDEV. *p≤0.05 vs. WT; †p≤0.05 vs. in-cage control. (n=7–13 per group). White and black squares represent male and female animals, respectively.

Protein levels of electron transport chain (ETC) complexes

Previously, our group determined that citrate synthase activity was increased in oim/oim mixed gastrocnemius muscle homogenates relative to WT mixed gastrocnemius muscle homogenates⁽²⁸⁾, which in light of the response of oim/oim mice to the exercise regimens, prompted the evaluation of the ETC complexes in the current study. Oim/oim mice exhibited reduced levels of ETC complex IV, mitochondria encoded cytochrome oxidase I [51% decrease in females and 20% decrease (which did not reach significance) in males] as compared to WT counterparts (Figure 3), while the protein levels of other complexes did not differ between genotypes.

Figure 3:

Female oim/oim mixed gastrocnemius muscle homogenate had reduced ETC complex IV mitochondrial encoded cytochrome c oxidase I (c4). Western blot analysis of electron transport chain (ETC) complexes measured in 4-month-old mixed gastrocnemius muscle of (A) male and (B) female WT (red diamonds) and oim/oim (blue squares) mice. (C) Representative blot images. Graphs are presented as MEANS±SE. *p≤0.05 vs. WT. (n=5–6 per group).

Skeletal Muscle Mitochondrial Function

To investigate skeletal muscle mitochondrial function more directly, mitochondria were isolated from 4-month-old male WT and oim/oim mice mixed gastrocnemius muscles, and oxygen consumption was measured using high resolution respirometry. Oim/oim skeletal muscle mitochondria demonstrated reductions in respiration during basal (−65%), state 2 (−52%), state 3-Complex I (−61%), state 3-Complex I+II (−53%), and maximal uncoupled respiration (−52%) as compared to WT skeletal muscle mitochondria (Figure 4A). These decreases clearly demonstrate inefficient mitochondrial function with compromised energy production in oim/oim gastrocnemius muscle.

Figure 4:

(A) Male oim/oim mice exhibit reduced mitochondrial respiration and (B) mitochondrial citrate synthase activity compared to WT littermates. (A) Basal, state 2, state 3, and maximal uncoupled mitochondrial respiration was significantly decreased in 4-month-old oim/oim (blue squares) male mice compared to WT (red diamonds) littermates. Oxygen consumption was measured with the addition of substrates to stimulate different respiration states: basal (mitochondria in MiPO₃ buffer), state 2 (Glutamate and Malate (GM)), state 3 respiration with electron flux through complex I (GM and ADP), state 3 respiration with electron flux through complex I and II (succinate), and maximal uncoupled respiration (FCCP). (B) Citrate synthase activity was decreased in isolated mitochondria from 4-month-old male oim/oim mice compared to WT littermates. Graphs are presented as Means±SE of the genotype values. *p≤0.05 vs. WT. (n=5–8 per group).

In addition, the isolated oim/oim gastrocnemius muscle mitochondria exhibited a 33% decrease in citrate synthase activity as compared to WT gastrocnemius muscle mitochondria (Figure 4B). While citrate synthase activity in whole muscle homogenate is often an indirect marker of mitochondrial content/mass, in isolated mitochondria, citrate synthase activity can provide insight into mitochondrial function ^(44–46).

Mitochondrial Biogenesis and Mitophagy

To determine if oim/oim mice have altered mitochondrial turnover, markers of mitochondrial biogenesis and autophagy/mitophagy were evaluated. Mitochondrial biogenesis markers, PGC1-α (a nuclear co-activator) and TFAM (a mitochondrial transcription factor) ⁽⁴⁷⁾ were increased in male oim/oim mixed gastrocnemius muscle while TFAM was increased in female oim/oim mixed gastrocnemius muscle compared to their WT counterparts (Figure 5A, Supplemental Figure 3A). Several markers of autophagy and mitophagy were also evaluated including LC3, a known component of autophagosome formation, as well as the LC3-II to LC3-I ratio, a method to assess autophagy; with increases in the LC3-II/I ratio being suggestive of elevated autophagy ^(47–50) and warranting direct studies of autophagy flux. In the skeletal muscle of male oim/oim mixed gastrocnemius, the LC3-II/I ratio was reduced by greater than 50% compared to WT counterparts; a decrease largely driven by increases in LC3-I (117%)(Figure 5B & C). The levels of other autophagy/mitophagy markers including PINK1, Parkin, BNIP3, and p62 did not differ between genotypes.

Figure 5:

Male oim/oim mice exhibit increased markers of mitochondrial biogenesis and decreased markers of autophagy and mitophagy. Western blot analysis of markers of (A) mitochondrial biogenesis (PGC-1α and TFAM) and (B) autophagy and mitophagy (PINK1, Parkin, BNIP3, LC3 II/I, and p62) in protein homogenate isolated from 4-month-old male oim/oim mixed gastrocnemius muscle. (A) TFAM and PGC-1α (p=0.07), a marker of mitochondrial biogenesis was increased in oim/oim (blue squares) mice compared to WT (red diamonds) mice and (B) LC3 II/I, a marker of mitophagy, was decreased in oim/oim mice compared to WT mice driven by an increase in (C) LC3-I. Graphs are presented as MEANS±SE. *p≤0.05 vs. WT; †p=0.07 vs WT. (n=5–6 per group).  

To determine if the decrease in the LC3-II/I ratio seen in the whole gastrocnemius muscle homogenate was indicative of decreased mitochondrial degradation rather than decreases in general autophagy, we measured LC3-II and BNIP3 levels in isolated mitochondria of mixed gastrocnemius muscle and found that LC3-II levels in the 4-month-old male oim/oim mice were dramatically decreased compared to their WT counterparts while BNIP3 levels appeared reduced but did not reach significance, further suggesting that mitophagy is decreased in the skeletal muscle of oim/oim mice (Figure 6).

Figure 6:

BNIP3 and LC3-II levels were measured in isolated mitochondria of 4-month-old male WT (red diamonds) and oim/oim (blue squares) mixed gastrocnemius muscle. LC3-II levels were significantly reduced in oim/oim mitochondrial compared to WT mitochondria. Graphs are presented as MEANS±SE *p≤0.05 vs WT. (n=6–8 per group).

Mitochondrial Content

Electron microscopic analyses of 4-month-old male oim/oim and WT soleus muscle found no evidence of gross abnormal morphological changes in oim/oim soleus muscle. In addition no differences were found in intermyofibrillar mitochondrial volume density, cross-sectional area, and numerical density (Figure 7). To investigate further, the relative mitochondrial DNA content was evaluated by the ratio of mitochondrial DNA to nuclear DNA in 4-month-old male and female oim/oim and WT gastrocnemius muscles. The male and female oim/oim gastrocnemius mitochondrial/nuclear DNA ratios were decreased by 40% and 34% relative to the mitochondrial/nuclear DNA ratios of sex matched WT gastrocnemius muscles (although not reaching significance), respectively (Figure 8).

Figure 7:

Ten representative images of 4-month-old WT (red diamonds) and oim/oim (blue squares) intermyofibrillar mitochondria from soleus muscle were assessed using Transmission Electron Microscopy and classical stereological methods. The intermyofibrillar (A) mitochondria volume density, (B) cross sectional area, and (C) mitochondrial numerical density as well as gross morphology were not different between (D) WT and (E) oim/oim soleus muscle. Graphs are presented as MEANS±SE. (n=3 per group). 

Figure 8:

Oim/oim mice exhibit a reduced mitochondrial DNA to nuclear DNA ratio. The ratio of mitochondrial DNA copy number to nuclear DNA copy number was decreased by 40% in (A) male and by 34% in (B) female 4-month-old oim/oim (blue squares) mixed gastrocnemius muscle as compared to WT (red diamonds) mixed gastrocnemius muscle. Graphs are presented as MEANS±SE. †p=0.08 vs. WT. (n=8–11 per group)

Skeletal Muscle Glycogen Content

Effective glycogen metabolism provides glucose to muscle cells to induce muscle contraction, and alterations could result in reduced muscle contractile forces. To investigate glycogen metabolism, skeletal muscle glycogen content was evaluated in WT and oim/oim gastrocnemius muscles. No differences between WT and oim/oim mice were found, further suggesting that the mitochondrial dysfunction itself is likely the major contributor to the reduced contractile force in oim/oim mice (Figure 9).

Figure 9:

Glycogen content is equivalent in oim/oim (blue squares) relative to WT (red diamonds) mixed gastrocnemius muscle. Glycogen content was measured in gastrocnemius muscle of 4-month-old (A) male and (B) female oim/oim and WT. Graphs are presented as MEANS±SE. (n=4–5 per group).

Discussion

The beneficial impact of exercise on overall health is well established, and particularly its positive impact on musculoskeletal health during the pre-pubertal/pubertal growth period ^{(8,13,14,51,52)}. We chose to test the impact of two different exercise modalities, voluntary wheel running (weight-bearing) and swimming (non-weight bearing)⁽⁵³⁾ on WT and oim/oim hindlimb muscle and bone. Voluntary wheel running provides unlimited access without aversive stimuli; but it does not control for duration and intensity. Swimming provides minimal ground reaction forces and aqua therapy is often a physiotherapeutic recommendation for individuals with osteogenesis imperfecta⁽⁵⁴⁾, however it may induce physiological stress responses, at least initially, in mice⁽⁵³⁾. Unlike previous reports of the positive influence of physical exercise on muscle and bone health ^(8,14), we found male WT and oim/oim mice exhibited decreases in muscle weight and function and slightly compromised bone microarchitecture and biomechanical properties, while female mice were minimally affected. Sex differences in bone structure and properties in the oim mouse model have previously been reported ⁽⁵⁵⁾. One caveat between our exercise study and the previous reports of the positive impact of physical exercises is that our study focused only on aerobic exercise (swimming and voluntary wheel running) while exercise studies in OI patients incorporated a variety of activity including aerobic and strength training exercises ^(8,14). In addition, our exercise study focused on mice modeling severe OI while previous reports were in children with mild and moderate OI ⁽⁸⁾. This suggests that both the type of exercise and the type of OI are important factors when considering exercise as a potential treatment option. WT mice ran greater than average distances compared to their age-matched oim/oim counterparts regardless of sex. This is consistent with previous studies of reduced physical activity levels in oim/oim mice as well as in Col1a1^Jrt/+mice, a model of moderately severe OI as a result of a Col1a1 gene defect ^(28,29). Growing evidence suggests bone is most responsive to dynamic moderate to high magnitude loads, which are quick, short in duration, and non-repetitive in direction ^(56,57). Interestingly, WT mice showed positive correlations between total distances run and bone strength, while oim/oim mice did not, which may reflect impaired musculoskeletal response to the weight bearing exercise or that the reduced distances ran, average speeds and time spent on the wheel by the oim/oim mice did not reach the minimal threshold for musculoskeletal responsiveness. The skeleton is striving for the optimum geometry and mass to be able to bear loads while remaining light enough for efficient mobility⁽⁵²⁾. Therefore, one should consider that the lack of an overall positive response to the extended running and swimming regimens may also reflect desensitization of the osteocytes to repetitive low magnitude loads ^(58–60). In separate studies with rats extended durations of running exercise have been shown to not change or to decrease BMD and serum alkaline phosphatase levels, a bone formation biomarker ^(15,20).

The lack of positive musculoskeletal effects in both WT and oim/oim animals that underwent training regimens suggests that perhaps a more controlled exercise regimen consisting of both endurance and strength training exercises with increasing loads is required to improve muscle and bone properties.

In a previous study, whole muscle homogenates from oim/oim gastrocnemius muscle had increased citrate synthase activity as compared to WT gastrocnemius muscle homogenates ⁽²⁸⁾, which led us to hypothesize that oim/oim mice have either 1) increased mitochondrial numbers (mitochondrial mass) or 2) increased mitochondrial function as compared to WT mice. However, our current study demonstrates that oim/oim mice have dramatic reductions in skeletal muscle mitochondrial respiration rates, markers of autophagy and mitophagy, ETC complex IV mitochondrial encoded cytochrome oxidase I, and mitochondrial citrate synthase activity with more mild reductions in mitochondrial DNA /nuclear DNA ratios compared to WT littermates. In addition, markers of mitochondrial biogenesis are increased in oim/oim mice, which may reflect compensatory increases in an attempt to overcome the severely reduced mitochondrial respiration or mitochondrial damage.

The decreases in mitochondrial respiration, ETC complex IV protein content, and mitochondrial DNA copy number may be reflective of the functional oim/oim muscle phenotype of decreased muscle contractile force in oim/oim mice ⁽²⁴⁾. The mitochondria are responsible for generating the energy necessary for muscle contraction ^(61,62), and are the primary source of ATP production; ATP is generated by complex V of the ETC chain, ATP synthase, via a proton gradient which is primarily generated and sustained as electrons are transferred through complex IV to oxygen while four protons are simultaneously pumped across the membrane ^(61,63). The decreased mitochondrial oxygen consumption in skeletal muscle of oim/oim mice suggests that the oim/oim mitochondrial function is not as efficient as WT mitochondria. Decreased ETC complex IV mitochondrial encoded cytochrome oxidase I in oim/oim gastrocnemius muscle is consistent with the decreased potential to transfer electrons to oxygen which may lead to the inability to maintain an adequate proton gradient.

Decreased mitochondrial DNA / nuclear DNA ratios suggests decreased mitochondria number ⁽³⁸⁾, although it is controversial as to whether mitochondrial DNA/nuclear DNA ratio is an adequate marker for mitochondrial content especially due to the large variation in DNA copy number per mitochondria⁽⁴⁵⁾. However, increased protein markers of mitochondrial biogenesis, PGC-1α and TFAM, and decreased LC3-II in isolated mitochondria favor increased mitochondrial number ^(45,47). Although, mitochondrial DNA appears reduced in oim/oim gastrocnemius, increased PGC-1α and TFAM support the hypothesis that oim/oim mice are trying to modulate the signaling pathways in favor of increasing mitochondria⁽⁴⁷⁾, likely to compensate for reduced function. This would be consistent with the indirect measure of mitochondrial content, whole muscle citrate synthase activity, previously observed to be increased in oim/oim whole muscle gastrocnemius protein homogenates⁽²⁸⁾.

This is the first report of severe muscle weakness in an OI mouse model being attributed, at least in part, to compromised mitochondrial function. Another moderately-severe OI mouse model, Col1a1^Jrt/+, which has an autosomal dominant 18 amino acid deletion in the triple helical domain of the proα1(I)collagen chain⁽⁶⁴⁾, exhibits reduced physical activity levels with increased oxygen consumption and CO₂ production concomitant with increased whole body energy expenditure and altered energy homeostasis ⁽²⁹⁾. Interestingly, we did not observe any changes in skeletal muscle glycogen content of oim/oim mice compared to WT littermates, further suggesting that the muscle defects, specifically reduced contractile generating force, may be associated with mitochondrial dysfunction, but future studies warrant the inclusion of fatiguing protocols to decipher the role of mitochondria in oim/oim muscle weakness more clearly. Further investigation into metabolic homeostasis and energy utilization and expenditure in the oim/oim mouse are critical to understanding the altered mitochondrial function.

Although the role of the extracellular matrix (ECM) as a potential contributing factor to the mitochondrial dysfunction observed in the oim/oim mouse remains to be investigated, the role of the ECM in reduced contractile force and the mechanotransduction of force also needs to be elucidated. The skeletal muscle ECM is hierarchical and can be divided into three structures: the endomysium, perimysium, and epimysium ^(65,66). The endomysium is a thin continuous extracellular matrix sheath containing collagen (types III, IV, V, VI, XII, and very little type I collagen) covering the myofibers. The endomysium interacts with the perimysium, which is composed largely of type I collagen and surrounds the muscle fascicles. The perimysial junctional plates (PJP) are the focal site of contact between the perimysium and the myofiber⁽⁶⁷⁾. Through the PJP the perimysium is responsible for the lateral transmission of the contractile forces to the epimysium (largely composed of type I collagen) and the myoteninous junction into the tendon and subsequent tendon-bone interface ^(65–68). The presence of not only reduced amounts of type I collagen⁽²⁴⁾, but also abnormal homotrimeric type I collagen in the oim/oim muscle contributes to an altered ECM, which likely disrupts force transmission and altering muscle-bone crosstalk. Passerieux et al. demonstrated intracellularly at the sites of the PJP the presence of type I collagen bundles (perimysial) in association with the subsarcolemmal mitochondria and myonuclei clusters⁽⁶⁸⁾. These findings suggest that although the intermyofibrillar mitochondria of the oim/oim soleus muscle did not appear quantitatively or morphologically abnormal by EM, it is important to evaluate the subsarcolemmal mitochondria. Abnormal muscle ECM has been demonstrated by defects in type VI collagen, a component of the perimysium, seen in Bethlem myopathy and Ullrich congenital muscular dystrophy, leading to mitochondrial dysfunction, calcium deregulation, elevated reactive oxygen species, increased apoptosis, and muscle weakness ^(69,70). Future investigation into the role of the ECM in oim/oim skeletal muscle weakness may provide insight into the mechanisms of mitochondrial dysfunction in the oim/oim skeletal muscle.

Understanding the mechanisms of muscle weakness in OI are an important consideration for OI treatment, as bone is responsive to the force it feels from muscle contraction. Physiotherapeutic studies that target the muscle weakness in OI are critically needed⁽⁷¹⁾. Severe muscle weakness or disuse hinder patient quality of life as well as negatively impacting bone health. Understanding the mechanism(s) of muscle weakness in OI will provide new therapeutic targets that will have the potential to improve both muscle and bone strength.

This study is one of the first to demonstrate that extended voluntary wheel running and swimming exercise regimens were well tolerated by the oim/oim mouse a model of moderately severe OI. Although wild-type mice exhibited positive correlations for total distance run and bone biomechanical strength, this was not true for the oim/oim mice, suggesting that the oim/oim mouse muscle-bone unit mechano-responsiveness is compromised. This study is the first to demonstrate reduced mitochondrial function in skeletal muscle as a potential contributing factor to muscle weakness in a mouse model of OI, as evidenced by alterations in mitochondrial content and respiratory function. Further studies including the evaluation of reactive oxygen species and calcium homeostasis are essential to characterize the mechanism of mitochondrial dysfunction in oim/oim mice and to understand the impact of the mitochondrial energy utilization and expenditure on muscle weakness. Exploring the mechanisms responsible for muscle weakness and whole-body metabolic alterations promises to provide important insight into the musculoskeletal pathogenesis of OI.