High doses of methylprednisolone causes substantial muscle loss with minimal impact on bone mass and architecture in C57BL/6JRj and RjOrl:SWISS mice

Abstract

Glucocorticoids are widely used to treat chronic autoimmune and inflammatory disorders. Unfortunately, glucocorticoid treatment frequently results in severe adverse effects, such as skeletal muscle atrophy and glucocorticoid-induced osteoporosis (GIO). Developing a reliable small animal model of GIO has been challenging, especially in achieving consistent bone loss, microstructural deterioration, and reduced bone strength. The study aimed to develop a robust small animal model to evaluate the effects of high doses of methylprednisolone (MP) on muscle and bone tissue simultaneously. Two strains of female mice, C57BL/6JRj (C57) and RjOrl: SWISS (SWISS), were divided into four groups each:  (1) Control, (2) MP 15 mg/kg/day, (3) MP 22.5 mg/kg/day, and (4) MP 30 mg/kg/day. While MP caused a reduction in muscle volume in both strains, it did not result in consistent bone deterioration in either strain. Furthermore, the sparse bone loss manifested differently in the two strains. Therefore, although high-dose MP effectively induced muscle atrophy, further refinement is needed in order to achieve consistent bone deterioration. Moreover, unlike previous studies, the musculoskeletal system of SWISS mice did not exhibit greater susceptibility to MP compared to that of C57 mice.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-29729-1.

Introduction

Glucocorticoids (GC) are widely used in the management of chronic autoimmune and inflammatory disorders¹. Between 1989 and 2008 the annual prevalence of oral GC use was 0.85% in the adult UK population². Unfortunately, systemic treatment with GCs is frequently accompanied by a rapid and profound bone loss. This adverse effect is termed glucocorticoid-induced osteoporosis (GIO), and it is characterized by a loss of bone mineral density (BMD) and a deterioration of the bone microarchitecture, increasing the risk of bone fractures³. As many as 30–50% of patients undergoing long-term systemic treatment with GCs will experience a bone fracture if preventive measures are not initiated at the onset of GC treatment⁴. Another common GC-induced adverse effect is skeletal muscle wasting, which can further exacerbate the risk of bone fractures⁵. The current pharmacological approach to prevent GIO involves the use of either anti-resorptive or osteoanabolic medications⁶. However, the latter is limited to a maximum of 24 months of treatment over a lifetime, after which preventive treatment of GIO is restricted to anti-resorptive medications only⁷. Importantly, the current preventive strategy focuses on minimizing the risk of bone fractures but does not address the accompanying muscle wasting. This limitation highlights the need for further research to develop new therapeutic interventions that simultaneously mitigate the detrimental effects of GCs on both skeletal muscle and bone tissue. To achieve this goal, a robust small animal model of skeletal muscle atrophy and GIO is essential.

In medical research, small animals such as mice and rats are preferred over larger animals as the initial species for investigation due to ethical and cost considerations. In addition, small animal models of both postmenopausal osteoporosis and disuse osteoporosis have already been established in mice⁸, making them an obvious choice for investigating skeletal muscle atrophy and GIO. However, to date, a robust small animal model of GIO has not yet been established. In a 2021 systematic review, Xavier et al. proposed that an appropriate model of GIO must include at least the following three outcome measures: (1) a loss of BMD, (2) changes in bone microstructural parameters, and (3) a reduction in mechanical bone strength⁹. They found that only four out of 18 studies using mouse models showed a reduction in all three outcome measures. In an earlier systematic review, Wood et al. examined 19 studies that were not included by Xavier et al. Of the studies included by Wood et al., only two met the later proposed criteria by Xavier et al. for an appropriate model of GIO¹⁰. Many of the examined studies did not investigate either BMD or mechanical bone strength, highlighting the need for a standardization of the outcome measures of GIO animal models. Among the studies that met these criteria, there was a lack of consistency regarding the strain and age of mice, the type of glucocorticoid used, and the duration of the study. The doses of GC used in the studies was relatively low, with 80% of the studies using methylprednisolone (MP) equivalent doses of 5 mg/kg/day or less, indicating that higher doses may be necessary to establish a robust model. Additionally, the studies did not explore the effects of GC on skeletal muscle, highlighting the need for further research into the concurrent effects of GC on both muscle and bone tissue in small animal models.

The C57BL/6JRj (C57) mouse is the most common mouse strain in laboratory research¹¹, making it an obvious candidate for a skeletal muscle atrophy and a GIO model. However, a study from Ersek et al. found that CD-1 mice – an outbred strain originating from Swiss Webster mice, similar to RjOrl: SWISS (SWISS) mice – showed greater susceptibility to GIO than C57 mice¹².

Therefore, the aim of the present study was to investigate the effect of high doses of MP (up to 30 mg/kg/day) on the musculoskeletal system in both C57 and SWISS mice and to evaluate the potential of both strains as a model of skeletal muscle atrophy and GIO using the outcome measures proposed by Xavier et al.

Materials and methods

Animal handling and procedures

The study comprised two strains of 16-week-old female mice: C57BL/6JRj (n = 32) and RjOrl: SWISS (n = 32) (Janvier Labs, Le Genest-Saint-Isle, France). The mice in both strains were stratified by weight into four groups (n = 8 per group): (1) Control, (2) MP 15 mg/kg/day, (3) MP 22.5 mg/kg/day, and (4) MP 30 mg/kg/day. The mice received subcutaneous (s.c.) injections, 7 days a week, of either MP (Solu-Medrol, Pfizer, New York City, NY, USA) or vehicle (phosphate-buffered saline).

The mice were housed in laboratory animal cages (n = 4 per cage) under controlled conditions: temperature of 22 °C, humidity of 50%, and a 12 h/12 h light/dark cycle. The mice had ad libitum access to water and standard rodent chow and were weighed twice weekly.

In order to study bone formation and resorption, fluorochrome labels were administered to the mice. The C57 mice were injected s.c. with tetracycline (20 mg/kg, T3383; Sigma-Aldrich, St. Louis, MO, USA) 5 days prior to initiation of the study, and calcein (20 mg/kg, C0875; Sigma-Aldrich, St. Louis, MO, USA) and alizarin (20 mg/kg, A3882; Sigma-Aldrich, St. Louis, MO, USA) 9 and 5 days prior to the end of the study, respectively. The SWISS mice were injected s.c. with calcein 5 days prior to initiation of the study, and alizarin 9 and 5 days prior to the end of the study, respectively.

After 28 days, the mice were euthanized by removal of the heart under general anesthesia with 2.5% inhaled isoflurane (IsoFlo Vet; Orion Pharma Animal Health, Copenhagen, Denmark). No mice died prematurely. The animal care and all the experimental procedures were conducted in accordance with the the ARRIVE guidelines and the guiding principles and regulations set out by the Danish Animal Experiment Inspectorate and were approved with the licensing number 2023-15-0201-01404.

In-vivo micro-computed tomography (µCT) of lower hindlimb musculature

On day zero and immediately before euthanasia, the volume of part of the lower hindlimb musculature was determined using in-vivo µCT (Fig. 1a). The mice were anesthetized with 2.5% inhaled isoflurane and placed on a horizontal bed in a scanner (Viva80, Scanco Medical AG, Brüttisellen, Switzerland). The right hindlimb was fully extended and securely fixed using a customized holder to minimize motion artifacts from respiration. The lower part of the right hindlimb was imaged in medium resolution mode (500 projections/180°) with an isotropic voxel size of 62.4 μm, an X-ray tube voltage of 45 kVp, an X-ray current of 177 µA, and an integration time of 200 ms. Subsequently, a 4,992-µm-high volume-of-interest (VOI) was positioned distally from the most proximal part of the fibular head. The outer boundary of the muscles was delineated using the contour tool (Scanco µCT Evaluation Program version 6.6), and the outer boundary of the bone (including bone and marrow volume) was delineated. The data were low-pass filtered with a Gaussian filter (σ = 0.8 and support = 1), and muscle volume was determined as the total tissue volume (muscle plus bone) minus the bone volume (including marrow).

Fig. 1.

Overview of the volume-of-interest (VOI) investigated using µCT. (a) VOI for in-vivo µCT analysis of the lower hindlimb musculature, comprising a 4,992-µm-high region placed distally from the most proximal part of the fibular head. The semitransparent red overlay represents the skeletal musculature. (b) VOI for µCT analysis of the L4 vertebral body, encompassing the entire region between the upper and the lower growth zones. (c) VOI for µCT analysis of the femoral mid-diaphysis, defined as an 820-µm-high region centered at the midpoint of the femur. (d) VOI for µCT analysis of the femoral metaphysis, comprising a 1,000-µm-high region positioned 200 μm proximally from the upper part of the growth zone. (e) VOI for µCT analysis of the femoral epiphysis, comprising the entire region from the first fusing of the femoral condyles to the lowest part of the growth zone.

Tissue extraction

After euthanasia, both hindlimbs and the L4 vertebra were removed. The right rectus femoris muscle was isolated and weighed. The femora of both hindlimbs were isolated, and the length of the right femur was measured from the head of the femur to the medial condyle using a digital sliding caliper. The right rectus femoris muscle and the left femur were immersion fixed in 0.1 M sodium phosphate-buffered formaldehyde (4% formaldehyde, pH 7.0) for 48 h and then stored in 70% ethanol. The right femur, right tibia, and L4 vertebra were stored in Ringer’s solution at −20 °C.

Rectus femoris muscle

The fixed rectus femoris muscle was dehydrated and embedded in plastic-based 2-hydroxyethyl methacrylate (Technovit 7100, Kulzer GmbH, Hanau, Germany), cut into 6-µm-thick cross-sectional sections on a microtome (Jung RM2065; Leica Instruments, Nussloch, Germany), and stained with Masson’s trichrome. The muscle sections were examined using a light microscope (Nikon Eclipse i80, Tokyo, Japan) connected to a computer equipped with Visiopharm (v. 2020. 09.0.8195, Visiopharm, Hørsholm, Denmark) at a final magnification of ×1027. The cross-sectional area (CSA) of the muscle cells was estimated using the nucleator principle¹³. On average, 222 and 236 muscle cells were counted per sample in C57 and SWISS mice, respectively.

Dual-energy X-ray absorptiometry (DEXA)

Femoral bone mineral content (BMC) and areal bone mineral density (aBMD) of the whole femur were determined using DEXA (pDEXA Sabre XL, Stratec, Birkenfeld, Germany) with a spatial resolution of 0.1 mm x 0.1 mm and a scan speed of 3 mm/s. The coefficient of variation (CV) for BMD in our laboratory is 6.2%.

Micro computed tomography

The right femur and L4 vertebra were placed in a desktop µCT (Scanco µCT 35, Scanco Medical, Brüttisellen, Switzerland) and imaged with 1,000 projections/180°, an X-ray tube voltage of 55 kVp, and an integration time of 800 ms. For the C57 mice, the femoral mid-diaphysis and the distal femoral metaphysis and epiphysis were imaged with an X-ray current of 145 µA, while L4 vertebra was imaged with an X-ray current of 72 µA. For the SWISS mice, all imaging were performed with an X-ray current of 72 µA. For both C57 and SWISS mice, the femoral mid-diaphysis was imaged with an isotropic voxel size of 7 μm, while the distal femoral metaphysis, the distal femoral epiphysis, and L4 vertebra were imaged with an isotropic voxel size of 3.5 μm.

The femoral mid-diaphysis was analyzed using a VOI that comprised an 820-µm-high region centered at the midpoint of the femur. The distal femoral metaphysis was analyzed using a VOI that comprised a 1,000-µm-high region, starting 200 μm proximal to the top of the growth zone to exclude primary spongiosa. The distal femoral epiphysis was analyzed using a VOI that started where the two condyles fused and ended where the growth zone first appeared. The L4 vertebral body was analyzed using a VOI that extended from the upper growth zone to the lower growth zone excluding primary spongiosa. The VOI of the femoral mid-diaphysis comprised cortical bone only, while the VOIs of the distal femoral metaphysis, epiphysis, and L4 vertebral body comprised trabecular bone only. Representation of the delineated VOIs can be seen in Fig. 1. The 3D data from the C57 mice were low pass filtered with a Gaussian filter (σ = 0.8 and support = 1) and segmented with a threshold of 426, 426, 553, and 553 mg hydroxyapatite (HA)/cm³ for the femoral mid-diaphysis, the distal femoral metaphysis, the distal femoral epiphysis, and the L4 vertebral body, respectively. The 3D data from the SWISS mice were low pass filtered with a Gaussian filter (σ = 1.3 and support = 2) and segmented with a threshold of 576, 576, 576, and 637 HA/cm³ for the femoral mid-diaphysis, the distal femoral metaphysis, L4 vertebral body, and the distal femoral epiphysis, respectively.

The cortical mid-diaphyseal microstructural analyses included bone area (B.Ar), marrow area (Ma.Ar), tissue area (T.Ar), cortical thickness (Ct.Th) and tissue mineral density (TMD). The trabecular microstructural analyses included bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), trabecular separation (Tb.Sp), structure model index (SMI), volumetric BMD (vBMD), and TMD.

Mechanical testing

The mechanical properties of the right femoral mid-diaphysis, femoral neck, and L4 vertebral body were determined using a materials testing machine (Instron 5566, Norwood, MA, USA). The femoral mid-diaphysis underwent a 3-point bending test, where the femur was placed on two rounded bars separated by 7.1 mm, with the anterior side of the femur facing upwards. Vertical load was applied to the mid-point of the femur with a third rounded bar at a constant deformation rate of 2 mm/s until fracture. Then, the proximal half of the femur was fixed in a custom-made testing jig, as previously described¹⁴. Vertical load was applied on the femoral head at a constant deformation rate of 2 mm/s until fracture of the femoral neck. The vertebral arch, along with the intervertebral discs and cartilage, was removed from the L4 vertebra. Vertical load was then applied to the vertebral body specimen at a constant deformation rate of 2 mm/s until fracture, as previously described¹⁵.

Dynamic bone histomorphometry

Cortical bone: A 200-µm-thick cross-sectional section was cut from the femoral mid-diaphysis and mounted unstained on glass slides. Fluorescent bone labels were analyzed using the microscope equipped with fluorescence at a final magnification of x1027. A 24-arm circular radiating grid was superimposed on the screen, and the intersections between the grid and bone or labels were counted.

Trabecular bone: The distal femoral metaphysis was embedded undecalcified in methacrylate, and 7-µm-thick longitudinal sections were cut using the microtome and mounted unstained on microscope slides. Using Visiopharm, a 1,000-µm-high region of interest (ROI) comprising trabecular bone only was delineated along the endocortical bone surface, starting 200 μm proximal to the growth place to avoid primary spongiosa. The fields of view were sampled to cover 100% of the ROI, and the labels were quantified using a randomly rotated, software-generated grid containing 10 parallel lines that covered each field of view.

Assessment of the formative bone parameters: mineralizing bone surfaces (MS/BS), mineral apposition rate (MAR), and bone formation rate (BFR/BS), was conducted in accordance with the guidelines established by the ASBMR Histomorphometry Nomenclature Committee¹⁶. However, in samples where no double labels were visible, an imputed value of zero was used to calculate the MAR if there were detectable double labels at other bone sites for the same animal.

Statistics

Data analysis was conducted using GraphPad Prism (v. 10.1.1, GraphPad Software, Boston, MA, USA). Normality was assessed graphically using a Q-Q plot. For normally distributed data, a one-way analysis of variation (ANOVA) was conducted, followed by a Holm-Sidak multiple comparisons test to identify differences between individual groups. For non-normally distributed data, a Kruskal-Wallis test was used, followed by Dunn’s multiple comparisons test.

Change in lower hindlimb muscle volume during the study were analyzed using multiple paired t-tests followed by Holm-Sidak multiple comparison tests.

Data are presented as mean ± standard deviation (SD). Results were considered statistically significant at p < 0.05. For statistically significant results, the exact p-value is reported unless p < 0.0001. When all MP doses yield statistically significant outcomes, p-values are presented in the following order: MP 15 mg/kg/day, MP 22.5 mg/kg/day, and MP 30 mg/kg/day.

Results

Body weight

C57 mice: The body weight (BW) was the same among all the groups of mice at day zero. However, from week one to week four, the BW was significantly lower in all MP-injected mice (p = 0.0003, p = 0.0003, p = 0.0004, respectively) compared to control mice (Fig. 2a). The most pronounced BW loss in MP-injected mice occurred during the first week, after which the BW stabilized, while the BW of control mice increased steadily throughout the study.

Fig. 2.

Temporal development of body weight of (a) C57BL6/JRj (C57) mice and (b) RjOrl: SWISS (SWISS) mice. Mean ± SD. *p ≤ 0.05 vs. Control.

SWISS mice: There was no difference in BW between MP-injected mice and control mice at any point during the study (Fig. 2b).

Muscle volume

C57 mice: In all MP-injected mice, the muscle volume of the right lower hindlimb at euthanasia was significantly reduced (p = 0.0001, p < 0.0001, p < 0.0001, respectively) compared to the muscle volume at day zero (Fig. 3a). In contrast, control mice showed a significant increase in muscle volume at euthanasia (p = 0.011) compared to day zero. Additionally, the muscle volume at euthanasia was significantly lower in all MP-injected mice (p < 0.0001, p < 0.0001, p < 0.0001, respectively) compared to control mice (Supplementary Fig. S1).

Fig. 3.

Skeletal muscle data. (a-b) Lower hindlimb muscle volume of C57BL/6JRJ (C57) mice and RjOrl: SWISS (SWISS) mice at day zero and at euthanasia. (c) Rectus femoris muscle mass of C57 mice and SWISS mice. (d) Rectus femoris muscle cell cross-sectional area (CSA) of C57 mice and SWISS mice. (e) Cross-sectional image of the rectus femoris muscle from a representative SWISS control mouse. (f) Cross-sectional image of the rectus femoris muscle from a representative SWISS mouse injected with methylprednisolone (30 mg/kg/day). Mean ± SD. # p ≤ 0.05 vs. muscle volume at day zero; ## p ≤ 0.01 vs. muscle volume at day zero; ### p ≤ 0.001 vs. muscle volume at day zero; *p ≤ 0.05 vs. Control; **p ≤ 0.01 vs. Control; ***p ≤ 0.001 vs. Control.

SWISS mice: In all MP-injected mice, the muscle volume at euthanasia was significantly reduced (p = 0.002, p = 0.0009, p = 0.0046, respectively) compared to the muscle volume at day zero (Fig. 3b). Furthermore, the muscle volume at euthanasia was significantly lower in all MP-injected mice (p = 0.0004, p = 0.0028, p = 0.0004, respectively) compared to the control mice (Supplementary Fig. S1).

Rectus femoris muscle

C57 mice: Neither the rectus femoris muscle mass nor the rectus femoris muscle cell CSA differed between MP-injected mice and control mice (Fig. 3c and d).

SWISS mice: The rectus femoris muscle mass did not differ between MP-injected mice and control mice. However, the rectus femoris muscle cell CSA was significantly reduced in all MP-injected mice (p = 0.0041, p = 0.0198, p = 0.006, respectively) compared to the control mice (Fig. 3d-f).

Femoral BMD, BMC, and length

C57 mice: There were no differences in femoral length, aBMD, or BMC in MP-injected mice compared to control mice.

SWISS mice: Femoral aBMD displayed a borderline significant reduction in mice injected with MP at 15 mg/kg/day (p = 0.0655) and significant reductions in mice injected with MP at 22.5 mg/kg/day (p = 0.0376) and MP 30 mg/kg/day (p = 0.0111) compared to control mice (Fig. 4a). Neither femoral BMC nor femoral length differed in MP-injected mice compared to control mice.

Fig. 4.

Areal bone mineral density (aBMD) and cortical bone parameters. (a) aBMD of C57BL/6JRj (C57) mice and RjOrl: SWISS (SWISS) mice. (b) Cortical thickness (Ct.Th) of the femoral mid-diaphysis of C57 mice and SWISS mice. (c) Representative cross section of the femoral mid-diaphysis of a C57 control mouse. (d) Representative cross section of the femoral mid-diaphysis of a C57 mouse injected with methylprednisolone (30 mg/kg/day). Mean ± SD. *p ≤ 0.05 vs. Control. **p ≤ 0.01 vs. Control. ***p ≤ 0.001 vs. Control.

Cortical bone microstructure at the femoral mid-diaphysis

C57 mice: Ct.Th was significantly reduced in all MP-injected mice (p < 0.0001, p = 0.0002, p = 0.0002, respectively) compared to control mice (Fig. 4b-d), while the other cortical microstructural parameters did not differ between MP-injected and control mice.

SWISS mice: None of the analyzed microstructural parameters of the cortical bone differed between MP-injected mice and control mice.

Trabecular bone microstructure at the distal femoral metaphysis

C57 mice: BV/TV did not differ between MP-injected mice and control mice. However, in mice injected with MP at 15 mg/kg/day, Tb.N was significantly increased (p = 0.0006), and Tb.Sp was significantly reduced (p = 0.0006) compared to control mice (Supplementary Table S1). Additionally, SMI was significantly reduced in all MP-injected mice (p = 0.0013, p = 0.0001, p = 0.0051, respectively) compared to control mice. This suggests that the trabecular bone network in MP-injected mice gradually became more plate-like than that of control mice.

SWISS mice: BV/TV did not differ between MP-injected mice and control mice. However, in mice injected with MP at 30 mg/kg/day, Tb.Th was reduced (p = 0.0246) compared to control mice. Metaphyseal BV/TV, Tb.Th. and SMI for C57 and SWISS mice can be seen in Fig. 5a-c.

Fig. 5.

Trabecular bone parameters. (a-c) Microstructural parameters of the distal femoral metaphysis. (d-f) Microstructural parameters of the L4 vertebral body. BV/TV = Bone Volume Fraction, Tb.Th = Trabecular Thickness, SMI = Structure Model Index. Mean ± SD.: *p ≤ 0.05 vs. Control. **p ≤ 0.01 vs. Control. ***p ≤ 0.001 vs. Control.

Trabecular bone microstructure at the distal femoral epiphysis

C57 mice: BV/TV did not differ between MP-injected mice and control mice. However, Tb.N showed a borderline significant increase in mice injected with MP at 15 mg/kg/day (p = 0.052) and MP 22.5 mg/kg/day (p = 0.058). In contrast, Tb.Sp was significantly reduced in all MP-injected mice (p = 0.0084, p = 0.0263, p = 0.0455, respectively) compared to control mice (Supplementary Table S1).

SWISS mice: None of the microstructural parameters differed between MP-injected mice and control mice.

Trabecular bone microstructure at L4 vertebral body

C57 mice: Surprisingly, in all MP-injected mice, BV/TV (p = 0.0003, p = 0.0022, p = 0.0105, respectively) and vBMD (p = 0.0003, p = 0.0007, p = 0.0122, respectively) were significantly increased compared to control mice (Fig. 5d and Supplementary Table S1). Similarly, Tb.N was significantly increased in mice injected with MP at 15 mg/kg/day (p = 0.0188) and MP 22.5 mg/kg/day (p = 0.0323) compared to control mice, accompanied by a corresponding significant reduction in Tb.Sp in mice injected with MP at 15 mg/kg/day (p = 0.0223 ) and MP 22.5 mg/kg/day (p = 0.0293). Tb.Th., however, did not differ between MP-injected mice and control mice. Additionally, in all MP-injected mice, SMI was significantly reduced (p = 0.0002, p = 0.0016, p = 0.0028, respectively) compared to control mice. TMD was slightly but significantly increased in all MP-injected mice (p = 0.0004, p = 0.0001, p = 0.0004, respectively) compared to control mice. Together, these findings suggest that the bone tissue in the MP-injected mice was more plate-like and more mineralized than in the control mice. Vertebral BV/TV, Tb.Th and SMI for C57 and SWISS mice can be seen in Fig. 5d-f.

SWISS mice: None of the microstructural parameters differed between MP-injected mice and control mice.

Mechanical testing

The mechanical properties of the femoral mid-diaphysis, the femoral neck, and the L4 vertebral body did not differ between MP-injected mice and control mice in either C57 mice or SWISS mice (Fig. 6).

Fig. 6.

Maximum bone strength (F_max) at the (a) femoral mid-diaphysis, (b) femoral neck, and (c) L4 vertebral body. Mean ± SD.

Dynamical bone histomorphometry of cortical bone

C57 mice: At the periosteal bone surface, MS/BS displayed a borderline significant reduction in mice injected with MP at 15 mg/kg/day (p = 0.0545) and significant reductions in mice injected with MP at 22.5 mg/kg/day (p = 0.0108) and MP 30 mg/kg/day (p = 0.0351). Additionally, in all MP-injected mice, MAR (p = 0.0032, p = 0.0032, p = 0.0037, respectively) and BFR/BS (p = 0.0032, p = 0.0010, p = 0.0032, respectively) were significantly reduced compared to control mice. Periosteal MS/BS, MAR, and BFR/BS for C57 and SWISS mice can be seen in Fig. 7a-c. At the endocortical bone surface, only BFR/BS was significantly reduced in mice injected with MP at 22.5 mg/kg/day (p = 0.0145) compared to control mice. Endocortical MS/BS, MAR and BFR/BS for C57 and SWISS mice can be seen in Fig. 7d-f.

Fig. 7.

Dynamic histomorphometry (a-c) cortical bone parameters of the femoral mid-diaphysis at the periosteal surface, (d-f) cortical bone parameters of the femoral mid-diaphysis at the endocortical surface, and (g-i) trabecular bone parameters at the distal femoral metaphysis. MS/BS = Mineralizing Bone Surfaces, MAR = Mineral Apposition Rate, BFR/BS = Bone Formation Rate. Mean ± SD.: *p ≤ 0.05 vs. Control. **p ≤ 0.01 vs. Control. ***p ≤ 0.001 vs. Control.

SWISS mice: At the periosteal bone surface, significant reductions in MS/BS and BFR/BS were observed only in mice injected with MP at 22.5 mg/kg/day (p = 0.0014 and p = 0.0242, respectively) compared to control mice. At the endocortical bone surface, the formative bone parameters did not differ between MP-injected mice and control mice.

Dynamical bone histomorphometry of trabecular bone at the distal femoral metaphysis

C57 mice: None of the formative bone parameters differed between MP-injected mice and control mice.

SWISS mice: MAR was significantly reduced in all MP-injected mice (p = 0.0111, p = 0.0023, p = 0.0034, respectively) compared to control mice, while the other formative bone parameters did not differ between MP-injected mice and control mice. Trabecular MS/BS, MAR, and BFR/BS for C57 and SWISS mice can be seen in Fig. 7g-i.

Discussion

The present study aimed to investigate and compare the effects of high doses of MP on the musculoskeletal system in C57 and SWISS mice, with the goal of developing a robust small-animal model of skeletal muscle atrophy and GIO. The primary findings revealed that high doses of MP resulted in modest but distinct reductions in skeletal muscle volume in both C57 and SWISS mice. In contrast, MP induced only minor and varying alterations in bone density, microstructure, and strength across both strains, failing to meet all three outcome measures suggested by Xavier et al. for an appropriate model of GIO⁹. Additionally, MP did not demonstrate superior efficacy in inducing osteopenia in SWISS mice compared to C57 mice as previously suggested by Ersek et al.¹².

In both strains, there was a modest reduction in skeletal muscle volume in MP-injected mice. However, the wet mass of the rectus femoris muscle did not differ from that of control mice in either mouse strain, despite a significant reduction in muscle cell CSA in SWISS mice. This discrepancy is likely due to the less precise nature of dissecting and removing a small muscle from a mouse, compared to the more accurate in-vivo µCT imaging of the same muscle in situ. This notion is supported by the observation that the coefficient of variation (CV) in the wet muscle mass (CV = 8.41%) was twice as large as that observed in the in situ determined muscle volume (CV = 4.29%).

Nonetheless, we expected a more pronounced effect of MP on the muscle volume as a previous study of ours demonstrated that implanting slow-release GC pellets (equivalent to 5.4 mg/kg/day of prednisolone) in SWISS mice resulted in substantial reductions of rectus femoris muscle mass and rectus femoris muscle cell CSA after four weeks¹⁷. This discrepancy in skeletal muscle effects may be due to implanted GC pellets, which, despite being designed for gradual release over 90 days, often discharge their full dose at once. This results in a sharp spike in plasma GC levels that quickly returns to baseline¹⁸. Such an initial surge of GC may exert a stronger impact on muscle tissue than the more continuous GC exposure seen in the present study.

In both strains, the relatively high doses of MP only precipitated modest effects on the bone tissue. In C57 mice, MP affected femoral cortical bone formation, reducing mineralizing bone surfaces, periosteal mineral apposition rate, and bone formation rate, along with a slight reduction in femoral cortical thickness. However, these alterations in the femoral cortical bone parameters did not translate into reductions of femoral BMD or bone strength. In contrast, MP increased the bone volume fraction and reduced trabecular separation of the L4 vertebral body, although this also did not manifest in an altered bone strength. These findings are consistent with those of previous studies of prednisolone and dexamethasone in C57BL/6 mice, which reported reductions in cortical bone parameters and increases in trabecular microstructural parameters^8,19.

In SWISS mice, MP mainly affected the trabecular bone formation of the distal femoral metaphysis, reducing trabecular mineral apposition rate and trabecular thickness, which did not translate into reductions in bone strength. However, bone mineral density of the femur was reduced at the highest MP doses. Taken together, these results suggest that trabecular deterioration may have begun but not yet fully materialized in the SWISS mice. Ersek et al. examined the effects of GC on bone tissue in CD-1 mice – an outbred strain derived from Swiss Webster mice²⁰ – and found that GC markedly reduced both trabecular bone volume fraction and bone strength¹². They used mice of similar age to those in the present study, but their study lasted eight weeks, twice as long as the present study. Therefore, it is possible that a longer study duration could have resulted in a more pronounced bone loss in SWISS mice.

MP was selected as the glucocorticoid based on prior unpublished pilot studies comparing different glucocorticoid types. The dosing regimen was determined from these pilot studies and relevant literature. In the reviews by Xavier et al.⁹ and Wood et al.¹⁰, glucocorticoids were administered at MP-equivalent doses of ≤ 5 mg/kg/day, corresponding to approximately 0.4 mg/kg/day in humans according to the FDA’s interspecies conversion factor²¹. This falls within the EULAR-recommended medium-to-high dose range for rheumatic diseases (0.09–1.1 mg/kg/day MP equivalent, depending on disease and severity)²². However, possibly due to physiological differences between humans and rodents, we hypothesized that higher doses were required to establish an effective GIO model, and thus the highest dose used in the present study was equivalent of up to MP 2.4 mg/kg/day in humans.

The s.c. administration route was preferred over implanted micro-osmotic slow-release pumps and slow-release pellets due to the questionable effectiveness of these methods in maintaining consistently elevated plasma GC levels¹⁸. Furthermore, in an unpublished study, where MP was administered in the drinking water, we found a strong positive correlation between water intake and MP dose, making this route of GC administration difficult to control.

The administration of GC via s.c. injections also has its limitations. Firstly, the initial peak of plasma GC levels quickly returns to baseline after just a few hours¹⁸. Secondly, repeated injections can stress the mice, potentially causing an initial surge in endogenous glucocorticoid release. This stress is expected to diminish over time due to the inhibition of the hypothalamic-pituitary-adrenal axis caused by the administration of high doses of exogenous GC²³. To minimize bias from the injections, control mice also received vehicle injections to standardize endogenous stress levels across all groups, and tunnel handling was employed to minimize handling-induced stress²⁴.

A study from Herrmann et al. suggested that the most effective method for maintaining constantly elevated plasma GC levels is s.c. implantation of slow-release GC pellets, which are replaced weekly to compensate for the previously mentioned initial burst-like GC release¹⁸. If confirmed, this approach could be utilized in future studies.

The current study has some limitations. Due to an aging X-ray tube, skeletal sites were imaged at two different X-ray currents; however, we do not believe this has had any effects on the results. Despite careful subcutaneous administration of GC, minor interindividual variations in systemic exposure cannot be excluded. To ensure uniform exposure, GC levels could be verified through serum measurements of methylprednisolone in future studies. Furthermore, we did not account for the potential interplay between MP and the underlying clinical condition it is intended to treat. Although glucocorticoid treatment is often aimed at suppressing inflammation, which itself can contribute to bone loss²⁵, it is possible that the complex interactions between inflammatory signaling and glucocorticoid-mediated pathways may exacerbate the bone loss in a clinical context²⁶.

In conclusion, the present study demonstrates that administering high doses of MP via s.c. injections in C57 and SWISS mice serve as a viable model for inducing skeletal muscle atrophy in both strains. However, the results do not unequivocally support the use of this method as model of GIO in either strain. Moreover, although previous studies have suggested that SWISS mice are more susceptible to GIO than C57 mice, our findings indicate that MP was not more effective at inducing osteopenia in SWISS mice compared to C57 mice. Therefore, further research is needed to develop a robust model of GIO in mice, with particular emphasis on the impact of study duration on the effects of glucocorticoid levels in these models.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Oliver Lassen Slavensky: concept and design, data collection, statistical analyses and manuscript writing. Annemarie Brüel: concept and design, critical revision, approval of final manuscript. Jesper Skovhus Thomsen: concept and design, critical revision, supervision, approval of final manuscript. Frederik Duch Bromer: concept and design, critical revision, statistical analyses, approval of final manuscript.

Funding

The study was kindly supported by the Department of Biomedicine, Health, Aarhus University and the Riisfort Foundation. The in-vivo µCT was acquired through a Novo Nordisk Foundation research infrastructure grant (grant NNF19OC0055801). The desktop µCT was donated by the VELUX Foundation.

Data availability

The datasets analyzed in this study can be requested from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.