Loss of glucocorticoid rhythm induces an osteoporotic phenotype in female mice

Abstract

Glucocorticoid (GC)-induced osteoporosis is a widespread health problem that is accompanied with increased fracture risk. Detrimental effects of anti-inflammatory GC therapy on bone have been ascribed to the excess in GC exposure, but it is unknown whether there is also a role for disruption of the endogenous GC rhythm that is inherent to GC therapy. To investigate this, we implanted female C57Bl/6J mice with slow-release corticosterone (CORT) pellets to blunt the rhythm in CORT levels without inducing hypercortisolism. Flattening of CORT rhythm reduced cortical and trabecular bone volume and thickness, whilst bone structure was maintained in mice injected with supraphysiologic CORT at the time of their endogenous GC peak. Mechanistically, mice with a flattened CORT rhythm showed disrupted circadian gene expression patterns in bone, along with changes in circulating bone turnover markers indicative of a negative balance in bone remodelling. Indeed, double calcein labelling of bone in vivo revealed a reduced bone formation in mice with a flattened CORT rhythm. Collectively, these perturbations in bone turnover and structure decreased bone strength and stiffness, as determined by mechanical testing. In conclusion, we demonstrate for the first time that flattening of the GC rhythm disrupts the circadian clock in bone and results in an osteoporotic phenotype in mice. Our findings indicate that at least part of the fracture risk associated with GC therapy may be the consequence of a disturbed GC rhythm, rather than excess GC exposure alone, and that a dampened GC rhythm may contribute to the age-related risk of osteoporosis.

Keywords: bone health; circadian rhythm; corticosteroids; fracture risk; osteoporosis.

FIGURE 1

Corticosterone pellets flatten the rhythm in plasma corticosterone and reduce lean body mass. (a) Plasma corticosterone (CORT) levels were measured in mice 1 week after implantation of a vehicle (n = 10), 4.5 mg CORT (n = 15) or 7.5 mg CORT pellet (n = 10), at regular intervals throughout the day and night. An area under the curve (AUC) of all individual plasma corticosterone (CORT) measurements was calculated to determine total CORT exposure (ng/ml*h). For each mouse, a CORT amplitude was calculated by subtracting the lowest value in plasma CORT from the peak value, and dividing this by 2. (b) After 5 weeks of pellet implantation, morning and evening plasma CORT levels were measured and used to calculate the amplitude in CORT rhythm. (c) At endpoint, after 7 weeks of pellet implantation, body weight, fat mass and lean mass were measured. Data represents means ±SEM, including individual data points. Time is denoted as Zeitgeber Time (ZT), where ZT0 = lights on and ZT12 = lights off. *p < 0.05, ***p < 0.001 compared to the vehicle control group, according to one‐way ANOVA with Dunnett's post hoc test. NS, non‐significant

FIGURE 2

Corticosterone pellets reduce both cortical and trabecular bone volume. Micro‐CT analyses were performed after 7 weeks of pellet intervention. (a) Representative structural images of the cortical bone in all groups, indicating cortical thickness in mm by colour codes. (B)–(D) Micro‐CT analysis was used to assess cortical bone volume (Ct.BV; b), cortical thickness (Ct. Th; c) and cortical bone mineral density (Ct.BMD; d). (e) Representative structural images of the trabecular bone in all groups. (f)–(h) Micro‐CT analysis of the relative trabecular bone volume (BV/TV; f), trabecular number (Tb.N; g) and trabecular thickness (Tb. Th; h). Data represents means ±SEM, including individual data points. *p < 0.05, ***p < 0.001 compared to the vehicle control group, according to one‐way ANOVA with Dunnett's post hoc test

FIGURE 3

Hypercortisolism with an intact corticosterone rhythm preserves lean mass and cortical bone structure. (a) Plasma corticosterone (CORT) levels were measured in mice 3 days after administration of vehicle (n = 10) or 3 mg/kg CORT via daily injection at 17 h (n = 10), at regular intervals throughout the day and night (n = 5 per group per time point). An area under the curve (AUC) of individual plasma corticosterone (CORT) measurements was calculated to determine total CORT exposure. (b)–(d) At endpoint, after 7 weeks, total body weight (b), fat mass (c) and lean mass (d) were measured. (e)–(g) Micro‐CT analysis was used to assess cortical bone volume (Ct.BV; e), cortical thickness (Ct. Th; f), and cortical bone mineral density (Ct.BMD; g). Data is expressed as means ±SEM, including individual data points. Time is denoted as Zeitgeber Time (ZT), where ZT0 = lights on and ZT12 = lights off. *p < 0.05 compared to the control group, according to unpaired t‐test

FIGURE 4

Corticosterone pellets disrupt rhythmic expression of bone‐related genes, glucocorticoid‐response genes, and clock genes. Mice were sacrificed at Zeitgeber time (ZT) 0, ZT6, ZT12, and ZT18 after 2 weeks of pellet intervention to assess circadian rhythmicity of bone‐related genes (Opg, Runx2, Blap, Col1a1, cFos, Sost, Rank, Rankl, Trap, Ctsk, Nfatc), glucocorticoid‐response genes (Fkbp5, Gilz, Sgk1, Mt2a) and clock genes (Per1, Per2, Bmal1, Clock, Reverbα) in tibiae (n = 6/group/timepoint). Data points were log‐transformed and double‐plotted from the dotted lines. Light and grey areas represent the light and dark phase, respectively. Rhythm analyses were performed by fitting a sine wave to the data and evaluating the confidence interval (CI) of the amplitude, as listed in Table S1. Significant amplitudes are shown in the figure. Amplitudes that differ significantly between the two groups are indicated with an asterisk. Data represent means ±SEM. NS, non‐significant or non‐detectable

FIGURE 5

Corticosterone pellets modulate bone turnover markers and reduce bone formation. (a) Plasma levels of tartrate‐resistant acidic phosphatase (TRAP) were evaluated in the morning and evening after 5 weeks of pellet intervention. (b), (c) At endpoint, after 7 weeks of intervention, the trabecular osteoclast surface area (b) was determined from femurs stained for TRAP and counterstained with Light Green (c). (d) Plasma levels of procollagen type 1 amino‐terminal propeptide (P1NP; b) were evaluated in the morning and evening after 5 weeks of pellet intervention. (e, f) After 7 weeks, the trabecular osteoblast surface area (e) was determined from femurs stained for osteocalcin and counterstained with haematoxylin (f). (g) Representative images of femoral trabecular bones double labelled with calcein (green), and counterstained with calcein blue (blue). (h)–(j) Calcein labelling was used to determine the MAR (h), mineralizing surface per bone surface (MS/BS; i) and bone formation rate per bone surface (BFR/BS; j). Data represent means ±SEM, including individual data points. Time is denoted as Zeitgeber Time (ZT), where ZT0 = lights on and ZT12 = lights off. *p < 0.05, **p < 0.01, ***p < 0.001 compared to the vehicle control group, according to two‐way ANOVA (a, d) or one‐way ANOVA with Dunnett's post hoc test (b, e, h–j)

FIGURE 6

Corticosterone pellets reduce bone strength and stiffness. (a)–(c) Polar moment of inertia (polar MOI; a) as well as two breaking strength parameters, Imin/Cmin (b) and Imax/Cmax (c), were calculated from micro‐CT data. (d) Three‐point bending tests of tibiae were performed to determine the maximum load (Fmax). (e), (f) The relationship between Fmax and femoral cortical bone volume (e) and cortical thickness (f) was evaluated by Pearson correlation analysis. (g) Bone stiffness was evaluated through three‐point bending tests. (h), (i) Bone stiffness was correlated to femoral cortical bone volume (h) and cortical thickness (i). Data represent means ±SEM, including individual data points. **p < 0.01 compared to the vehicle control group, according to one‐way ANOVA with Dunnett's post hoc test