Effects of eldecalcitol on cortical bone response to mechanical loading in rats

Animals

Six-month-old female Wistar rats (retired breeder; Shimizu Laboratory Supply, Kyoto,
Japan), with initial body weights ranging between 255 g and 355 g, were used in this
study. During the experimental period, water and commercially available food (CE-2;
CLEA Japan, Tokyo, Japan; calcium content 1.18 g/100 g, phosphorus content 1.09 g/100 g,
vitamin D3 content 250 IU/100 g) were given ad libitum. The duration of daily light exposure in the breeding room was 12 h (7:00 AM to 7:00 PM),
and the room temperature was maintained at 24 °C.

After a 7-day acclimation period, rats were randomized into four groups based on eldecalcitol
dose (n?=?10 per group), each with the same mean body weight, as follows: (1) vehicle
administration (VEH), (2) low dose (ED-L), (3) middle dose (ED-M), and (4) high dose
(ED-H). Rats were allowed normal cage activity between loading sessions.

Eldecalcitol administration

We prepared 0.025–0.1 ?g/ml solutions of eldecalcitol by dissolving in medium-chain
triglyceride. Eldecalcitol or vehicle (medium-chain triglyceride) was administered
orally via gastric lavage 3 days per week for 3 weeks. The rats received vehicle alone
(VEH) or eldecalcitol at doses of 0.025 ?g/kg (ED-L), 0.05 ?g/kg (ED-M), or 0.1 ?g/kg
(ED-H). After administration of eldecalcitol or vehicle, tibial mechanical loading
was performed on the same day.

In vivo external mechanical loading

In vivo mechanical loading involved load application using a four-point bending device (developed
and assembled in the Biomechanics Laboratory, Creighton University) 10], 11]. Each rat was anesthetized with ether, and its right lower leg was placed between
the pads of the device. The right tibia was loaded at 38 N for 36 cycles at 2 Hz,
3 days per week for 3 weeks, for a total of 9 days. The left tibia was not loaded.

The force applied during loading was monitored by a strain gauge attached to the lever
arm as previously reported 10]–13]. Before the experiment, the four-point bending device was calibrated with a load
cell that had been previously calibrated by application of forces ranging from 0 to
70 N, using a mechanical testing machine (MTS810; MTS, Minneapolis, MN, USA). The
actual applied load during in vivo four-point bending was calculated based on this calibration 10]–13].

Bone histology

Rats received calcein injections (6 mg/kg BW, i.p.) on experimental days 13 and 19.
On day 20, rats in all four groups were anesthetized with 50 mg/kg BW ketamine hydrochloride
and 10 mg/kg BW xylazine, and were sacrificed by exsanguination. Both loaded (right)
and non-loaded (left) tibiae were removed, placed in 10 % phosphate-buffered formalin
for 24 h, and then transferred to 70 % ethanol. The tibiae were cut into three pieces:
(a) the proximal 1 cm, (b) the distal 5 mm, and (c) the remaining central diaphysis.
Central regions were stained with Villanueva bone stain for 72 h 11], 14]. Specimens were dehydrated with increasing concentrations of ethanol and acetone
and then embedded in methyl methacrylate. The region of maximum bending was located
in the central diaphysis, 3–13 mm proximal to the tibio-fibular junction (TFJ) 10]–13]. Two cross-sections were prepared from the region of maximum bending, specifically
at 4 mm and 4.5 mm proximal to the TFJ. These cross-sections were then ground to a
thickness of 60 ?m and mounted on glass slides. Histomorphometric data were collected
from these two sections and mean values were calculated.

Calculation of in vivo strain

The in vivo strain was calculated using the moment of inertia of each central diaphyseal cross-section,
as previously reported 11]. The outline of the cortical bone on each slide was traced, and the moment of inertia
and section modulus for each cross-section were calculated using Bone Histomorphometry
Software (System Supply, Nagano, Japan). The peak compressive strain on the lateral
surface was calculated using beam-bending theory:

(1)

where ?c?=?calculated peak compressive strain on the lateral periosteal surface, M?=?bending
moment (N-m), E?=?longitudinal Young’s modulus (estimated as 29?×?10
⁹
 N/m
²
), I?=?moment of inertia, and C?=?distance from the centroid to the lateral surface.

The in vivo peak compressive strain (?p) at the lateral periosteal surface was then predicted
from ?c using the following formula:

(2)

Equation (2) was derived from the in vivo strain gauge measurement 10].

Histomorphometry

A camera connected to a personal computer was used to run Bone Histomorphometry Software
(System Supply Co. Ltd.). The standard nomenclature for bone histomorphometry variables
was used 15].

The total tissue area (TtT.Ar, mm
²
) and marrow area (Ma.Ar, mm
²
) were measured, and the difference between them was reported as the cortical area
(Ct.Ar, mm
²
). The woven bone contained irregular collagen bundles and a diffuse fluorochrome
label, which was identified under conventional polarized and ultraviolet light. The
Ct.Ar did not include the woven bone area (Wo.Ar).

For both the periosteum and endosteum, we measured the single-labeled perimeter (sL.Pm,
%), double-labeled perimeter (dL.Pm, %), and woven bone perimeter (Wo.Pm, %; defined
as the perimeter with overlying woven bone). The Wo.Pm was not included in calculations
of the sL.Pm or dL.Pm. The formation perimeter (F.Pm) was defined as (dL.Pm?+?Wo.Pm?+?sL.Pm/2)/B.Pm
14], 16], 17]. The mineral apposition rate (MAR; ?m/d) and surface-based bone formation rate (BFR;
?m
³
/?m
²
/d) were calculated. A minimum MAR value of 0.3 ?m/d was used in rats showing only
sL.Pm 18]. The BFR was calculated using the formula BFR?=?MAR?×?F.Pm 13], 14], 16].

Histomorphometric data were collected from the periosteal and endocortical surfaces
of the tibia. The tibial periosteal surface was subdivided into lateral and medial
surfaces in the same manner as in our previous studies (Fig. 1) 14], 17], 19], because the type of stress applied by the four-point bending device differed between
the two surfaces (compression vs. tension) 10].

Fig. 1. Three surfaces of the tibia for histomorphometric analysis. The tibial periosteal
surface was divided into lateral and medial regions as indicated for histomorphometric
analysis

Statistical analysis

Data were analyzed by repeated two-way analysis of variance (ANOVA) for the effects
of external loading (loaded and non-loaded) and eldecalcitol treatment, and their
interaction. Post hoc multiple comparisons among eldecalcitol groups were performed
using the Dunnett’s test. SPSS Statistics software (version 21, SPSS Inc., Chicago,
IL, USA) was used for the analyses and P??0.05 was considered to be statistically significant.

Animal ethics

Our procedures were approved by the Committee on Laboratory Animals, Faculty of Medicine,
Tottori University, Japan.