Figure 1.
Experimental condition and fluorescent image of tibial sections.
(A) Schematic view of 3-point-bending method. The lateral surface of tibia is supported by 2 rubber-padded points set 9 mm apart. The load cell applies the medial surface of tibia in the point set 3.6 mm distal from the proximal supporting point. In the experiment, a static pre-load is applied to the tibia. This load was set to be approximately 1 newton (N). (B, C) Designed wave forms of the loads applied to the mouse left tibia. (B) The short-pulse mode consists of a constant load (0.5 s) and a rest period (39.5 s). (C) The long-pulse mode consists of a much longer period of constant load (39.5 s) and a much shorter rest period (0.5 s). A bout of the mechanical loading comprises 36 cycles lasting a total of 24 minutes. No loads are applied to the right tibia. The magnitude of each load is approximately 6 N. In the experiment, a static pre-load is applied to the tibia. This load is set to be approximately 1 N. (D) Time schedule and fluorescent image of tibial sections (a) The experimental design for the loading and calcein labeling. Loading was performed 3 times (red arrows). Calcein labeling was carried out 4 times (green triangles). (b–e) Calcein fluorescent images of tibial sections (green colored) were examined 14 days after loading. (b, d) In the right tibia, used as a control, new bone formation was not observed in either short-pulse or long-pulse group. (c) In the left tibia in the short-pulse group, bone formation was prominent on the periosteal surface around the loading point (white arrow). (e) In the left tibia in the long-pulse group, bone formation was significant on the periosteal surface at the side opposite to the loading point (white asterisks), but not around the loading point. White arrowheads indicate loading direction.
Figure 2.
Stress induced in the long-pulse mode induced cell proliferation in the periosteum.
Hematoxylin and eosin (H&E)-stained histological sections of the side opposite to the loading point in the long-pulse group at days 1, 3, 5, and 7 after loading to the left tibia. In the right tibia, used as the control, the periosteum consisted of a thin layer (A). In the left tibia treated in the long-pulse mode, periosteal hypertrophy (purple colored nuclei) first occurred at day 1 (B), and became prominent at day 3 (C). The woven bone (pale pink color) first appeared at day 5 (D) and became prominent at day 7 (E). CB, cortical bone; PL, periosteal cell layer; WB, woven bone; Scale bar, 50 µm. A representative result from 4 individual experiments is shown.
Figure 3.
Induction of hypertrophic periosteum by the long-pulse mode.
(A) Immunological staining after treatment in the long-pulse mode. Histological sections around the side opposite to the loading point of the long-pulse treated tibiae were prepared at days 1, 3, and 7. These sections were stained with anti-Ki-67 antibodies (a–c) and anti-periostin antibodies (g–i), and ALP activity was also assessed (d–f). In the right tibia, taken as the control, Ki-67-positive cells were rarely detected (a), low ALP activity was observed (d), and restricted periostin signals were observed (g). In the left tibia treated in the long-pulse mode, Ki-67-positive cells were detected at days 3 and 7 (b, c, arrows), and high ALP activity was detected in the osteogenic layer, but not in the fibrous layer, at days 3 and 7 (e, f, arrows). Expression of periostin was detected throughout the hypertrophic periosteum at days 3 and 7 (h, i). CB, cortical bone; POL, periosteal osteogenic layer; PFL, periosteal fibrous layer; WB, woven bone. Scale bar, 50 µm. A representative result from 4 independent experiments is shown. (B) Increased BMP2 protein expression in hypertrophic periosteal cells. Lysates of periosteal cells around the side opposite to the loading point from the loaded left tibia in the long-pulse mode and the control right tibia at day 3 were separated by SDS-PAGE and blotted with anti-BMP-2 antibodies. The arrow indicates precursor BMP-2 bands, and the arrowhead indicates the mature BMP-2 band. The active mature BMP-2 was preferentially expressed in the loaded left tibia, whereas equivalent signals of the precursor BMP-2 were detected for both right and left tibiae. Moreover, in the cytochalasin-D-treated left tibia 3 days after the treatment as shown in Fig. 7, active mature BMP-2 was also significantly expressed.
Figure 4.
in-situ immunofluorescence imaging method.
(A) Schematic view of in-situ immunofluorescence imaging method for analyzing the periosteum. The calcein-labeled bone surface and immunostained perisoteum were examined by laser-scanning confocal microscope. (B) Immunostained periosteal cells were distinguished from the calcein-labeled bone surface. Periosteal cells were stained with Alexa Fluor 568-labeled phalloidin and TOPRO3 for detecting the actin cytoskeleton and nuclei, respectively; and a merged image is shown in the right column. (C) Osteocytes, embedded in the calcein-labeled bone, were stained with phalloidin for actin and TOPRO3 for nuclei; and a merged image is shown in the right column. (D) Periosteal cells were stained with anti-N-cadherin antibodies, and this photo was merged with that of TOPRO3-labeled nuclei. (E) Periosteal cells were stained with anti-periostin antibodies, and this photo was merged with that of TOPRO3-stained nuclei.
Figure 5.
Disorganization and remodeling of actin cytoskeleton in periosteal cells by mechanical loading.
Periosteal cells subjected to mechanical loading in the long-pulse mode were stained with Alexa Fluor 568-labeled phalloidin for actin cytoskeleton and TOPRO3 for nuclei one hour after onset of the mechanical loading. The right tibia was used as the unloaded control sample. On the bone surface of the right tibia, the actin cytoskeleton was detected as having the shape of stress fibers (right tibia, control). On the side opposite to loading point of the left tibia treated in the long-pulse mode, the signals of actin stress fibers were decreased in intensity, and these fibers were disorganized (left tibia, 1 hr). Following the first observation after 1 hr in the long-pulse mode, we continuously manipulated this mode to left tibia for 4 days. Periosteal cells in the left tibia subjected to mechanical loading in the long-pulse mode were stained similarly at day 1, day 2 and day 4 with Alexa Fluor 568-labeled phalloidin for actin cytoskeleton and TOPRO3 for nuclei, and both signals were merged. At day 1, actin stress fibers were rarely detected, and round-shaped cells were observed (red arrows). At day 2, actin stress fibers were partially restored (red arrows). Interestingly, nuclei with holes and abnormally shaped nuclei were detected at days 1 and 2 (red asterisks). Disorganized actin stress fibers and abnormally shaped nuclei disappeared at day 4. Scale bar, 20 µm. For all groups, n = 6.
Figure 6.
Induction of periosteal hypertrophy by treatment of cytochalasin-D.
Periosteal hypertrophy was induced by injection of cytochalasin-D into the periosteum. (A) Periosteal hypertrophy was revealed by observing three-dimensional reconstructed immunofluorescent images. In the control right tibia (right), the calcein-labeled layer (calcein) and a TOPRO3-labeled thin periosteal cell layer (nuclei) were detected. In the 200 µM cytochalasin-D-treated left tibia (left, cyto-D), the calcein-labeled layer (calcein) was again seen, but the hypertrophic periosteal cell layer (nuclei) had become markedly hypertrophic. (B) Thickness of periosteum treated with cytochalasin-D. Thickness was measured from the calcein-labeled bone surface to the superficial aspect of the layer of TOPRO3-labeled nuclei. In a non-treated right tibia, the periosteal thickness was approximately 40 µm (right tibia). In the left tibia treated with DMSO in PBS, the periosteal thickness was about 100 µm (control). In the cytochalasin-D-treated left tibia, the periosteal thickness was increased to about 140 µm (100 µM) and 150 µm (200 µM), respectively. Periosteal thickness was measured by laser scanning confocal microscope (Olympus). CB; cortical bone. For all groups, n = 4. Error bars represent standard deviation. Significant difference (indicated by asterisk, paired t-test) are detected in periosteal thickness between the cytochalasin-D treated groups (100 µM+200 µM, n = 8) and the control (p<0.05).
Figure 7.
Suppression of periosteal hypertrophy by inhibition of remodeling of actin cytoskeleton by ROCK inhibitor.
During the mechanical loading to the left tibia in the long-pulse mode, a ROCK inhibitor, which blocks the actin polymerization in actin remodeling, was injected into mice via a tail vein. The thickness of periosteum was measured at day 3 post loading. In unloaded and non-injected right tibia, the periosteal thickness was about 40 µm. In the loaded and vehicle-injected left tibia, the periosteal thickness was approximately 120 µm (0 mg/kg). In the loaded and ROCK inhibitor-injected left tibia, the periosteal thickness was decreased in a dose-dependent manner (25 mg/kg, 50 mg/kg, 75 mg/kg). Periosteal thickness was measured by observation with a laser scanning confocal microscope (Olympus). For all groups, n = 5. Error bars represent standard deviation. *Significant deference versus non-injected control (0 mg/kg) (p<0.05).