Degeneration of the osteocyte network in the C57BL/6 mouse model of aging

MicroCT shows changes in bone structure and geometry with aging

MicroCT analysis showed changes in femoral bone structure and geometry with aging. Reconstructed images of the whole bones and of cortical bone at the midshaft showed increased cortical diameter in aged compared to young femurs in both genders (Fig. 1C). Quantitative analysis at the midshaft showed no significant change in cortical bone volume/total volume (BV/TV), which is essentially a measure of cortical porosity, or cortical bone area (Fig. 1D, 1E). However, aging was associated with a significant increase in cortical bone perimeter of 28% in females and 23% in males (Fig. 1F) and a significant decrease in cortical bone thickness of 14% in females but not males (Fig. 1G). In aged males, there was considerable variation in cortical thickness around the bone perimeter, (Fig. 1C, second row, right panel), resulting in the difference in the average thickness being non-significant. Midshaft cortical bone parameters were not significantly different between genders.

Figure 1. Aging is associated with structural changes in bone that are more pronounced in females than males. (A) X-ray of mouse femur showing regions used for midshaft cortical and distal cortical and trabecular analysis. (B) Bright field montaged images of midshaft and distal sections used for confocal imaging of 5 and 22mo mouse femurs with imaging regions indicated by numbered black boxes (Bars = 500μm). (C) MicroCT reconstructions for whole femurs and midshaft cortical and distal cortical and trabecular bone. Graphs in (D-G) show microCT quantitation of midshaft cortical bone parameters, (H-K) show distal cortical bone parameters and (L-O) show quantitation of trabecular bone parameters in young and aged mice. (Data are mean ± SEM, * = p< 0.05, ANOVA/Tukey’s, females n= 8, males n= 6-7) (Note: n=6 was used for the aged female group in graphs (M-O) because in two aged females the trabecular number was so low that the software cannot reliably compute values for trabecular spacing).

Cortical porosities were observed in the femurs in several of the aged mice, particularly in females at the distal end (see Fig. 2A arrow and 2B asterisks). Therefore, an additional cortical analysis was performed on the distal femur. This showed that the distal cortical BV/TV was significantly decreased by 6.3% in females but not males (Fig. 1H), consistent with increased cortical porosities. Females but not males showed a significant 19% decrease in distal cortical bone area (Fig. 1I) and 20% increase in distal cortical bone perimeter (Fig. 1J). The reduced cortical thickness was more pronounced in the distal region than the midshaft, with a significant 36% reduction in females and a 19% reduction in males (Fig. 1K). Distal cortical bone parameters were not statistically significant between genders except for distal cortical bone thickness, which was significantly lower in males vs. females at 5mo (Fig. 1K).

Figure 2. Aging is associated with cortical porosities, regional variation in mineral density and growth plate closure. (A) microCT reconstruction of distal femur cortical bone in an aged female showing cortical porosity (arrow). (B) BSEM image of the distal femur (Bar = 500μm) and midshaft cortical bone (Bar = 50μm) in young and aged C57BL/6 mice. Arrow indicates growth plate closure in the aged male and * indicates cortical porosities. (C) Quantitation of cortical porosity from n=5 animals showing a significant increase in cortical porosity in female C57BL/6 mice with age.

Micro CT analysis of trabecular bone parameters showed that by 5mo, there was a significant gender difference in trabecular BV/TV, with males having ~2-fold higher BV/TV compared to females (Fig. 1C, L). A significant reduction in BV/TV was seen with aging in both genders (66% in females and 40% in males) (Fig. 1C, 1L). This was due to decreased trabecular number and increased spacing (Fig. 1M, N) rather than changes in thickness (Fig. 1O). No significant differences were seen in BMD between genders or age groups in cortical or trabecular bone (data not shown).

Backscattered SEM (BSEM) shows cortical porosities in aged mice and closure of the growth plate in aged males

BSEM confirmed reduced cortical thickness and expansion of cortical diameter in aged compared to young mice and showed cortical porosities in several of the aged mice, particularly females (Fig. 2B, asterisks). This was more prevalent at the distal end of the femur. In all male mice, the growth plate was either completely or partially closed (Fig. 2B, arrow). BSEM also showed greater regional variability in bone density in aged mice, indicated by variation in greyscale intensity of the cortical bone (Fig. 2B). Quantitation of the cortical porosities revealed a significant increase in porosity in females but not males with age (Fig. 2C).

Degenerative changes in osteocyte connectivity and density in aged mice

Confocal imaging of phalloidin/DAPI stained thick sections was performed in the same cortical regions to determine how aging affects the osteocyte networks (Fig. 3A-D). This revealed a dramatic reduction in dendrite connectivity with aging in both genders (Fig. 3A and Supplementary Movie 1 and Supplementary Movie 2), which was more severe in females. Gaps were seen in the osteocyte network where an osteocyte would be expected to be located but was absent (Fig. 3A, arrowheads). We also observed “islands” of osteocytes disconnected from other osteocytes (Fig. 3B, C). These osteocyte “islands” were often located near discontinuities or cement lines as shown in Fig. 3B (arrow), suggesting remodeling. These data show not only reduced dendrite number but also reduced osteocyte connectivity with aging. Occasionally in aged animals, dendrites were seen without an osteocyte cell body (Fig. 3D, circle), suggesting the osteocyte had died, leaving its dendrites behind. Supplementary Figure S1 shows additional confocal images of osteocytes in all three imaged regions of the cortex (see Fig. 1B for locations of imaged regions) to illustrate the regional variation in osteocyte dendrite connectivity.

Figure 3. Degeneration of the osteocyte network in aged bone. (A) Maximal Z-projections of 250 planes (32.5μm) from 100x confocal images of phalloidin (green) and DAPI (blue) stained midshaft femur sections showing osteocyte connectivity in young and aged male and female mice. (B) Dendrites do not cross over discontinuities in the bone (arrow) although matrix is present as seen in the corresponding brightfield image. (C) Some osteocytes in aged animals are found in “islands” with few or no connections to the surrounding osteocytes (dashed box). Enlarged images in (C) show a 3D render (front and back) of the same osteocyte confirming its lack of connectivity with surrounding osteocytes. (D) Occasionally dendrites with no visible cell body were seen in the aged mice (dashed circle) suggesting the dendrites may be left behind after apoptosis. (E) Maximal Z-projections of 20 planes (2.6μm) from DiI (red) and DAPI (blue) labeled femur sections showing staining of the osteocyte cell membrane in young and aged mice. Black and white insets show lipid material in the matrix around the dendrites. (Bars = 10μm).

Observations of phalloidin/DAPI staining were confirmed by confocal imaging using osteocyte cell membrane staining with DiI (Fig. 3E), showing that the observed reduction in dendrite number was not just due to decreased levels of actin in aged osteocytes. Imaging with DiI also showed decreased dendrite connectivity in aged mice that was more severe in females than males. The membrane dye also revealed extensive amounts of lipid material in the matrix that decreased in aged compared to young bone (Fig. 3E, insets). This lipid material is consistent with extracellular vesicle-like structures shed by osteocytes that we have described previously [35].

Quantitation of confocal phalloidin/DAPI stained image stacks showed a significant reduction in osteocyte number per mm³ of 18% in females and in 27% in males in the midshaft (Fig. 4A). A significant 21% decrease was seen for males in the distal region as well, with females trending downwards (Fig. 4B). Differences in osteocyte density between genders were not significant. Accompanying decreased osteocyte density, there was a dramatic decrease in dendrite number per osteocyte in aged mice (Fig. 4C, D). The effect was more severe in females (45% reduction) compared to males (27% reduction), particularly in the midshaft, however, a significant decrease was also found for both genders in the distal region. Frequency distribution plots of dendrite number per osteocyte (Fig. 4E, F) showed that in females the osteocyte populations in young and aged mice were more clearly separated than in males. Male mice had a wider distribution of dendrites per cell compared to females and in both genders there was a shift in the population towards lower numbers of dendrites per osteocyte with aging. Additionally, the number of canaliculi per osteocyte was quantified from confocal image stacks in mice injected with fixable 10kDa Texas Red-dextran and stained with phalloidin/DAPI. The Texas Red-dextran permeates the lacunar and canalicular fluid space and can be used to image the canaliculi and lacunae [35]. This revealed that the number of canaliculi per lacuna in the midshaft also decreased significantly with age, by 26% in females and 30% in males (Fig. 4G). Interestingly, the canaliculi outnumbered primary dendrites by 1.2-1.4 fold in the young mice and by 1.5-1.7 fold in the aged mice, showing that not every canaliculus is occupied by a dendrite.

Figure 4. Aging is associated with reduced osteocyte dendrite connectivity, which precedes a decline in osteocyte number. Quantitation of osteocyte density (A, B) and dendrite number per osteocyte (C, D) in the midshaft and distal femur from phalloidin stained sections in young and aged mice. (E, F) show frequency distribution plots of the number of dendrites per osteocyte in young and aged female (E) and male (F) mice. (G) Quantitation of the number of canaliculi per lacuna from Texas Red-dextran labeled midshaft femur sections in young and aged mice. (H) Time course showing a decline in dendrite number per osteocyte in female mice over 5, 12, 18 and 22mo. (I) Regression plot showing the linear decline in osteocyte dendrite number between 5 and 22mo in female mice. (J) Time course showing changes in osteocyte density in female mice at 5, 12, 18 and 22 mo. (K) Quantitation of empty lacunae in young and aged mouse femurs. (L-N) Scatterplots of osteocyte dendrite number versus cortical thickness (L), cortical perimeter (M) and trabecular BV/TV (N) with Spearman’s correlation coefficient and significance indicated. (Data are mean ± SEM, * = p< 0.05, ANOVA/Tukey’s) (A-D and K), females n=5, males n=6; G, n=5; (H, J, n= 3-5).

To further understand the kinetics of the decline in osteocyte density and dendrites, an extended time course of 5, 12, 18 and 22mo was analyzed in females (Fig. 4H). This showed a linear decline in dendrite number per osteocyte from 5 through 22 mo. A linear regression plot of age against dendrite number confirmed a highly significant negative linear correlation (R²= 0.9858, p<0.01) (Fig. 4I). Interestingly, osteocyte number per mm³ bone peaked at 12mo (Fig. 4J), when dendrite number had already started to decline (Fig. 4H). The osteocyte density then steadily declined from 12 to 22 mo. Therefore the decline in dendrite number preceded the decline in osteocyte number. The number of empty osteocyte lacunae was measured in undecalcified plastic sections of the femur (Fig. 4K). The overall percentage of empty lacunae was low (≤3% in all groups) and increased significantly with age in females, doubling from ~1.5% to 3%. Males did not show a significant increase in empty lacunae, which may be partly because the percentage started out low (less than 1%) in young males. Fig. 4L-N show scatter plots of osteocyte dendrite number versus cortical thickness, cortical perimeter and trabecular BV/TV. Spearman correlation tests showed a significant positive correlation between dendrite number and cortical thickness (r_s = 0.5631, p<0.01) and a negative correlation between dendrite number and cortical perimeter (r_s = -0.7251, p<0.001), with no significant correlation between trabecular BV/TV and dendrite number.

Changes in osteocyte cell body and lacunar volume with age

In mice injected with fixable Texas Red-dextran and stained with the membrane dye DiO, confocal image stacks were used to create 3D volumes of the osteocyte cell body and its corresponding lacuna (Fig. 5A). Quantitative analysis showed that the osteocyte cell body volume in female and male mice decreased significantly with age (15% and 19%, respectively) (Fig. 5B) in the midshaft with a similar reduction in the distal region (Fig. 5C). While the cell body volume decreased in aged female mice the average lacunar volume was not significantly changed (Fig. 5D, E). In contrast, males showed a small decrease in lacunar volume with aging which was significant in the distal region (Fig. 5D, E). This results in an increased lacunar fluid space in female mice but no change in males with aging (Fig. 5F, G).

Figure 5. Decreased osteocyte cell volume with decreased lacunar volume in aged males but not females leads to gender differences in lacunar fluid volumes. (A) Examples of 3D renderings of the osteocyte cell bodies labeled with DiO and lacunae labeled with fixable Tx-Red-dextran for volumetric calculations. (B and C) quantitation of osteocyte cell body volume in the midshaft (B) and distal (C) region. (D and E) quantitation of the corresponding lacunar volumes. (F and G) show calculation of the lacunar fluid space derived from subtracting the osteocyte cell body volume from its corresponding lacunar volume. (Data are mean ± SEM, * = p< 0.05, ANOVA/Tukey’s; n= 5).

Histology reveals cortical porosities, loss of growth plate in males and altered osteoblast and osteoclast activity in aged animals

Histology using von Kossa tetrachrome staining confirmed the age related loss of femoral trabecular bone and expansion of cortical diameter in males and females (Fig. 6A). Loss of the growth plate was apparent in all 6 aged male femurs with both von Kossa Tetrachrome and Safranin O staining and ranged from partial to complete loss (Fig. 6A, B, arrow). In contrast, all females retained the growth plate although it was thinner with aging (Fig. 6B). Histological examination of the cortical porosities revealed the presence of extramedullary marrow (Fig. 6A, C, asterisks). TRAP staining suggested increased numbers of osteoclasts in male and female aged mice (Fig. 6C), which was particularly prominent in extramedullary marrow spaces associated with cortical porosities (asterisk). However, the TRAP staining in aged mice was variable between animals (Fig. 6C).

Figure 6. Histomorphometry shows increased extramedullary marrow, decreased osteoblast activity and decreased growth plate area and thickness in aged mice. (A) von Kossa Tetrachrome stained sections (Bar = 1000μm) and (B) Safranin O stained sections (bar = 500μm) from the distal femur in young and aged mice. (C) TRAP stained sections at the midshaft. The images on two rows show different individual mice to illustrate the variability of TRAP staining in aged mice (Bar = 200μm). Arrows indicate regions where the growth plate has closed and * indicates extramedullary marrow spaces. (D) Quantitation of the extramedullary marrow space area, (E) growth plate area and (F) growth plate thickness from von Kossa Tetrachrome stained sections in young and aged mice. (G-L) show histomorphometric quantitation of osteoblast and osteoclast parameters in femoral trabecular bone of young and aged mice. (M and N) show histomorphometric quantitation of osteoclast parameters in cortical bone of young and aged mice. (Data are mean ± SEM, * = p< 0.05, ANOVA/Tukey’s; females n=5, males n=6).

The area of extramedullary marrow spaces (marrow filled cortical porosities) was measured as a percentage of total bone area. A significant increase in extramedullary marrow spaces was seen in females with age (Fig. 6D). Males trended towards an increase that did not reach statistical significance due to the variability in aged males. Measurements from safranin O stained sections showed a significant decrease in growth plate area and thickness in both genders (Fig. 6E, F). No significant differences between genders were seen although males showed a larger reduction in growth plate area than females, due to complete or partial loss of the growth plate in males.

Histomorphometric analysis in von Kossa tetrachrome stained sections showed a significant loss of trabecular BV/TV in young vs. old mice of both genders (Fig. 6G) in agreement with microCT data. The number of osteoblasts per bone perimeter (Ob/B.Pm), osteoblast surface per bone perimeter (ObS/B.Pm), and osteoid surface/B.Pm in trabecular bone were highest in young male mice and were significantly reduced in aged males (Fig. 6H-J). Females already had significantly lower values for all these trabecular osteoblast parameters than males at 5mo and showed no additional significant change with aging (Fig. 6H-J). No significant differences were observed in osteoclast number (NOcl/B.Pm) or osteoclast surface (OCLS/B.Pm) in the trabecular bone compartment (Fig. 6K-L). In the cortical bone there was a significant increase in the total number of osteoclasts in aged male mice (from 36.33±10.48 at 5mo to 105.9±36.81 at 22mo, p<0.05) but not females (58.20±30.97 at 5mo and 161.0±59.15 at 22mo). However, since much of this osteoclast activity was associated with remodeling within the cortical porosities, the bone surface was also increased and therefore differences in the NOcl/B.Pm and OCLS/B.Pm did not reach significance (Fig. 6M, N).

Aging in the C57BL/6 mouse model is associated with calcification of the achilles tendons

Radiographs of the hindlimbs showed that calcification of both Achilles tendons occurred in 100% of male and female aged mice. In contrast, none of the 5mo old mice showed tendon calcification (Suppl. Figure S2, arrows).

Animals

The C57BL/6 mouse model of aging was used in which 5 month old (5mo) mice represent young adults and 22-23mo old mice represent aged (geriatric) adults [6,51]. 12 and 18mo mice were used in some experiments to model early aging. Mice of both genders were obtained from the NIH/NIA aged rodent colony (sample sizes for experiments are indicated on the accompanying figure legends). Animal experiments and euthanasia were performed under an approved IACUC protocol at the University of Missouri Kansas City (UMKC), and conformed to relevant federal guidelines. The UMKC animal facility is AAALAC approved and animal care and husbandry meets requirements in the Guide for the Care and use of Laboratory Animals (8th Ed.), National Research Council. Animals were group housed and maintained on a 12hr light/dark cycle with ad libitum food and water at 22^oC constant temperature and 45-55% humidity.

Bone sample preparation

Unless stated otherwise, reagents and chemicals were purchased from Sigma Aldrich, St. Louis, MO. Mice were humanely euthanized and the femurs were fixed in 4% paraformaldehyde (PFA) in PBS, pH 7.4 at 4°C for 24hr. One femur was stored in 70% ethanol at 4°C prior to microCT analysis (see below). After microCT, the undercalcified femurs were dehydrated in graded ethanols. They were then infiltrated with acetone, then 1:1 followed by 1:2 acetone with methyl methacrylate infiltration solution [84% methyl methacrylate (MMA), 14% dibutyl phthalate, 1% polyethylene glycol, 0.7% benzoyl peroxide], and then 100% MMA for 5 days before plastic embedding in MMA with 0.033% N,N-dimethyl p-toluidine at -20°C for 3-5 days. The other femur was prepared for cryosectioning as described previously [35]. Briefly, after PFA fixation the femurs were decalcified in 10% EDTA, pH 7.4 for 1-2 wks, then equilibrated in 15% then 30% sucrose in PBS before embedding and freezing in Tissue-Tek O.C.T compound (Sakura Finetek USA Inc., Torrence CA).

MicroCT

MicroCT analysis was performed with a vivaCT40 (Scanco Inc, Basel, Switzerland) in accordance with recommended guidelines [55] using an X-ray energy of 55kV (145µA), a voxel resolution of 10.5µm, 200ms integration time with the number of projections set at 1000/180 degrees and using a 0.5mm aluminum low pass filter. The threshold was set to 316 (equivalent to 498.5 mgHA/ccm) for both cortical and trabecular bone to distinguish mineralized from non-mineralized tissue. Two VOIs were chosen for cortical bone analysis. First, 50 slices (525µm) were contoured in the midshaft starting at the distal end of the third trochanter and progressing toward the knee joint. Second, 100 slices (1050µm) were contoured starting at the metaphyseal end of the distal growth plate, below the primary spongiosa, and progressing toward the midshaft. This same VOI was used for contouring of trabecular bone with cortical bone excluded. Bone properties were analyzed using Scanco bone evaluation software.

3D Multiplexed confocal imaging

For 3D multiplexed confocal imaging of osteocyte networks, 50μm transverse sections were cut from decalcified O.C.T. embedded femurs on a Leica CM3050S cryomicrotome (Leica Microsystems, Wetzlar, Germany). Sections were taken from the midshaft and distal regions as shown in Fig. 1A and stained en block. For full details of multiplexed staining methods see [35]. Briefly, to stain the actin cytoskeleton, sections were incubated in 165nM AlexaFluor-488-phalloidin (ThermoFisher Scientific, Waltham, MA) in PBS overnight followed by three PBS washes then a 30min incubation in 4μg/ml DAPI (ThermoFisher) in PBS. Stained sections were coverslip mounted in 1:1 glycerol:PBS with 1mM MgCl₂. For visualization of osteocyte membranes, sections were incubated for 2 days in 1,1'-dioctadecyl-3,3,3'3'-tetramethylindocarbocyanine perchlorate (DiI) or 3,3'-dioctadecyloxacarbocyanine perchlorate (DiO) at 100μM (ThermoFisher) in 9:1 ethanol:DMSO. Sections were washed briefly in ethanol followed by three PBS washes prior to labeling with DAPI as above. DiI or DiO labeled sections were mounted in 97% TDE (2,2’-thiodiethanol) mounting buffer as described previously [35], which matches the refractive index of the tissue samples to the immersion oil, thereby minimizing light spreading in the Z-axis. For experiments in which the LCS was imaged simultaneously with the osteocyte cell membrane, mice were injected intravenously with a Texas Red conjugated lysine fixable 10kDa dextran (ThermoFisher) (32 mg/kg in PBS) 4 min prior to euthanization. This dye distributes into the lacunae and canaliculi and is fixable with aldehyde based fixatives, so it survives decalcification.

For confocal imaging, two sections per animal for each type of staining were imaged on a Leica TCS Sp5 II confocal microscope (Leica Microsystems, Wetzlar, Germany) in resonant scanner mode with line averaging set to 96. A tiled image of the entire transverse section was obtained with the 5x objective (NA 0.15 Zoom 1.7) followed by collecting detailed Z-stacks of 250-350 Z-planes with the 100x oil objective (NA 1.44 zoom 1.7 with a 0.13 μm step size) from three standardized regions in the midshaft sections and 4 regions in the distal sections as indicated in Fig. 1B. Using tranverse sections means that the osteocytes are imaged with their longest axis as the Z-axis. For phalloidin imaging 488nm laser excitation was used with an emission collection window of 493-580 nm acquired together with a brightfield image. For imaging DiI, laser excitation was 543 nm with a collection window of 553-650 nm. For DiO/Texas Red-dextran dual imaging the DiO was excited at 488 nm with an emission window of 494-564 nm and Texas Red was excited in the same scan at 594 nm with a 604-684nm collection window. In all confocal experiments DAPI signal was acquired in a separate scan using 405nm laser excitation and a collection window of 410-480 nm.

Quantitation of 3D osteocyte parameters

For determination of osteocyte density, 3D reconstructions were generated of phalloidin/DAPI stained 100x image stacks with 250 slices (32.5 μm tissue depth) (see Supplementary Movie 1 and Supplementary Movie 2). These 3D reconstructions were used to count osteocyte number in the 272,217 μm³ volume using ImageJ software (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, MD, http://imagej.nih.gov/ij/, 1997-2015) with the cell counting plugin (Author: Kurt De Vos, University of Sheffield Academic Neurology). To avoid over or under estimation of osteocyte numbers, osteocytes caught in partial section at the edges of the 3D volume were included in the count on only three of the six faces of the cube volume. The mean number of osteocytes per mm³ of bone was determined for each animal in both midshaft and distal regions.

To measure the number of dendrites per osteocyte, 100x stacks of phalloidin labeled sections were contrasted in ImageJ to visualize the dendrites in the entire stack. Dendrites were only counted if the entire osteocyte cell body was captured in the Z stack and all dendrite sprouting points at the cell body could be discerned. Using the Image J Cell Counting plugin, dendrites were counted by scrolling through the Z stack and counting each dendrite at its sprouting point from the cell body. At least 60 osteocytes were counted per animal (i.e. per data point) for both midshaft and distal regions to obtain the average number of dendrites per osteocyte for each animal. Quantitation of canaliculi per lacuna was performed in a similar manner using stacks with Texas Red Dextran labeling of the lacunocanalicular system.

Image stacks from dextran, DiO and DAPI multiplexed imaging were used for simultaneous determination of lacunar and cell body volumes. To determine osteocyte cell body volume RGB image stacks of DiO images merged with DAPI were converted to 8 bit greyscale images to create a “filled in” image of the intact osteocyte. The 8 bit stacks were then normalized to correct for signal loss with tissue depth, and thresholded in the Object Counter 3D plugin [56] so that just the cell body was measured and not cell processes. A lower volume limit of 10 μm³ was used. Numbered volumes were inspected in a 3D projection and volumes not representing an intact osteocyte (i.e. partially sectioned osteocytes) were excluded. Lacunar volumes were determined using the same method from dextran/DiO/DAPI greyscale images. By subtracting the lacunar volume from its corresponding cell body volume a fluid space volume was determined for each lacuna. Only lacunae occupied by an osteocyte were measured and 40-60 lacunae were measured for both the midshaft and distal regions to determine the average cell body and lacunar size for each animal.

Histology

Undecalcified longitudinal plastic sections of the femur (5µm thickness) were cut on a Microm HM 355S microtome with tungsten carbide blade. After the sections were deplasticized and rehydrated, von Kossa tetrachrome staining was performed using standard histological procedures. Histochemical staining for TRAP was done using standard procedures with an extended incubation time of 5hrs at 37°C in Naphthol AS-BI Phosphate substrate and using toluidine blue as the counterstain. Safranin O staining was performed with Fast Green counterstaining using standard histological staining methods. Stained sections were coverslip mounted in permount.

Histomorphometry

For all histomorphometric quantitations, three non-consecutive longitudinal sections were measured per animal and mean values from these sections used as the data point for one animal. Osteoblast and osteoclast quantitation was performed on von Kossa Tetrachrome or TRAP stained sections, respectively using a Nikon E800 microscope (Nikon Instruments, Inc. Melville NY) and SONY Exwave HAD camera (Sony Corp., New York, NY) interfaced with an OsteomeasureXP bone histomorphometry system (OsteoMetrics Decatur, GA). Measurements were made from trabecular bone in the secondary spongiosa at a standard location 80μm below the growth plate extending ~450μm using a 20x objective. The following parameters were measured directly or derived from measured indices: bone volume per total volume (BV/TV), osteoblast number, osteoblast surface and osteoid surface per bone surface (N.Ob/B.Pm, Ob.S/B.Pm, OS/B.Pm), osteoclast number and osteoclast surface per bone surface (N.Ocl/B.Pm, OclS/B.Pm). For quantitation of empty osteocyte lacunae and TRAP positive osteoclasts in cortical bone, three non-consecutive longitudinal TRAP-stained femur sections were used per animal. For each section, twelve images were taken of the central midshaft, six from the medial and six from the lateral cortex using a 20x objective. The number of empty lacunae and lacunae occupied by an osteocyte were counted using the Cell Counter Plugin in ImageJ. Lacunae with a long axis <5μm were excluded as they represent sections that caught the outer edge of a lacuna and it could not be reliably determined if they were occupied. Osteoclast parameters for cortical bone were measured using the Cell Counting plugin and Measure functions in ImageJ.

Backscattered SEM (BSEM)

For BSEM, the MMA embedded blocks remaining after sectioning for histology were polished with graded sandpapers (600, 800, 1200 grit [Buehler, Lake Bluff IL]) followed by Metadi Supreme Polycrystalline Diamond Suspensions of 1, 0.25, and 0.05μm (Buehler). Observation of backscattered electrons was performed on a field-emission SEM XL-30 scanning electron microscope (FEI, Hillsboro, OR) equipped with a solid state BSE detector on Au-Pd coated specimens using 15.0kV energy at 40x and 100x magnification. Quantitation of the cortical porosity as a percentage of the bone volume was done from 40x BSEM images of the distal region of the femur in ImageJ and using the measure analysis tool to determine the areas.

Statistical analysis

For statistical comparisons, one way analysis of variance (ANOVA) was used followed by Tukey’s post hoc test. Linear regression and Spearman’s Correlation tests were used to test for significant correlations between datasets. Data were considered statistically significant with a p value ≤0.05.