Differential responses of mechanosensitive osteocyte proteins in fore- and hindlimbs of hindlimb-unloaded rats

Accepted Manuscript Differential responses of mechanosensitive osteocyte proteins in forelimbs and hindlimbs in hindlimb unloaded rats

Corinne E. Metzger, Jessica E. Brezicha, Jon P. Elizondo, S. Anand Narayanan, Harry A. Hogan, Susan A. Bloomfield PII: DOI: Reference:

S8756-3282(17)30266-1 doi: 10.1016/j.bone.2017.08.002 BON 11388

To appear in:

Bone

Received date: Revised date: Accepted date:

13 April 2017 12 July 2017 2 August 2017

Please cite this article as: Corinne E. Metzger, Jessica E. Brezicha, Jon P. Elizondo, S. Anand Narayanan, Harry A. Hogan, Susan A. Bloomfield , Differential responses of mechanosensitive osteocyte proteins in forelimbs and hindlimbs in hindlimb unloaded rats, Bone (2017), doi: 10.1016/j.bone.2017.08.002

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Differential responses of mechanosensitive osteocyte proteins in forelimbs and hindlimbs in hindlimb unloaded rats Corinne E. Metzger1, Jessica E Brezicha2, Jon P. Elizondo3, S. Anand Narayanan4, Harry A. Hogan2,3, Susan A. Bloomfield1. Department of Health and Kinesiology, Texas A&M University, College Station, TX, 77843

2

Department of Biomedical Engineering, Texas A&M University, College Station, TX, 77843

3

Department of Mechanical Engineering, Texas A&M University, College Station, TX, 77843

4

Department of Medical Physiology, Texas A&M University Health Science Center, Temple, TX, 76504

ED

M

AN

US

CR

IP

T

1

PT

Corresponding Author: Susan A. Bloomfield

CE

Department of Health & Kinesiology, MS 4243 Texas A&M University

AC

College Station, TX 77843-4243

[email protected], 979-845-2871

ACCEPTED MANUSCRIPT Abstract Osteocytes are believed to be the primary mechanosensors of bone tissue, signaling to osteoblasts and osteoclasts by releasing specific proteins. Sclerostin, interleukin-6 (IL-6), and insulinlike growth factor-I (IGF-I) are osteocyte proteins that signal to osteoblasts. This study’s goals are to

T

determine if osteocyte protein response to mechanical unloading is restricted to the unloaded bone using

IP

the hindlimb unloading (HU) rodent model. We also examined tumor necrosis factor-α (TNF-α) due to

CR

its interactions with all three osteocyte proteins. We hypothesized that unloaded hindlimb cancellous

US

bone would have an altered osteocyte protein (sclerostin, IL-6, and IGF-I) response compared to controls, while the response in the weight-bearing forelimb would not differ from ambulating controls.

AN

Male Sprague Dawley rats (7-mo old) experienced either HU (n=7) or normal cage activity (CON; n=7) for 28 days. The unloaded distal femur and the weight-bearing proximal humerus were compared in HU

M

vs CON. Metaphyseal bone density was reduced in the HU rats’ hindlimb, but not in the proximal

ED

humerus, compared to CON values. Osteocyte density was 30% lower in the HU distal femur, but not

PT

different from CON in the proximal humerus. %Sclerostin+ osteocytes in the distal femur were higher in HU compared to CON, but lower in the proximal humerus. Both %IGF-I+ and %IL-6+ osteocytes were

CE

lower in the distal femur for HU, but higher in the proximal humerus for HU. Osterix surface, a marker

AC

of osteoblasts, was lower in HU in the distal femur; however, the proximal humerus had more %osterix+ surface in HU. In HU %Cathepsin K+ surface, a marker of osteoclasts, was higher in the distal femur and lower in the proximal humerus. %TNF-α+ osteocytes were no different from CON in either bone site. HU proximal humerus osteocyte protein responses of sclerostin, IL-6, and IGF-I changed in the opposite direction as observed in the distal femur within the same animal. The opposite response of osteocyte proteins and osteoblast surface in hind- and forelimb bones within the same animal suggests that, while osteocytes in the unloaded hindlimb sense a lack of mechanical strain, osteocytes in the

ACCEPTED MANUSCRIPT weight-bearing forelimb in HU animals sense some increase in local strain and generate molecular signaling to osteoblasts. Key Words: Sclerostin, Interleukin-6, Insulin-like Growth Factor-I, mechanical loading

T

1. Introduction

IP

Bone is a dynamic tissue sensitive to environmental stimuli such as loading-induced stresses and

CR

strains. A reduction in mechanical loading on bone results in bone loss and increased risk of fractures. For example, astronauts during long duration (4-6 month) spaceflight missions experience site-specific

US

bone loss at weight-bearing bone sites (32) while there is no difference in bone in non-weight-bearing

AN

sites like the radius (17, 40). More relevant to clinical populations on Earth, 17 weeks of bed rest can cause 1.4%-10.4% bone mineral content loss in lower limb skeletal sites (16). Additionally, many

M

hindlimb unloading (HU) rodent studies have demonstrated bone loss due to disuse (1, 5, 19, 23, 35).

ED

Cancellous bone is particularly sensitive to disuse, with 28 days of HU in male rats resulting in significant losses in cancellous volumetric bone mineral density and cancellous bone volume (5, 35).

PT

While the mechanical regulation of bone mass is well known, the exact mechanisms that orchestrate the

CE

site-specific bone adaptations to loading and unloading are less well understood. Research over the past decade has shown how osteocytes, bone cells embedded in the bone

AC

matrix, are the primary regulators of bone cell activity and also key mechanosensors of bone tissue (6, 7, 13, 39). These cells sense stresses on the bone via fluid shear stress as well as sensing the load on their cell bodies, dendritic processes, or cilia (7). Osteocytes communicate via dendritic processes and also by the release of proteins that send messages to, directly or indirectly, stimulate or halt osteoblastogenesis and osteoclastogenesis (7, 13). Consequently, osteocytes may orchestrate alterations in bone mass seen with loading and unloading based on the proteins they bind and/or release.

ACCEPTED MANUSCRIPT One of the most well-known osteocyte proteins is sclerostin, a product of the SOST gene, which inhibits bone formation (38). With mechanical unloading, sclerostin gene and protein expression increase; with increased loading, sclerostin decreases (19, 28). Two less-understood osteocyte proteins that appear to be mechanically regulated are insulin-like growth factor-I (IGF-I) and interleukin-6 (IL-

T

6). Mechanical loading increases in vivo osteocyte-derived IGF-I measured by gene expression (15, 18,

IP

20, 27). An in vitro study of cultured osteocytes demonstrated that shear-loaded osteocytes produced IL-

CR

6, which was proposed to signal to osteoblasts (4); however, the exact role of IL-6 in osteocyte signaling and mechanotransduction is equivocal. Furthermore, factors beyond mechanical stimuli, such as

US

inflammation (2, 21), can also alter osteocyte proteins. Sclerostin, IGF-I, and IL-6 are influenced by the

AN

pro-inflammatory cytokine, tumor necrosis factor-α (TNF-α). TNF-α is a transcriptional activator of sclerostin (2). In osteoblasts, TNF-α suppresses IGF-I secretion (30); it also synergistically works with

M

IL-6 to enhance osteoclast resorption of bone (34). Additionally, TNF-α stimulates osteocyte apoptosis,

ED

but this mechanism is inhibited by shear loading stress on osteocytes (37). While not known to be expressed by osteocytes, TNF-α positively stains in osteocytes (2, 21) indicating it may bind to

PT

osteocytes, potentially modulating the release of osteocyte proteins (e.g. sclerostin).

CE

The goals of this current study were threefold: firstly, examine the roles of osteocyte proteins sclerostin, IGF-I, and IL-6 in disuse using the HU model and correlate these changes with alterations in

AC

bone turnover. We hypothesized that in the unloaded hindlimb, osteocyte sclerostin would increase and osteocyte IGF-I and IL-6 would decrease after 28 days of disuse, concurrent with the expected reductions in bone mass, osteocyte density, and osteoblast-covered surfaces. Secondly, we aimed to determine if the response of all three osteocyte proteins is confined to the site-specific unloaded bone or if osteocytes in bone sites experiencing usual loading patterns would have similar changes to the unloaded bone. Sclerostin is only altered in an externally loaded ulna and not the contralateral untreated

ACCEPTED MANUSCRIPT ulna (28) supporting its local response to mechanical stimuli; however, no such data exists for osteocyte IL-6 and IGF-I. For this second aim, we examined osteocyte proteins in the weight-bearing humeri in rats experiencing HU compared to the unloaded femur within the same animal and also compared to normally-ambulating control animals. We hypothesized that cancellous bone in humeri of HU rats

T

would exhibit no changes in osteocyte density and osteocytes positive for sclerostin, IL-6, and IGF-I

IP

compared to ambulating control rats. Finally, we examined if osteocyte-associated TNF-α was different

CR

in unloaded femurs or weight-bearing humeri as compared to corresponding ambulating control bone sites and if these changes corresponded with changes in the other three proteins.

US

2. Methods

AN

2.1 Animals: Adult male Sprague Dawley rats (Harlan, Houston, TX, USA) were obtained at 5 months old and allowed to acclimate for 8 weeks before the initiation of the study; therefore, rats were 7 months

M

old at the beginning of hindlimb unloading. Tissues for this study were taken from a larger study with

ED

separate hypotheses that required a long acclimation period. All rats were singly housed, fed standard rat chow (Teklad 2018; Envigo, Houston, TX, USA), and had free access to water for the duration of the

PT

study. The rats were kept in a 12-hour light/dark cycle at a constant room temperature of 23° C in an

CE

institutionally approved animal facility. All rats were weighed twice weekly, and health was monitored twice daily. Animals were block assigned by weight and total volumetric bone mineral density at the left

AC

proximal tibia metaphysis (vBMD; measured in vivo via peripheral quantitative computed tomography) to ensure equal variance and no statistical differences in these two parameters between groups in body weight and proximal tibia vBMD at the beginning of the experimental period. The two groups consisted of normally ambulating controls (CON; n=7) and a hindlimb unloaded treatment group (HU; n=7). Statistical power analyses for our outcomes showed n=6 as sufficient power (β=0.8) for all bone mass, histomorphometric, and histological analyses. The HU group underwent 28 days of hindlimb unloading

ACCEPTED MANUSCRIPT starting at 7 months of age. Following the 28 day period of HU, animals were anesthetized via intraperitoneal injection of ketamine and dexmedetomidine (at a ratio of 3:2; Henry Schein Animal Health, Dublin, OH, USA) and euthanized via exsanguination and decapitation. All animal care and experimental procedures for this study were approved and conducted in accordance with the Texas

T

A&M University Institutional Animal Use and Care Committee and conform to the NIH Guide for the

IP

Care and Use of Laboratory Animals.

CR

2.2 Hindlimb Unloading: Hindlimb unloading was achieved by tail suspension based on the methods

US

developed by Morey-Holton et al. (24). Rats were anesthetized via intraperitoneal injection of ketamine and dexmedetomidine (at a ratio of 3:2). Tails were cleaned and sterilized before application of a custom

AN

harness to the lateral sides of the tail via a thin layer of adhesive (Amazing Goop, Eclectic Products, Los

M

Angeles, CA, USA). The harness was allowed to dry for ~30 minutes. After the harness was dry, the anesthesia was reversed via intramuscular injection of atipamexole hydrochloride (Henry Schein Animal

ED

Health, Dublin, OH, USA). Rats were then placed individually in a 46x46x46 centimeter cage with a

PT

pulley system at the top. The tail harness was linked to the pulley system, but rats were initially allowed to ambulate normally on all four limbs to acclimate to the cage and harness. After 24 hours, the harness

CE

was lifted on the pulley system to suspend the hindlimbs off the ground at approximately a 30º head-

AC

down tilt. The forelimbs remained weight-bearing, allowing the rat to ambulate on the cage floor. Ad libitum access to food and water was maintained throughout the entire suspension period. Every 12 hours throughout the 28 day period, the health of the rats was monitored and the suspension height was adjusted, if needed, to ensure hindlimbs remained elevated off the cage floor. During the unloading period, rats had access to environmental enrichment devices and minimal bedding. Currently, our group has a 95% success rate of rats completing the 28-day protocol without complications that require removal from the study. In the current study, all HU rats completed 28 days of hindlimb unloading.

ACCEPTED MANUSCRIPT 2.3 Peripheral Quantitative Computed Tomography (pQCT): Tomographic scans were performed ex vivo on right humeri and left femurs (stored frozen in gauze soaked in phosphate-buffered saline at 35ºC) using a Stratec XCT Research-M device (Norland Corp., Fort Atkinson, WI). The scanning fan beam thickness was 0.5 mm, voxel resolution was 70 μm, and scanning speed was 2.5 mm/s. Before

T

each use, the machine was calibrated with a hydroxyapatite standard cone and cortical phantom. Thawed

IP

humeri and femurs were placed in a vial filled with phosphate-buffered saline to maintain proper

CR

hydration during the scan, after which they were returned to a −35°C freezer. For each animal, transverse images were taken at the distal femur metaphysis and proximal humerus metaphysis. Four

US

images spaced 0.5 mm apart were taken of the femur, and three images spaced 1.0 mm apart were taken

AN

of the humerus. Femur metaphysis scans were centered 5.25 mm proximal from the point midway between the intercondylar fossa and medial/lateral condyles. Humerus metaphysis scans were centered

M

1.0 mm distal to the most distal point of the humeral head. Analyses were performed using

ED

Stratec software (v6.00, Norland Corp., Fort Atkinson, WI). A standardized analysis for metaphyseal bone (contour mode 3, peel mode 2, contour threshold 450 g/mm3, peel threshold 800 g/mm3) was

PT

applied to each scan. For each bone, total (cancellous + metaphyseal cortical bone) bone mineral content

CE

(BMC), total volumetric bone mineral density (vBMD), cancellous vBMD, and cortical vBMD were calculated. Machine precision (based on manufacturer data) is ±3.0 mg/cm3 for cancellous bone and

AC

±9.0 mg/cm3 for cortical bone. Reproducibility was measured by randomly selecting three animals from the control group and taking five images at the same metaphysis location for each bone type. Coefficients of variation were calculated from these measurements (Table 1). 2.4 Static histomorphometry of the proximal tibia: To validate disuse-induced bone loss and alterations in forming/resorbing surfaces, histomorphometry was completed on the proximal tibia metaphysis. For cancellous histomorphometry measures, undemineralized right proximal tibia were fixed in 4%

ACCEPTED MANUSCRIPT phosphate-buffered formalin for 24 hours and then subjected to serial dehydration and embedded in methyl methacrylate (Aldrich M5, 590-9, St. Louis, MO, USA). Serial frontal sections were cut 4 μmthick and treated with von Kossa stain and tetrachrome counterstain. The histomorphometric analyses were performed using OsteoMeasure Analysis System, version 3.3 (OsteoMetrics, Inc., Atlanta, GA,

T

USA). A defined region of interest was established approximately 500 μm from the growth plate and

IP

within the endocortical edges, encompassing approximately 4 mm2 at 40x magnification. Cancellous

CR

bone volume (BV/TV), osteoid (OS/BS) and osteoclast (Oc.S/BS) surfaces as a percent of total cancellous surface were measured at 40x magnification. All nomenclature for cancellous

US

histomorphometry follows standard usage (11).

AN

2.5 Immunohistochemistry: Right distal femurs and right proximal humeri were fixed in 4% phosphate-

M

buffered formalin for 24 hours at 4○C and then decalcified in a sodium citrate/formic acid solution for approximately 18 days. Samples were then stored in 70% ethanol. Sections were processed, and 8 µm

ED

sections underwent immunohistochemistry (IHC) as previously described (21). Sections were incubated

PT

with the following primary antibodies: polyclonal rabbit anti-IL-6 (Abcam, Inc, Cambridge, MA), rabbit polyclonal anti-IGF-I (Abcam), polyclonal goat anti-mouse sclerostin (R&D Systems, Minneapolis,

CE

MN), polyclonal rabbit anti-TNF-α (LifeSpan Biosciences, Inc., Seattle, WA), polyclonal rabbit anti-

AC

cathepsin K (Abcam), and polyclonal rabbit anti-Osterix/Sp7 (Abcam). Negative controls for all antibodies were completed by omitting the primary antibody. Sections stained for IGF-I, IL-6, and sclerostin were analyzed by quantifying the proportion of all osteocytes staining positively for the protein in the cancellous bone (~500 microns from the growth plate, an area of approximately 4 mm2). Osteocyte density was calculated by quantifying the number of osteocytes normalized per mm2 of area analyzed and also normalized to the bone volume/total volume within the bone sample (to give a density within the bone tissue only). Due to other requirements of the larger project, only one humerus bone was

ACCEPTED MANUSCRIPT available per animal, which disallowed performing both standard histology and IHC for the forelimb. Hence, we used IHC-determined osterix surface as a surrogate for osteoblasts and cathepsin K surface for osteoclasts. Sections stained for cathepsin K and osterix were analyzed by quantifying the percent cancellous surface covered by positive stain in a region of interest beginning ~500 microns from the

T

growth plate, an area of approximately 4 mm2 in both the distal femur and proximal humerus. All stains,

CR

analyses were completed by the same individual for consistency.

IP

antibodies, and analyses were identical between the distal femur and proximal humerus, and image

US

2.6 Statistical analyses: A paired t-test was used to compare pre- to post-HU body weights in CON and HU. A t-test was completed between the hindlimb unloading (HU) and ambulatory control (CON)

AN

groups for each variable. Effect size (partial eta-squared) was determined for values of p<0.05.

M

Statistical analyses were completed on SPSS (IBM; Armonk, NY). All data are represented as mean ±

ED

standard deviation. 3. Results

PT

All HU rats in this study successfully completed 28 days of hindlimb unloading without any health

CE

concerns requiring removal from the study. Comparing pre- to post-HU body weight, CON animals gained weight over the 28 days (p=0.002 in paired t-test) while the HU group had no difference in body

AC

weight over time (p=0.560). At the end of HU, hindlimb unloaded rats had significantly (p=0.028, effect size=0.342) lower body weight (476±29g) than did the ambulatory controls (CON; 515±28g). 3.1 Histomorphometry at the proximal tibia metaphysis validated disuse-induced bone loss in the unloaded hindlimb (Fig. 1). Static histormorphometry revealed lower cancellous bone volume in HU compared to CON (p=0.036; effect size=0.315). Additionally, HU had lower osteoid cancellous surface

ACCEPTED MANUSCRIPT (p=0.003; effect size=0.533) and higher cancellous osteoclast surface (p=0.001; effect size=0.632) compared to CON. 3.2 In the unloaded distal femur, bone mineral content and bone mineral density were lower in HU; however, the humeri of HU rats were not significantly different from those of control animals (Table 2).

IP

T

Total (cancellous + metaphyseal cortical bone) BMC and vBMD at the distal femur were lower in HU (p=0.006, effect size=0.483 and p<0.0001, effect size=0.694, respectively). Cancellous vBMD was also

CR

lower in HU compared to CON in the distal femur (p=0.007, effect size=0.462). The cortical shell

US

vBMD of the metaphysis in the distal femur was not significantly different (at p<0.05) between the groups (p=0.051). Total BMC and vBMD at the proximal humerus were not different in HU vs CON

AN

(p=0.785 and p=0.983, respectively). Cancellous vBMD was not different at the proximal humerus of

M

HU rats compared to CON (p=0.065). There were no differences in the cortical shell vBMD of the proximal humerus between HU and CON (p=0.617). Overall, HU resulted in decrements in bone mass

ED

and density in the unloaded distal femur, but no differences were observed from control in the weight-

PT

bearing humeri.

CE

3.3 Osteocyte sclerostin, IGF-I, and IL-6 changed in opposite directions in the unloaded hind-limb compared to the weight-bearing forelimb. %Sclerostin-positive osteocytes were higher in the distal

AC

femur in HU rats compared to CON (Fig. 2A; p=0.011; effect size=0.400), but lower in HU in the proximal humerus compared to CON (p=0.006; effect size=0.477). %IGF-I-positive osteocytes were lower in HU compared to CON in the distal femur (Fig. 2B; p=0.022; effect size=0.302), while %IGF-Ipositive osteocytes were higher in HU in the proximal humerus (p=0.028; effect size=0.342). In the distal femur in HU rats, %IL-6-positive osteocytes were lower compared to CON (Fig 2C; p=0.036; effect size=0.278), while %IL-6-positive osteocytes were higher in the proximal humerus in HU rats (p=0.015; effect size=0.399).

ACCEPTED MANUSCRIPT 3.4 Osteocyte density was lower in the unloaded distal femur, but not different in the proximal humerus. In the distal femur, osteocyte density per mm2 was not statistically lower in HU compared to CON (Fig. 2E; p=0.061) but when normalized to bone volume was significantly lower than CON (Fig. 2F; p=0.024, effect size=0.414). In the proximal humerus there was no difference in osteocyte density per mm2

T

(p=0.842) nor normalized to bone volume (p=0.515) between HU and CON.

IP

3.5 %TNF-α positive osteocytes were no different in HU vs. CON at either bone site. Osteocyte

CR

prevalence of TNF-α was no different from CON in the distal femur (Fig. 2D; p=0.207) or in the proximal humerus (p=0.878).

US

3.6 Osterix and cathepsin K showed opposite responses in %covered cancellous surfaces in the HU rats

AN

in distal femur compared to the proximal humerus. %Cathepsin K cancellous surface was higher in the distal femur (Fig. 3A; p=0.002; effect size=0.599) but lower in the proximal humerus in HU compared

M

to CON (p=0.003; effect size=0.541). %Osterix cancellous surface was lower in the distal femur (Fig.

ED

3B; p=0.004, effect size=0.551), but higher in the proximal humerus in HU compared to CON rats

PT

(p=0.028; effect size=0.343).

CE

4. Discussion

This is the first study to report on site-specific changes of mechanosensitive osteocyte proteins in

AC

response to weight-bearing and disuse that are consistent with increases or decreases, respectively, observed in osteoblast-covered surfaces. The alterations in osteocyte proteins suggest site-specific catabolic or anabolic environments within the same animal depending on mechanical stimuli. These responses appear to be TNF-α independent, given no observed change in osteocyte TNF-α at either the forelimb or hindlimb. Secondly, this study confirms the structural impact of local loss of mechanical signaling on bone as evidenced by decrements in hindlimb bone mass and density but no change in forelimb bone mass and density of HU rats.

ACCEPTED MANUSCRIPT Disuse-induced bone loss is associated with both a decline in bone formation and an increase in bone resorption. After 12 weeks of bedrest in healthy young adults, bone biopsies reveal 40% lower osteoblast surface and doubled osteoclast surface in cancellous bone (45). Numerous studies in rodents have shown lower bone formation rate and osteoid surface in HU (19, 35, 36) as well as decreased

T

osteoblast differentiation (46). HU studies show differing results in osteoclast number or surface with

IP

some showing increases (1) and some showing no changes (35, 36). In our unloaded rats, osteoid surface

CR

was 63% lower and cancellous osteoclast surface was 113% higher in the proximal tibia compared to ambulatory controls. In the unloaded distal femur, cathepsin K-positive cancellous surfaces (staining

US

osteoclasts) were 80% higher and osterix-positive cancellous surfaces (staining osteoblasts) were 60%

AN

lower than in ambulatory control rats. Concurrently, we found significantly lower bone mass and density in the unloaded distal femur of the unloaded rats compared to the ambulatory controls. The loss of

M

mechanical loading stimuli on the hindlimb bones resulted in bone loss due to increased resorption and

ED

decreased formation.

PT

In designing this study, we presumed the humerus would experience weight-bearing during hindlimb unloading similar to that borne by the humerus of ambulating control rats. We hypothesized

CE

osteocyte proteins in the unloaded femur, but not in the weight-bearing humerus, would be altered in HU

AC

rats compared to normally ambulating controls. While our hypothesis about the femur was correct, our data suggest HU humeri osteocytes are responding to a modulated loading environment distinctly different from corresponding control humeri. The unloaded femur demonstrated osteocyte protein and osterix changes consistent with an expected unloading paradigm compared to ambulating controls; however, we observed distinctly opposite changes in the HU humeri compared to the ambulating controls which are consistent with increased loading compared to ambulatory controls. Osteocytes in the femur exhibited higher sclerostin and lower IGF-I and IL-6 in response to the lack of mechanical loads,

ACCEPTED MANUSCRIPT while the humerus osteocytes had lower sclerostin and higher IGF-I and IL-6, suggesting increased local strain magnitude or altered strain distribution loading in the forelimbs. This surprising finding demonstrates that the signaling of osteocytes via IL-6, and IGF-I is a precise adaptation specific to the local bone site similar to that exhibited by sclerostin and that the humerus of an HU rat experiences

IP

T

loads different from the humerus in an ambulatory control. While the response of osteocyte proteins in the humerus and femur did not match our original

CR

hypothesis, osteocyte density data did; in the unloaded bone, osteocyte density normalized to bone

US

volume was approximately 30% lower in the distal femur compared to ambulatory control. This is consistent with previous research demonstrating lower osteocyte number and increased osteocyte

AN

apoptosis due to mechanical unloading (1, 35, 36). Osteocyte apoptosis is a signal to increase osteoclast

M

activity (1), and, in the distal femur of our HU rats, there was indeed an increase in cathepsin K-positive cancellous surface. In the weight-bearing proximal humerus, there was less than a 5% difference (n.s.) in

ED

osteocyte density between the HU and CON groups, but cathepsin K surface was lower in the unloaded

PT

rats than in the ambulatory control rats. Among the mechanisms proposed for osteocyte apoptosis are both altered shear/interstitial flow changes as well as TNF-α (3, 8). TNF-α alone induces osteocyte

CE

apoptosis in a dose-dependent manner (8). Additionally, reduced flow and shear forces elevate osteocyte

AC

apoptosis regardless of presence of TNF-α (8). In a fluid shear stress study in which mechanical bone loading was mimicked by applying pulsating fluid flow to cultured osteocytes, the TNF-α induced apoptosis observed in osteocytes was actually inhibited by mechanical loading (38). While we observed no changes in osteocytes positive for TNF-α, we do see a decline in osteocyte density in the unloaded bone, suggesting a potential role of shear/interstitial flow induction of apoptosis due to HU as observed by others (1, 35). A time course study including measurements of osteocyte apoptosis would fully

ACCEPTED MANUSCRIPT address this question. The fact that osteocytes are preserved in the humeri also supports the data showing no changes in bone mass and density in this site. Our data indicating increased or altered loading state in the humerus of HU rats are contrary to the view that the forelimbs are normally-loaded during HU. Using force plates under the cage floor with

IP

T

HU rats, Hargens et al. reported that 50% of the rat’s bodyweight was borne by the forelimbs (12). This experiment assumed the forelimbs carry 50% of body weight under normal (non-HU) conditions.

CR

However, this study did not include comparative force plate measurements of control animals. Other

US

studies have reported measurements of ground reaction forces in healthy rats. In one such study, young rats were placed in specially made cages restricting movement beyond turning. It was found that these

AN

rats, standing on all four limbs, placed 80% of their bodyweight on the hindlimbs (31). Based upon this,

M

50% bodyweight on the forelimbs in HU may represent increased loading compared to ambulatory controls. Furthermore, the dynamic nature of loading during ambulation and weight-bearing should also

ED

be considered. Both of these studies measured only static loads, which does not account for the

PT

frequency of loading, changes in the direction of loading, and postural changes during HU. A study that examined the vertical dynamic loads on the hind- and forelimbs of rats during normal ambulation found

CE

equally distributed bodyweight between the forelimbs and hindlimbs at approximately 50% bodyweight

AC

on each (10). In HU however, rats stabilize posture utilizing one forelimb at a time to propel themselves when walking, perhaps causing that one limb to experience enhanced dynamic loading when compared to normal ambulation on all four limbs. It thus remains unclear if external loading on the forelimbs is altered in HU or not. Regardless of the exact mechanism, however, our osteocyte immunohistochemical data demonstrating a decrease in sclerostin and increased IGF-I, IL-6, and osterix-positive surfaces provide biological evidence for altered loading or strain distributions in the forelimb of HU rats compared to ambulatory controls.

ACCEPTED MANUSCRIPT Studies examining sclerostin have demonstrated the local adaptation to specific load where ulnar loading in mice and rats resulted in declines in sclerostin compared to the contralateral unloaded ulna (28). This indicates the loading effects on sclerostin were not generalized to the osteocytes of the contralateral ulna. In this current study comparing ambulating controls to HU rats, we found opposite

T

responses of osteocyte sclerostin in the distal femur vs. the proximal humerus with 19% higher

IP

sclerostin-positive osteocytes in the distal femur of HU rats and 30% lower sclerostin in the proximal

CR

humerus. To our knowledge, this is the first study to look at sclerostin in the forelimb of HU rats. The decrease in osteocyte sclerostin in the weight-bearing humerus also suggests increased or an altered

US

mechanical environment sensed by the osteocytes in that region. The increase in osteocyte sclerostin in

AN

the femur was consistent with lower osterix-positive surface staining, while the decrease in sclerostin prevalence in the humerus was associated with higher osterix-positive staining. As sclerostin acts to

M

suppress bone formation, these changes in osterix staining, a molecular marker for osteoblasts, are

ED

biologically consistent with the changes in sclerostin. Contrary to previous findings (2, 21) from systemic inflammation models (i.e., diet-induced obesity and inflammatory bowel disease), sclerostin in

PT

this model of mechanical unloading/loading does not correlate with changes in TNF-α. Indeed, we

CE

observe here altered osteocyte IL-6 protein in combination with changes in sclerostin in line with observations of IL-6 family cytokines as negative regulators of SOST gene expression (41). This is also

AC

supported by our observed changes in IGF-I, also a known negative regulator of SOST (15). Osteocyte-derived IGF-I has been studied in the context of loading, but minimally addressed in the context of disuse or unloading. Mice with a conditional knockout of osteocyte Igf1 exhibit impaired upregulation of loading-induced genes, including those in the Wnt signaling pathway, after two weeks of four-point-bending of the tibia (15). In rats, four-point-bending of the tibia results in a 75% increase in IGF-I mRNA positive osteocytes compared to the contra-lateral unloaded tibia (27). Within 30 minutes

ACCEPTED MANUSCRIPT of axial loading of the caudal vertebrae in rats, IGF-I mRNA in osteocytes increases (18). In our study, we found 94% higher IGF-I-positive osteocytes in the humerus of HU rats compared to the humerus of normally ambulating controls. Based on the previous literature on osteocyte-derived IGF-I in conditions of loading, this provides further independent evidence of increased loading on the humerus in HU rats.

T

In the unloaded distal femur of HU rats there was 30% lower prevalence of IGF-I-positive osteocytes

IP

compared to controls. To our knowledge, this is the first study to examine osteocyte IGF-I in conditions

CR

of unloading. In the humerus, higher osteocyte IGF-I was matched by higher osterix surface while lower osteocyte IGF-I in the unloaded femur corresponded with lower osterix-staining. These data indicate

US

osteocyte IGF-I may be signaling to increase osteoblasts in cases of increased/altered loading, whereas

AN

in unloading, a lack of osteocyte IGF-I may be associated with lower osteoblast number and bone

M

formation.

The other osteocyte protein we measured in this study, IL-6, is less well understood. The role of

ED

IL-6 in bone physiology is complex in nature, as it can sometimes stimulate resorption and sometimes

PT

formation (33). In vitro results demonstrate that shear-loading on cultured osteocytes, used to simulate mechanical loading, stimulates osteocyte production of IL-6, which was proposed to communicate with

CE

osteoblasts (4). In addition, osteocytes induced to undergo apoptosis via TNF-α stimulation in vitro

AC

have been shown to release IL-6 as well (9). Since reduced shear loading on cultured osteocytes has been shown to cause apoptosis (3), osteocytes could conceivably release IL-6 in both conditions of increased and decreased loading. In our study, the percentage of IL-6-positive osteocytes was lower in the unloaded distal femur (25% lower than ambulatory controls) coincident with lower osterix-positive surface, but the percentage of IL-6-positive osteocytes was 48% higher in the humerus of HU rats corresponding with elevated osterix surfaces. Therefore, we hypothesize IL-6 is a mechanosensitive

ACCEPTED MANUSCRIPT osteocyte protein that, in conditions of unloading and loading seen in our animals, signals to osteoblasts to decrease or increase bone formation, respectively. We have previously observed changes in osteocyte IL-6 in animals with systemic inflammatory conditions (21); however, in that model both osteocyte IL-6 and TNF-α were elevated and correlated

IP

T

with lower bone formation and higher bone resorption. The function of IL-6 has been shown to be tightly linked with TNF-α, and its role can have different implications depending on the presence or

CR

absence of TNF-α. For example, in the context of skeletal muscle, IL-6 production post-exercise plays

US

significant roles in skeletal muscle metabolism and anti-inflammatory effects, and the increase in IL-6 generally does not correspond with increased TNF-α (25, 26). Given that in our study we see no increase

AN

in osteocytes positive for TNF-α in the humeri (coinciding with no change in osteocyte density), IL-6

M

could play an anabolic, anti-inflammatory role analogous to what is seen in skeletal muscle. Therefore, we believe in this model of unloading/loading osteocyte IL-6 is functioning in a primarily anabolic role

ED

due to no concurrent changes in TNF-α and similar changes in osteocyte IL-6 to IGF-I and osterix

PT

surfaces.

CE

For the purpose of this study, we chose to focus our measures on osteocyte proteins and their relation to osteoblasts. However, it must be noted that all of these osteocyte proteins have the ability to,

AC

directly or indirectly, trigger osteoclastogenesis and may play a role in osteoclast function (14, 22, 42). The main purpose of this study was to examine sclerostin, IL-6, and IGF-I in relation to osteoblasts and address whether these changes were associated with changes in TNF-α, which has clear roles in osteoclastogenesis and interactions with all three osteocyte proteins we focused on here (2, 34). Additionally, it must be noted that the key osteoclastogenesis regulator, receptor activator of nuclear factor κB ligand (RANKL), and its decoy receptor, osteoprotegerin (OPG), are osteocyte proteins with clear roles in mechanical loading/unloading (39, 43, 44). We hypothesize the robust increase in

ACCEPTED MANUSCRIPT osteoclast surface (measured by cathepsin K) is likely due in part to increases in osteocyte RANKL in the unloaded bone as previous studies indicate (44). Additionally, the suggested increase in mechanical loading on the HU humerus likely decreased osteocyte RANKL, contributing to lower cathepsin K in those bones compared to the ambulating controls. Further studies should address the interaction between

T

RANKL and OPG with sclerostin, IL-6, and IGF-I in conditions of loading and unloading as previous

IP

literature has highlighted interactions between these factors and the RANKL/OPG axis (29, 34, 42).

CR

Limitations of this study include not being able to obtain dynamic measures of bone formation

US

rate on both the unloading hindlimb and the forelimb. Due to constraints with the parent protocol, we were restricted to measuring osterix surface as a surrogate measure of osteoblasts. Therefore, we cannot

AN

confirm changes in bone formation rate in this study. Additionally, factors like IL-6 and IGF-I are

M

difficult to address due to many different cell types producing, secreting, and binding these factors. We cannot disregard what may be happening systemically especially with factors that are much more

ED

ubiquitous throughout systemic physiology. We restricted our analysis to osteocytes to focus on the

PT

signaling role of these important bone cells in unloading and loading conditions, as previous literature

CE

has confirmed osteocyte release of IGF-I and IL-6 as playing a role in mechanotransduction (4, 15, 27). In conclusion, we found that 28 days of hindlimb unloading in skeletally mature male rats

AC

resulted in bone loss, increased osteoclast surfaces, and decreased osteoblast surfaces in the unloaded hindlimb. These changes in the hindlimb corresponded with lower osteocyte IGF-I and IL-6 and higher osteocyte sclerostin. While the weight-bearing humerus of the HU rats had no statistical changes in bone mass or density, we measured lower osteocyte sclerostin and higher IGF-I and IL-6, contrary to our original hypothesis (Fig. 4). These changes in the distal femur and proximal humerus were consistent with decreased and increased, respectively, osterix-positive bone surfaces. Additionally, we propose that these changes in osteocyte sclerostin, IL-6, and IGF-I are independent of TNF-α. Therefore, our results

ACCEPTED MANUSCRIPT provide indirect evidence that the humerus of HU rats, contrary to previous opinion, may experience increased loading, or at least altered strain distributions as sensed by the osteocytes, during tail suspension interventions. Additionally, this study demonstrates a precise, local response of osteocytes to mechanical stimuli via the release of proteins to orchestrate changes in bone turnover in response to

T

those stimuli. Treatments that could influence osteocytes in specific regions could, therefore, provide

IP

targeted treatments for disuse-induced bone loss.

CR

Acknowledgements

US

This study was supported by NASA Space Biology #NNX13AM43G. JEB and SAN were supported via the NASA-National Space Biomedical Research Institute Predoctoral Fellowship by

AN

NASA Cooperative Agreement NCC 9-58. JPE was supported via the National Science Foundation

M

Award No. HRD-1612776. The authors thank Scott Lenfest, Jeremy Black, Jennifer Kosniewski,

ED

Cameron Schaefer, Will Reyna, and Coleman Leach for assistance with hindlimb unloading procedures and animal husbandry, Michael Junior for assistance with ex vivo pQCT scans and sclerostin

PT

immunohistochemistry, and Jennifer Kosniewski for rat HU artwork.

CE

Author roles: CEM performed histological analyses, ran statistical analyses, and wrote the manuscript. HAH designed the parent animal project from which the tissues were collected. JEB and JPE assisted

AC

with writing the manuscript and compiling and analyzing data. SAN assisted with immunostaining protocols and writing the manuscript. CEM, JEB, JPE, SAN, HAH, and SAB contributed to data interpretation, read and revised the manuscript, and approved the final version.

ACCEPTED MANUSCRIPT References 1. Aguirre JI, Plotkin LI, Stewart SA, Weinstein RS, Parfitt AM, Manolagas SC, Bellido T. Osteocyte apoptosis is induced by weightlessness in mice and precedes osteoclast recruitment and bone loss. J Bone Miner Res. 2006; 21(4): 605-615.

T

2. Baek K, Hwang HR, Park HJ, Kwon A, Qadir AS, Ko SH, Woo KM, Ryoo HM, Kim GS, Baek

IP

JH. TNF-α upregulates sclerostin expression in obese mice fed a high-fat diet. Cell Physiol.

CR

2014; 229: 640-650.

US

3. Bakker AD, Klein-Nulend J, Burger E. Shear stress inhibits while disuse promote osteocyte apoptosis. Biochem Biophys Res Commun. 2004; 320: 1163-1168.

AN

4. Bakker AD, Kulkarni RN, Klein-Nulend J, Lems WF. IL-6 alters osteocyte signaling toward

M

osteoblasts but not osteoclasts. J Dent Res. 2014; 93(4): 394-399. 5. Bloomfield SA, Allen MR, Hogan HA, Delp MD. Site- and compartment-specific changes in

ED

bone with hindlimb unloading in mature adult rats. Bone. 2002; 31: 149-157.

PT

6. Bonewald LF. The amazing osteocyte. J Bone Miner Res. 2011; 26(2): 229-238.

3(10): 7-15.

CE

7. Bonewald LF. Mechanosenstion and transduction in osteocytes. BoneKEy-Osteovision. 2006;

AC

8. Cheung W, Liu C, Tonelli-Zasarsky RML, Simmons CA, You L. Osteocyte apoptosis is mechanically regulated and induces angiogenesis in vitro. J Orthop Res. 2011; 29: 523-530. 9. Cheung W, Simmons CA, You L. Osteocyte apoptosis regulates osteoclast precursor adhesion via osteocytic IL-6 secretion and endothelial ICAM-1 expression. Bone. 2012; 50: 104-110. 10. Clarke KA. Differential fore- and hindpaw force transmission in the walking rat. Physiol Behav. 1995; 58(3); 415-419.

ACCEPTED MANUSCRIPT 11. Dempster DW, Compston JE, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR, Parfitt AM. Standard nomenclature, symbols, and units for bone histomorphometry: a 2012 update of the report of the ASMBR histomorphometry nomenclature committee. J Bone Miner Res 2013; 28(1): 1-16.

IP

tension in tail-suspended rats. Physiologist, Suppl. 1984; 27: S37.

T

12. Hargens AR, Steskal J, Johansson C, Tipton CM. Tissue fluid shift, forelimb loading, and tail

CR

13. Klein-Nulend J, Bakker AD. Osteocytes: Mechanosensors of bone and orchestrators of mechanical adaptation. Clinic Rev Bone Miner Metab. 2007; 5: 195–209.

US

14. Kudo O, Sabokbar A, Pocock A, Itonaga I, Fujikawa Y, Athanasou NA. Interleukin-6 and

AN

interleukin-11 support human osteoclast formation by a RANKL-independent mechanism. Bone. 2003; 32: 1-7.

M

15. Lau KHW, Baylink DJ, Zhou XD, Rodriguez D, Bonewald LF, Li Z, Ruffoni D, Muller R,

ED

Kesavan C, Sheng MHC. Osteocyte-derived insulin-like growth factor I is essential for determining bone mechanosensitivity. Am J Physiol Endocrinol Metab. 2013; 305: E271-E281.

PT

16. LeBlanc AD, Spector ER, Evans HJ, Sibonga JD. Skeletal responses to space flight and the bed

CE

rest analog: a review. J Musculoskelet Neuronal Interact. 2007; 7(1): 33-47. 17. LeBlanc AD, Schneider V, Shackelford L, West S, Oganov V, Bakulin A, Voronin L. Bone

AC

mineral and lean tissue loss after long duration space flight. J Musculoskel Neuron Interact. 2000; 1(2): 157-160. 18. Lean JM, Mackay AG, Chow JWM, Chambers TJ. Osteocytic expression of mRNA for c-fos and IGF-I: an immediate early gene response to an osteogenic stimulus. Am J Physiol. 1996; 270: E937-E945.

ACCEPTED MANUSCRIPT 19. Macias BR, Swift JM, Nilsson MI, Hogan HA, Bouse DS, Bloomfield SA. Simulated resistance training, but not alendronate, increases cortical bone formation rate and suppresses sclerostin during disuse. J Appl Physiol. 2012; 112(5): 918-925. 20. Mason DJ, Hillam RA, Skerry TM. Constitutive in vivo mRNA expression by osteocytes of β-

T

actin, osteocalcin, connexin-43, IGF-I, c-fos and c-jun, but no TNF-α and tartrate-resistant acid

IP

phosphatase. J Bone Miner Res. 1996; 11(3): 350-357.

CR

21. Metzger CE, Narayanan A, Zawieja DC, Bloomfield SA. Inflammatory bowel disease in a rodent model alters osteocyte protein levels controlling bone turnover. J Bone Miner Res. 2017; 32(4):

US

803-813.

AN

22. Mochizuki H, Hakeda Y, Wakatsuki N, Usui N, Akashi S, Sato T, Tanaka K, Kumegawa M. Insulin-like growth factor-I supports activation and formation of osteoclasts. Endocrinol. 1992;

M

131(3): 1075-1080.

ED

23. Morey-Holton ER, Globus RK. Hindlimb unloading of growing rats: a model for predicting skeletal changes during space flight. Bone. 1998; 22(5): 83S-88S.

PT

24. Morey-Holton ER, Globus RK. Hindlimb unloading rodent model: technical aspects. J Appl

CE

Physiol. 2002; 92: 1367-1377.

25. Paulson G, Mikkelsen UR, Raastad T, Peake JM. Leukocytes, cytokines and satellite cells: what

AC

role do they play in muscle damage and regeneration following eccentric exercise? Exerc Immunol Rev. 2012: 18(1); 42-97. 26. Pedersen BK, Febbraio MA. Muscle as an endocrine organ: focus on muscle-derived interleukin6. Phys Rev. 2008: 88(4); 1379-1406.

ACCEPTED MANUSCRIPT 27. Reijnders CMA, Bravenboer N, Tromp AM, Blankenstein MA, Lips P. Effect of mechanical loading on insulin-like growth factor-I gene expression in rat tibia. J Endocrinol. 2007: 192; 131-140. 28. Robling AG, Niziolek PJ, Baldridge LA, Condon KW, Allen MR, Alam I, Mantila SM, Gluhak-

T

Heinrich J, Bellido TM, Harris SE, Turner CH. Mechanical stimulation of bone in vivo reduces

IP

osteocyte expression of Sost/sclerostin. J Biol Chem. 2008; 283(9): 5866-5875.

CR

29. Rubin J, Ackert-Bicknell CL, Zhu L, Fan X, Murphy TC, Nanes MS, Marcus R, Holloway L, Beamer WG, Rosen CJ. IGF-I regulated Osteoprotegerin (OPG) and receptor activator of nuclear

US

Factor κB ligand in vitro and OPG in vivo. J Clin Endocrinol Metab. 2002; 87(9): 4273-4279.

AN

30. Scharla SH, Strong DD, Mohan S, Chevalley T, Linkhart TA. Effect of tumor necrosis factor-α on the expression of insulin-like growth factor I and insulin-like growth factor binding protein 4

M

in mouse osteoblasts. Eur J Endocrinol. 1994; 131: 293-301.

ED

31. Schött E, Berge OG, Ängeby-Möller K, Hammarström G, Dalsgaard CJ, Brodin E. Weight bearing as an objective measure of arthritic pain in the rat. J Pharmacol Toxicol Meth. 1994;

PT

31(2): 79-83.

CE

32. Sibonga JD, Evans HJ, Sung HG, Spector ER, Lang TF, Oganov VS, Bakulin AV, Shackelford LC, LeBlanc AD. Recovery of spaceflight-induced bone loss: bone mineral density after long-

AC

duration missions as fitted with an exponential function. Bone. 2007; 41: 973-978. 33. Sims NA. Cell-specific paracrine actions of IL-6 family cytokines from bone, marrow and muscle that control bone formation and resorption. Int J Biochem Cell Biol. 2016; 79: 14-23. 34. Steeve KT, Marc P, Sandrine T, Dominique H, Yannick F. IL-6, RANKL, TNF-alpha/IL-1: interrelations in bone resorption pathology. Cytokine Growth Factor Rev. 2004; 15: 49-60.

ACCEPTED MANUSCRIPT 35. Swift JM, Swift SN, Allen MR, Bloomfield SA. Beta-1 adrenergic agonist treatment mitigates negative changes in cancellous bone microarchitecture and inhibits osteocyte apoptosis during disuse. PLOS One. 2014; 9(9). 36. Swift JM, Swift SN, Nilsson MI, Hogan HA, Bouse SD, Bloomfield SA. Cancellous bone

IP

alendronate treatment. J Bone Miner Res. 2011; 26(9): 2140-2150.

T

formation response to simulated resistance training during disuse is blunted by concurrent

CR

37. Tan SD, Kuijpers-Jagtman AM, Semeins CM, Bronckers ALJJ, Maltha JC, Von den Hoff JW, Everts V, Klein-Nulend J. Fluid shear stress inhibits TNF-α induced osteocyte apoptosis. J Dent

US

Res. 2006; 85(10): 2006.

AN

38. ten Dijke P, Krause C, de Gorter DJJ, Lowick CWGM, van Bezooijen RL. Osteocyte-derived sclerostin inhibits bone formation: its role in bone morphogenetic protein and Wnt signaling. J

M

Bone Joint Surg. 2008; 90(1): 31-35.

ED

39. You L, Temiyasathit S, Lee P, Kim CH, Tummala P, Yao W, Kingery W, Malone AM, Kwon RY, Jacobs CR. Osteocytes as mechanosensors in the inhibition of bone resorption due to

PT

mechanical loading. Bone. 2008; 42: 172-179.

CE

40. Vico L, Collet P, Guignandon A, Lafage-Proust MH, Thomas T, Rehailia M, Alexandre C. Effects of long-term microgravity exposure on cancellous and cortical weight-bearing bones of

AC

cosmonauts. Lancet. 2000; 355: 1607-1611. 41. Walker EC, McGregor NE, Poulton IJ, Solano M, Pompolo S, Fernandes TJ, Constable MJ, Nicholson GC, Zhang JG, Nicola NA, Gillespie MT, Martin TJ, Sims NA. Oncostatin M promotes bone formation independently of resorption when signaling through leukemia inhibitory factor receptor in mice. J Clin Invest. 2010; 120: 582-592.

ACCEPTED MANUSCRIPT 42. Wijenayaka AR, Kogowa M, Lim HP, Bonewald LF, Findlay DM, Atkins GJ. Sclerostin stimulates osteocyte support of osteoclast activity by a RANKL-dependent pathway. PLOS One. 2011; 6(10): e25900. 43. Xiong J, O’Brien CA. Osteocyte RANKL: new insights into the control of bone remodeling. J

T

Bone Miner Res. 2012; 27(3): 499-505.

IP

44. Xiong J, Onal M, Jilka RL, Weinstein RS, Manolagas SC, O’Brien CA. Matrix-embedded cells

CR

control osteoclast function. Nature Med. 2011; 17(10): 1235-1241.

45. Zerwekh JE, Ruml LA, Gottschalk F, Pak CYC. The effects of twelve weeks of bed rest on bone

US

histology, biochemical markers of bone turnover, and calcium homeostasis in eleven normal

AN

subjects. J Bone Miner Res. 1998; 13(10): 1594-1601.

46. Zhang R, Supowit SC, Klein GL, Lu Z, Christensen MD, Lozano R, Simmons DJ. Rat tail

M

suspension reduces messenger RNA level for growth factors and osteopontin and decreases the

AC

CE

PT

ED

osteoblastic differentiation of bone marrow stromal cells. J Bone Miner Res. 1995; 10: 415-423.

ACCEPTED MANUSCRIPT Total vBMD

Total BMC

Cancellous vBMD

Femur

0.23%

0.41%

0.64%

Humerus

0.18%

0.28%

0.46%

Table 1. Coefficients of variation for pQCT measures of the distal femur and proximal humerus.

CON

HU

11.44 ± 0.9*

6.30 ± 0.8

615.05 ± 26.8

523.39 ± 36.0*

620.32 ± 27.7

620.29 ± 65.1

340.02 ± 47.2

281.04 ± 19.6*

222.39 ± 30.2

188.44 ± 31.3

1056.53 ± 38.5

1002.55 ± 50.8

1092.41 ± 32.2

1108.76 ± 72.3

CR

13.10 ± 0.8

US

Total BMC (g) Total vBMD (mg/cm3) Cancellous vBMD (mg/cm3) Cortical vBMD mg/cm3)

HU

IP

CON

Proximal Humerus

T

Distal Femur

6.17 ± 0.8

Table 2. pQCT measures of the metaphysis regions of the distal femur and proximal humerus. Data are

AN

represented as mean ± SD. *Indicates difference from CON. p<0.05.

M

Figure Legends

ED

Figure 1 – Cancellous histomorphometry of the unloaded proximal tibia. Data are represented as mean

PT

± SD. *Indicates difference from CON (p<0.05). A) Cancellous bone volume was lower in HU compared to CON (p=0.036). B) Osteoid surface was lower in HU compared to CON (p=0.003). C)

CE

Cancellous osteoclast surface was higher in HU versus CON (p=0.001). D) Representative image of cancellous osteoclast (left) and cancellous osteoid surface (right).

AC

Figure 2 – Immunohistochemistry of osteocytes in the cancellous bone distal femur and proximal humerus. Data are represented as mean ± SD. *Indicates difference from CON (p<0.05). A) %Sclerostin-positive osteocytes were higher in HU in the distal femur (p=0.011) but lower in HU in the proximal humerus (p=0.006). B) %IGF-I-positive osteocytes were lower in the distal femur in HU (p=0.022) but higher in the proximal humerus in HU (p=0.028). C) %IL-6-positive osteocytes were lower in the distal femur in HU (p=0.036) but higher in the proximal humerus in HU (p=0.015). D)

ACCEPTED MANUSCRIPT %TNF-α+ osteocytes were no different between CON and HU in either bone site. E) Osteocyte density per mm2 at the distal femur in HU was not statistically different than CON, and there were no differences in the proximal humerus. F) Osteocyte density normalized to bone volume was significantly lower in the HU distal femur compared to CON (p=0.024). There were no differences between CON and

T

HU at the proximal humerus.

IP

Figure 3 – Immunohistochemistry of surface proteins in the distal femur and proximal humerus. Data

CR

are represented as mean ± SD. *Indicates difference from CON (p<0.05). A) Cancellous surface positively stained for cathepsin K was higher in the distal femur in HU (p=0.002) and lower in the

US

proximal humerus in HU (p=0.003). B) Cancellous surface positively stained for osterix was lower in

AN

the distal femur in HU (p=0.004) but higher in the proximal humerus in HU (p=0.028). C) Representative image of cathepsin K staining in cancellous bone. D) Representative image of osterix

M

staining in cancellous bone.

AC

CE

PT

hindlimb and loaded forelimb.

ED

Figure 4 – Proposed mechanism for changes in osteocyte proteins and osteoblasts in the unloaded

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Highlights for Review: BONE-D-17-00283R1 Changes in this version of the manuscript - Addition of discussion of osteoclast changes and the likely contribution of RANKL/OPG axis - Addition of supplemental figure with immunohistochemistry representative images - Significant editing of entire manuscript and shortening of parts of the discussion