Osteocyte Viability and Regulation of Osteoblast ... - Wiley Online Library

315 downloads 2806 Views 2MB Size Report
Jun 1, 2004 - gradients within the bone pores (which drives fluid flow .... live/dead cell viability stain (Molecular Probes) consisting ... using a hard tissue microtome (SM2500S; Leica). ... data not shown). ... was lost on days 15 and 22.
JOURNAL OF BONE AND MINERAL RESEARCH Volume 19, Number 9, 2004 Published online on June 1, 2004; doi: 10.1359/JBMR.040516 © 2004 American Society for Bone and Mineral Research

Osteocyte Viability and Regulation of Osteoblast Function in a 3D Trabecular Bone Explant Under Dynamic Hydrostatic Pressure Erica Takai,1 Robert L Mauck,2 Clark T Hung,2 and X Edward Guo1

ABSTRACT: A new trabecular bone explant model was used to examine osteocyte– osteoblast interactions under DHP loading. DHP loading enhanced osteocyte viability as well as osteoblast function measured by osteoid formation. However, live osteocytes were necessary for osteoblasts to form osteoids in response to DHP, which directly show osteoblast– osteocyte interactions in this in vitro culture. Introduction: A trabecular bone explant model was characterized and used to examine the effect of osteocyte and osteoblast interactions and dynamic hydrostatic pressure (DHP) loading on osteocyte viability and osteoblast function in long-term culture. Materials and Methods: Trabecular bone cores obtained from metacarpals of calves were cleaned of bone marrow and trabecular surface cells and divided into six groups, (1) live cores ⫹ dynamic hydrostatic pressure (DHP), (2) live cores ⫹ sham, (3) live cores ⫹ osteoblast ⫹ DHP, (4) live cores ⫹ osteoblast ⫹ sham, (5) devitalized cores ⫹ osteoblast ⫹ DHP, and (6) devitalized cores ⫹ osteoblast ⫹ sham, with four culture durations (2, 8, 15, and 22 days; n ⫽ 4/group). Cores from groups 3– 6 were seeded with osteoblasts, and cores from groups 5 and 6 were devitalized before seeding. Groups 1, 3, and 5 were subjected to daily DHP loading. Bone histomorphometry was performed to quantify osteocyte viability based on morphology and to assess osteoblast function based on osteoid surface per bone surface (Os/Bs). TUNEL staining was performed to evaluate the mode of osteocyte death under various conditions. Results: A portion of osteocytes remained viable for the duration of culture. DHP loading significantly enhanced osteocyte viability up to day 8, whereas the presence of seeded osteoblasts significantly decreased osteocyte viability. Cores with live osteocytes showed higher Os/Bs compared with devitalized cores, which reached significant levels over a greater range of time-points when combined with DHP loading. DHP loading did not increase Os/Bs in the absence of live osteocytes. The percentage of apoptotic cells remained the same regardless of treatment or culture duration. Conclusion: Enhanced osteocyte viability with DHP suggests the necessity of mechanical stimulation for osteocyte survival in vitro. Furthermore, osteocytes play a critical role in the transmission of signals from DHP loading to modulate osteoblast function. This explant culture model may be used for mechanotransduction studies in long-term cultures. J Bone Miner Res 2004;19:1403–1410. Published online on June 1, 2004; doi: 10.1359/JBMR.040516 Key words:

trabecular bone, explant model, hydrostatic pressure, osteoblasts, osteocytes INTRODUCTION

ONE ADAPTS TO its mechanical environment so that its form follows function, a mechanism known as Wolff’s law, or bone adaptation.(1) Although the basic concepts of Wolff’s law have been generally accepted, the regulatory biochemical signaling pathways that mediate this adaptive process are unknown. It is widely believed that an understanding of bone adaptation mechanisms will contribute to the fundamental knowledge of bone formation, bone quality maintenance, and the etiology, prevention, and treatment of age-related fractures, as well as improvements in many orthopedic and dental implants.(2)

B

The authors have no conflict of interest.

Bone adaptation processes involve intimate coordination of three distinct bone cell types: osteocytes, osteoblasts, and osteoclasts. Osteocytes form gap junctions with osteoblasts(3,4) and are thought to play an important role in the coordination of osteoclast and osteoblast activities in bone adaptation. Furthermore, osteocytes form extensive interconnected 3D cellular networks that position them to be suitable sensors of changes in the local mechanical and hormonal environment in bone tissue. However, few studies have shown direct interactions between osteoblasts and osteocytes in vitro,(5) and no studies have examined these interactions with osteocytes in their native 3D environment. A variety of mechanical stimuli such as hydrostatic pressure, fluid shear, and load-induced strain have been shown

1 Bone Bioengineering Laboratory, Department of Biomedical Engineering, Columbia University, New York, New York, USA; Cellular Engineering Laboratory, Department of Biomedical Engineering, Columbia University, New York, New York, USA.

2

1403

1404

to elicit a response from osteocytes both in vitro and in vivo.(6 –9) Unlike other in vitro applications of mechanical stimuli such as membrane stretch, which can induce concomitant fluid flow-induced effects (e.g., shear stress, streaming potentials, nutrient convection), and fluid shear stress, which is driven by a pressure gradient, hydrostatic pressure can be applied uniformly to all cells without other modes of mechanical stimulation. Also, because the effect of hydrostatic pressure can be studied independently of deformation, there are minimal secondary effects from factors such as nutrient transport and streaming potentials. Hydrostatic pressure has also been shown to play an important role in bone mass maintenance in vivo. For example, with long-term bedrest, hydrostatic pressure is redistributed from the lower extremities to the skull, leading to a decrease in bone mass in the calcaneus and an increase in the skull.(10) Also, femoral vein ligation, which increases intramedullary pressure, has been found to alleviate loss of bone mass because of disuse.(11) Furthermore, load-induced deformation of bone tissue leads to a shift in pressure gradients within the bone pores (which drives fluid flow through the pores).(12,13) Although pore pressures within the lacunar-canalicular pores have yet to be determined experimentally, computational simulations estimate peak pore pressures to be as high as 2.2 MPa when bone is subjected to a step load of physiologic magnitude.(14) Despite the importance of hydrostatic pressure in bone mass maintenance and large changes in pressure to which osteocytes are subjected with normal activity, no studies address the effects of hydrostatic pressure on osteocytes in their native 3D lacunar-canalicular network. Unlike 2D in vitro cell studies, explant cultures of bone have the advantage of allowing osteocytes to maintain their 3D morphology and intercellular networks and to be surrounded by their native extracellular matrix environment. Explant cultures also simplify the complexities of in vivo studies, so that individual chemical and mechanical factors can be studied independently. Such bone explant culture models have been used to study osteocyte response to various chemical and mechanical stimuli in their native bone environment, both in cortical and trabecular bone.(15–18) A novel trabecular bone explant culture first developed by el Haj et al.(15) provides an important in vitro model for mechano-signal transduction. The culture model developed by el Haj et al.(15) consists of a trabecular bone cylinder cleaned of bone marrow and cultured for a maximum of 1 day. The advantages of this explant culture model include the ability to apply controlled mechanical stimuli, such as deformational, hydrostatic pressure, and fluid shear loads, and/or chemical stimuli to the explant while maintaining the natural 3D environment of the osteocytes. In this study, we have modified this trabecular explant culture model to include controlled seeding of surface cells such as osteoblasts or osteoclasts and have extended the culture period for up to 22 days. Our long-term objective is to use this modified trabecular bone explant model to examine the interaction of intact osteocytes with osteoblasts and osteoclasts, as well as to study mechano-signal transduction between bone cells when subjected to controlled mechanical or chemical stim-

TAKAI ET AL.

uli in long-term culture. Controlled seeding of surface cells facilitates the examination of osteoblast– osteocyte interactions with osteocytes in their native bone environment. This novel approach also provides a venue to explore tissueengineered interfaces between bone and cartilage.(19) The first objective of this study was to characterize our modified trabecular bone explant culture model in long-term culture, in terms of osteocyte viability and controlled surface cell seeding, under static culture conditions. This is a crucial experiment because no long-term trabecular bone explant culture studies have been previously performed. Furthermore, co-culture of those explants with osteoblasts has not yet been pursued. Therefore, the initial goal is critical in establishing this novel model for future mechanobiology studies of trabecular bone. The second goal was to determine osteocyte viability in this explant model when subjected to dynamic hydrostatic pressure (DHP) loading and/or the presence of seeded osteoblasts. The third aim was to examine the effects of osteocyte– osteoblast interactions and/or DHP loading on osteoblast function.

MATERIALS AND METHODS Bone core preparation The bone cores were divided into six groups as follows: (1) live cores ⫹ hydrostatic pressure loading, (2) live cores ⫹ sham, (3) live cores ⫹ seeded osteoblasts ⫹ hydrostatic pressure loading, (4) live cores ⫹ seeded osteoblasts ⫹ sham, (5) devitalized cores ⫹ seeded osteoblasts ⫹ hydrostatic pressure loading, and (6) devitalized cores ⫹ seeded osteoblasts ⫹ sham. Cylindrical trabecular bone cores 5 mm in diameter and 4 mm in height were harvested from the epiphyses of metacarpal bones of 3- to 4-month-old calves using a diamondtipped coring tool (Starlite Industries) and an Isomet lowspeed saw (Buehler) under sterile conditions. Tissue, ⬃4 mm from the articular surface and 5–7 mm from the marrow cavity, was removed from the end of the bone cores. The cores were cleaned of bone marrow and most trabecular surface cells with sterile PBS using a dental water pick. One-third of the cores (groups 5 and 6) were devitalized by repeated freeze-thaw. Complete devitalization of the cores was confirmed by the fact that all lacunae were empty in hematoxylin and eosin–stained sections of devitalized cores.

Seeding of primary osteoblasts Primary osteoblasts were obtained from trabecular bone chips harvested from the same region as the bore cores. Trabecular bone cores were obtained as described above, minced into small chips in Hank’s buffered salt solution, and rinsed several times. The bone chips were transferred to culture dishes with ␣-MEM supplemented with 10% FBS, and 1% penicillin/streptomycin. After ⬃1 week of culture, the first batch of cells that migrated out of the trabecular bone chips were discarded, because they are a more heterogeneous population. The bone chips were then moved to a new culture dish, where the second batch of osteoblasts were allowed to migrate out for approximately another 1–2 weeks, at which time they were used. All devitalized cores

BONE CELL INTERACTIONS UNDER HYDROSTATIC PRESSURE

1405

FIG. 1. (A) Custom-made cell seeder. (B) Custom feedback controlled bioreactor. (C) Bone cores sealed in a sterile plastic bag.

(groups 5 and 6) and one-half of the live cores (groups 3 and 4) were seeded the day after harvest with these primary osteoblasts, using a custom-made cell seeder, at an initial seeding density of ⬃5 ⫻ 105 cells/core (Fig. 1A). The cell seeder suspends the bone cores in a solution of osteoblasts at a concentration of 1 ⫻ 105 cell/ml, which is mixed with a magnetic stir bar. All cores were cultured in 6-well plates with 4 ml/core of ␣-MEM supplemented with 10% FBS and 1% penicillin/streptomycin. To characterize the cells on the surface of the trabecular bone of seeded cores, a long-term cell-labeling membrane dye (CM-DiI; Molecular Probes) was used to track seeded osteoblasts (on a subset of live bone cores not used in this study), according to manufacturer’s instructions. Primary osteoblasts were stained with the membrane dye and seeded on live bone cores on day 1, and these seeded cores were cultured for 8 days. On days 2 and 8, the bone cores were additionally stained with calcein-AM (Molecular Probes), which dyes all live cells, including osteocytes. The cores were imaged using an Olympus confocal microscope system at 10-␮m increments for 160 ␮m, and these images were superimposed.

DHP loading DHP loading was applied using a custom feedbackcontrolled bioreactor(20) (Fig. 1B) to groups 1, 3, and 5. The device is comprised of a 63-mm-inner-diameter stainless steel pressure vessel with a pressure transducer connected to a stainless steel piston. The piston is driven by an air cylinder controlled by double-acting solenoid valves in line with a compressed air source, which results in a 64-fold pressure amplification, allowing rapid pressurization of the vessel. A custom LabView program (National Instruments) was written using feedback control to regulate the peak pressure and frequency of pressurization. The uniformity and magnitude of pressurization within the vessel has been verified using the built-in pressure transducer and calibrated contact film. For the duration of mechanical loading, the bone cores were placed in sterile plastic bags, two to six

cores per bag, with 8 ml of supplemented ␣-MEM per bag, air bubbles were removed, and the bags were sealed with a bag sealer (Fig. 1C). The bags were placed in the pressure vessel filled with distilled water. All air was evacuated from the pressure vessel and tightly sealed. Loaded cores were subjected to a peak 3-MPa load at 0.33 Hz with a triangle waveform for 1 h/day, starting on day 2. Sham-loaded cores (groups 2, 4, and 6) were similarly sealed in plastic bags but were not placed inside the pressure chamber. All experiments were performed at 37°C. Because the specimens were sealed in the plastic bags with no air pockets, the amount of dissolved gases was minimally affected by pressurization. Bone cores were cultured in 6-well plates, except during the daily pressure loading duration, with media exchange every other day.

Histology Bone cores were harvested on days 2, 8, 15, and 22 (n ⫽ 4). At each time-point, cores were vertically cut in half. One-half of each of the cut cores was stained using a live/dead cell viability stain (Molecular Probes) consisting of calcein-AM (live) and ethidium homodimer-1 (dead), and imaged using a confocal microscope at 10-␮m increments for 160 ␮m. The images were superimposed into one image. The specimens were fixed in a 10% neutral buffered formalin solution. These specimens were decalcified in buffered formic acid, paraffin embedded, and sectioned at 10 ␮m thickness. To quantify the number of live osteocytes based on cell morphology, nonconsecutive sections of live cores (groups 1– 4) were stained with hematoxylin and eosin (H&E), where osteocytes with darkly stained round to oval intact nuclei and smooth margins were considered live and those with condensed, fragmented, or pale indistinct nuclei were considered dead.(21,22) An area 400 ␮m wide by 4000 ␮m tall in the center of each section was analyzed under a light microscope with a 20⫻ objective, using the bone histomorphometry software OsteoMeasure (4.00c; Osteometrics). The number of live osteocytes was normalized to the bone

1406

TAKAI ET AL.

FIG. 2. Cell labeling membrane stain on seeded osteoblasts. Red/yellow, seeded osteoblasts; green, remaining surface cells not removed by water pick. At day 2, most cells on the trabecular surface are seeded osteoblasts. At day 8, the majority of the cells on the surface are still the seeded osteoblasts. Note that osteocytes also appear green.

area within the measured region to give the number of live osteocytes per bone area. To detect apoptotic cells, TUNEL staining was performed using the ApopTag apoptosis detection kit (Serologicals) on nonconsecutive sections of live cores (groups 1– 4). The sections were counterstained with methyl green. DNasetreated bone sections were used as positive controls, and samples without terminal deoxynucleotidyl transferase were used as negative controls. An area 400 ␮m wide by 4000 ␮m tall in the center of each section was analyzed under 20⫻ magnification to obtain the number of TUNELpositive osteocytes. The number of necrotic osteocytes was also counted using the same morphological criteria as dead osteocytes in H&E-stained sections, but excluding TUNELstained cells. The number of TUNEL-positive osteocytes was divided by the total number of dead osteocytes (i.e., TUNEL-stained plus necrotic cells) to obtain the percentage of apoptotic osteocytes per total dead osteocytes. The other half of each specimen was fixed in a 10% neutral buffered formalin solution, embedded undecalcified in methylmethacrylate, and sectioned at 8 ␮m thickness using a hard tissue microtome (SM2500S; Leica). Nonconsecutive sections of all bone cores were stained with Goldner’s trichrome to measure the osteoid surface and the bone surface using OsteoMeasure. One-half of each section was analyzed, and the osteoid surface measurements were normalized to the bone surface in the analyzed region to obtain the osteoid surface per bone surface (Os/Bs). For TUNEL-, H&E-, and Goldner’s trichrome–stained sections, three sections per specimen were analyzed in a blind fashion and averaged to obtain the values for each specimen. To determine statistical significance, two-way ANOVA with Fisher’s posthoc analysis was performed (Systat), where p ⱕ 0.05 was considered significant.

RESULTS Osteoblast seeding Images of bone cores seeded with primary osteoblasts stained with membrane dye showed that, at day 2, nearly all of the cells on the surface were the seeded osteoblasts (Fig. 2, red/yellow). At day 8, the majority of the cells on the

surface were still seeded osteoblasts, rather than repopulation of surface cells that were not removed by washing. It should be noted that osteocytes also appear (green) in this cell staining protocol.

Osteocyte viability Qualitative live/dead cell viability staining indicated many live osteocytes at day 2, with many cells on the surface of the seeded cores and very few on the surface of unseeded cores (Figs. 3A and 3B). By day 8, substantially fewer live osteocytes could be seen for all groups, with many live cells on the trabecular surface (Figs. 3C–3F). Interestingly, many cells repopulated trabecular surfaces even in the unseeded cores by day 8. Live cores subjected to DHP loading without seeded osteoblasts (Fig. 3D, group 1) showed the most live osteocytes, whereas cores subjected to sham loading and seeded with osteoblasts showed the least (Fig. 3E, group 4). For day 15 or longer, the surface cells dominated the images and hindered clear visualization of the osteocytes, which reside inside bone tissue (images not shown). More quantitative morphological assessments of osteocyte viability by H&E-stained sections showed, in general, a decrease in osteocyte viability with time, which was stabilized between days 15 and 22 (Fig. 4). It is interesting to note that a good portion (⬃30%) of osteocytes remained viable even after 22 days in culture. Up to day 8, DHP loading significantly attenuated osteocyte death (p ⱕ 0.008), whereas its effect was negligible for longer culture periods (days 15 and 22; Figs. 4A and 4B). Seeding osteoblasts, in general, significantly reduced osteocyte viability (p ⱕ 0.05) up to day 15. The interaction between pressure loading and the presence of seeded osteoblasts resulted in some intriguing effects on osteocyte viability. With pressure loading, significant differences in osteocyte viability between seeded and unseeded cores were observed at both days 8 and 15, whereas without pressure loading, the significant difference was detected only at day 15 (Figs. 4C and 4D). Also, without seeded osteoblasts, the application of DHP actually maintained osteocyte viability between days 2 and 8 (Fig. 4B).

BONE CELL INTERACTIONS UNDER HYDROSTATIC PRESSURE

1407

FIG. 3. Live/dead viability stain. Green, live; red, dead cells. Live cores, day 2, (A) with and (B) without seeded osteoblasts. (B) Very few cells remain on the surface of unseeded cores after PBS wash, indicated by arrowheads. Day 8 DHP loaded (C) with and (D) without seeded osteoblasts. Day 8 sham loaded (E) with and (F) without seeded osteoblasts. OCY, osteocytes.

These data obtained from H&E-stained sections were in agreement with the qualitative assessment of osteocyte viability using live/dead staining.

Mode of osteocyte death: TUNEL Positive and negative controls confirmed that the TUNEL assay was functional. The percentage of TUNELpositive cells of the total number of dead osteocytes was ⬃50% for the duration of the experiment, regardless of the treatment group or time-point of sampling (p ⬎ 0.05; data not shown). Therefore, although the total number of dead osteocytes was different among different treatment groups, one-half of the dead osteocytes were apoptotic for all groups.

Osteoblast function The presence of live osteocytes, in general, significantly increased the percentage of osteoid surface (p ⱕ 0.04), and this increase became larger with time (Figs. 5A and 5B). DHP loading resulted in a trend of increased osteoblast function, as indicated by the osteoid surface (Fig. 5C). Without live osteocytes, the application of DHP did not affect osteoblast function (Fig. 5D). In addition, there were interactions between the application of DHP and the presence of live osteocytes. With applied pressure loading, significant differences in osteoid surface between live and devitalized cores were observed throughout the culture period (days 8 –22), whereas the significant difference was

1408

TAKAI ET AL.

FIG. 6. Osteoid surface. (A) Live cores with vs. without seeded osteoblasts, with DHP (*p ⫽ 0.004, ⫹p ⱕ 0.02 with day2, ⫹⫹p ⱕ 0.04 with all seeded). (B) Live cores with vs. without seeded osteoblasts, with sham load (⫹p ⱕ 0.02,⫹⫹p ⫽ 0.0004). OCY, osteocytes. Results are expressed as means ⫾ SD.

FIG. 4. Osteocyte viability assessed by morphology. (A) Live cores DHP vs. sham load with seeded osteoblasts (*p ⫽ 0.008, ⫹p ⱕ 0.04 with all). (B) Live cores DHP vs. sham load without seeded osteoblasts (*p ⫽ 0.0008, ⫹p ⱕ 0.00007 with all except DHP day 8). (C) Live cores with vs. without seeded osteoblasts with DHP (*p ⫽ 0.05, **p ⫽ 0.03). (D) Live cores with vs. without seeded osteoblasts with sham load (*p ⫽ 0.009). Results are expressed as means ⫾ SD.

only noticed at day 22 without pressure loading (sham loading; Figs. 5A and 5B). Bone cores with live osteocytes seeded with osteoblasts showed significantly more osteoid surfaces (p ⫽ 0.004) compared with live bone cores without seeded osteoblasts at day 8 (Fig. 6A) when subjected to DHP. This significance was lost on days 15 and 22. On the other hand, there was no difference in osteoid formation between live cores that were seeded or not seeded with osteoblasts in the absence of pressure loading (Fig. 6B) at any time-point.

DISCUSSION

FIG. 5. Osteoid surface. (A) Live vs. devitalized cores with DHP load (*p ⫽ 0.01, **p ⫽ 0.001, ***p ⫽ 0.00003). (B) Live vs. devitalized cores with sham load (*p ⫽ 0.04). (C) Sham vs. DHP loading of live cores with seeded osteoblasts (⫹p ⫽ 0.0004, ⫹⫹p ⱕ 0.04 with all DHP loaded). (D) Sham vs. DHP loading of devitalized cores with seeded osteoblasts (⫹p ⫽ 0.04). OCY, osteocytes. Results are expressed as means ⫾ SD.

In this study, we have characterized an explant culture model of trabecular bone with the ability to culture live osteocytes, and with controlled surface cell seeding, under static culture conditions. This trabecular bone explant model was used to examine the effects of hydrostatic pressure loading and osteoblast– osteocyte interactions on osteocyte viability, as well as osteoblast function. In this explant culture, both osteoblasts and some osteocytes can be kept alive for an extended period of time (22 versus 1 day) under static culture conditions. Also, we have showed that surface cells, such as osteoblasts, can be seeded onto the surfaces of the trabecular bone to examine osteoblast– osteocyte interactions. A membrane dye used to track seeded osteoblasts showed that, at days 2 and 8, nearly all cells on the trabecular surface were indeed the seeded osteoblasts, showing that most cells on the surface of the bone were derived from the seeded osteoblasts rather than repopulation of cells that were not completely removed from washing with the water pick. Osteocyte viability was decreased by the presence of seeded osteoblasts with both DHP and sham-loading. The presence of seeded osteoblasts may decrease osteocyte viability by decreasing nutrient and waste exchange to the osteocytes, because the osteoblasts proliferate to cover the entire trabecular surface between days 8 and 15 of culture. The decrease was significant at both days 8 and 15 for pressure-loaded cores compared with only day 15 for shamloaded cores. In our study, we have shown that dynamic pressure loading increases osteoblast functions, as indicated

BONE CELL INTERACTIONS UNDER HYDROSTATIC PRESSURE

by the significant increase in osteoid surface in the presence of live osteocytes (Fig. 6A). This induced increase in osteoblast activity should result in increased consumption of nutrients, and therefore, further reduce nutrient delivery to osteocytes inside bone tissue.(15,16,23,24) Thus, the greater difference in osteocyte viability between seeded and unseeded cores subjected to DHP compared with sham loading may be caused by exacerbated nutrient deficiencies of osteocytes. Interestingly, skeletal unloading has been found to induce osteocyte hypoxia in vivo, which can be reversed with brief deformational loading.(25) The results from this study suggest that the usage of this co-culture model for relatively long-term studies (⬎8 days) is limited by the nutrient deficiency of the osteocytes, which needs to be addressed in future studies. Perfusion of the bone cores may alleviate nutrient deficiencies caused by the increase in number of osteoblasts with culture time and mechanical stimulation.(26,27) DHP significantly enhanced osteocyte viability both with and without seeded osteoblasts, but only up to 8 days. This effect of pressure loading is in agreement with previous studies using long bone (cortical bone) explant cultures and in vivo models, which showed higher osteocyte viability with deformational loading.(28,29) These studies also support the idea that osteocytes require mechanical stimuli to survive. Because hydrostatic pressure is a mechanical stimulus that does not incur enhancement of nutrient transport through convection, an increase in osteocyte viability with DHP loading suggests that mechanical stimuli alone can enhance osteocyte viability. It is also important to note that there was no change in soluble gas concentrations in the culture medium caused by pressurization, because the specimens were sealed in fluid-filled bags. Therefore, it is unlikely that the increased osteocyte viability in the loaded cores was caused by increased gas exchange. Recently, an analytical model of nutrient transport through the lacunarcanalicular pores has predicted that diffusive and loadinduced transport of nutrients alone may be insufficient to maintain osteocyte viability, and therefore, load-induced active transport of nutrients by osteocytes may be necessary to supplement nutrient transport.(30) Their theory is consistent with our observation that DHP keeps osteocytes alive. TUNEL staining showed that the percentage of apoptotic osteocytes of the total number of dead osteocytes was the same regardless of treatment or culture duration. This indicates that, although the overall osteocyte viability is affected by DHP loading and the presence of seeded osteoblasts, the mode of osteocyte death is unaffected by these factors. Osteoblast function measured by osteoid surface per bone surface was significantly increased by the presence of live osteocytes in seeded cores. This suggests that there is intercellular communication between osteocytes and osteoblasts in this explant culture. This increased osteoblast function was significantly enhanced by the application of pressure loading. However, dynamic hydrostatic loading did not increase osteoblast function in devitalized bone cores seeded with osteoblasts compared with sham. Taken together, these in vitro results provide direct evidence that osteocytes play an important role in sensing DHP loading and in coordinating subsequent osteoblast functions.

1409

Several 2D in vitro studies have shown that DHP loading of various magnitudes elicits responses from osteoblasts such as increased calcium signaling and increased expression of c-fos, osteopontin, collagen, and alkaline phosphatase mRNA, even in the absence of osteocytes.(31–35) From these studies, DHP loading has been shown to promote an osteoblastic phenotype, but not direct osteoblast bone formation activities (such as mineralization). In this study, we have shown that the presence of osteocytes and hydrostatic pressure loading work in concert to increase osteoblast anabolic (bone formation) activities. Therefore, hydrostatic pressure may be simultaneously stimulating osteocytes to communicate with osteoblasts as well as to promote osteoblast maturation. One limitation of this explant culture model is the repopulation of the trabecular surface with the very few surface cells that were not removed by vigorous flushing with PBS in bone cores not seeded with osteoblasts by 8 –15 days of culture (Figs. 3B, 3D, and 3F). Nevertheless, the addition of seeded osteoblasts on live cores significantly enhanced osteoid surface per bone surface compared with that of live cores without seeded osteoblasts (Fig. 6A). A chemical method of removing the surface cells such as soaking the cores in trypsin may improve the removal of remaining surface cells. It should be emphasized that the majority of surface cells in the seeded cores came from controlled seeding of osteoblasts rather than repopulation of existing bone surface cells. Under the current culture conditions, although a good portion of osteocytes survived for a long period in this in vitro culture (⬎15 days), the apparent nutrition deficiency– dependent osteocyte death limits this in vitro model’s applicability in longer culture studies. This limitation may be easily rectified by media perfusion of cultured bone cores.(26,36) Despite this limitation, the new co-culture model of trabecular bone explants has revealed some interesting findings regarding mechanotransduction mechanisms in bone. With the caveat that the data are robust up to day 8, our finding suggests that DHP loading keeps osteocytes alive and enhances osteoblast function through osteocyte– osteoblast interactions. It would be of interest to study the effect of deformational loading,(36) as well as to examine osteoclast– osteocyte interactions in this explant culture.

ACKNOWLEDGMENTS This work is supported by National Institutes of Health Grants AR48287 and AR46568. The authors thank Michelle Huang and Garrett Kinnebrew and thank Dr Helen H Lu for technical advice.

REFERENCES 1. Wolf J 1892 Das Gesetz der Transformation der Knochen. Hirchwild, Berlin, Germany. 2. Burr DB, Martin RB 1989 Errors in bone remodeling: Toward a unified theory of metabolic bone disease. Am J Anat 186:186 –216. 3. Marotti G, Ferretti M, Muglia MA, Palumbo C, Palazzini S 1992 A quantitative evaluation of osteoblast-osteocyte relationships on growing endosteal surface of rabbit tibiae. Bone 13:363–368. 4. Doty SB 1981 Morphological evidence of gap junctions between bone cells. Calcif Tissue Int 33:509 –512.

1410 5. Yellowley CE, Li Z, Zhou Z, Jacobs CR, Donahue HJ 2000 Functional gap junctions between osteocytic and osteoblastic cells. J Bone Miner Res 15:209 –217. 6. Cheng B, Kato Y, Zhao S, Luo J, Sprague E, Bonewald LF, Jiang JX 2001 PGE(2) is essential for gap junction-mediated intercellular communication between osteocyte-like MLO-Y4 cells in response to mechanical strain. Endocrinology 142:3464 –3473. 7. Inaoka T, Lean JM, Bessho T, Chow JW, Mackay A, Kokubo T, Chambers TJ 1995 Sequential analysis of gene expression after an osteogenic stimulus: C-fos expression is induced in osteocytes. Biochem Biophys Res Commun 217:264 –270. 8. Klein-Nulend J, van der Plas A, Semeins CM, Ajubi NE, Frangos JA, Nijweide PJ, Burger EH 1995 Sensitivity of osteocytes to biomechanical stress in vitro. FASEB J 9:441– 445. 9. Pitsillides AA, Rawlinson SC, Suswillo RF, Bourrin S, Zaman G, Lanyon LE 1995 Mechanical strain-induced NO production by bone cells: A possible role in adaptive bone (re)modeling? FASEB J 9:1614 –1622. 10. Turner CH 1999 Site-specific skeletal effects of exercise: Importance of interstitial fluid pressure. Bone 24:161–162. 11. Bergula AP, Huang W, Frangos JA 1999 Femoral vein ligation increases bone mass in the hindlimb suspended rat. Bone 24:171– 177. 12. Weinbaum S, Cowin SC, Zeng Y 1994 A model for the excitation of osteocytes by mechanical loading-induced bone fluid shear stresses. J Biomech 27:339 –360. 13. Knothe Tate ML, Knothe U, Niederer P 1998 Experimental elucidation of mechanical load-induced fluid flow and its potential role in bone metabolism and functional adaptation. Am J Med Sci 316:189 –195. 14. Zhang D, Weinbaum S, Cowin SC 1998 Estimates of the peak pressures in bone pore water. J Biomech Eng 120:697–703. 15. el Haj AJ, Minter SL, Rawlinson SC, Suswillo R, Lanyon LE 1990 Cellular responses to mechanical loading in vitro. J Bone Miner Res 5:923–932. 16. Dallas SL, Zaman G, Pead MJ, Lanyon LE 1993 Early strainrelated changes in cultured embryonic chick tibiotarsi parallel those associated with adaptive modeling in vivo. J Bone Miner Res 8:251–259. 17. Lozupone E, Favia A, Grimaldi A 1992 Effect of intermittent mechanical force on bone tissue in vitro: Preliminary results. J Bone Miner Res 7:S2;S407–S409. 18. Rawlinson SC, Mosley JR, Suswillo RF, Pitsillides AA, Lanyon LE 1995 Calvarial and limb bone cells in organ and monolayer culture do not show the same early responses to dynamic mechanical strain. J Bone Miner Res 10:1225–1232. 19. Hung CT, Lima E, Mauck RL, Takai E, LeRoux MA, Lu HH, Stark RG, Guo XE, Ateshian GA 2003 Anatomically shaped osteochondral constructs for articular cartilage repair. J Biomech 36:1853–1864. 20. Mauck RL, Soltz MA, Ateshian GA, Hung CT 2001 Dynamic hydrostatic pressurization increases matrix gene expression by chondrocytes in 3D culture. Adv Bioeng 51:23147. 21. Usui Y, Kawai K, Hirohata K 1989 An electron microscopic study of the changes observed in osteocytes under ischemic conditions. J Orthop Res 7:12–21. 22. Hamaya M, Mizoguchi I, Sakakura Y, Yajima T, Abiko Y 2002 Cell death of osteocytes occurs in rat alveolar bone during experimental tooth movement. Calcif Tissue Int 70:117–126.

TAKAI ET AL. 23. Rawlinson SC, Pitsillides AA, Lanyon LE 1996 Involvement of different ion channels in osteoblasts’ and osteocytes’ early responses to mechanical strain. Bone 19:609 – 614. 24. Skerry TM, Bitensky L, Chayen J, Lanyon LE 1989 Early strainrelated changes in enzyme activity in osteocytes following bone loading in vivo. J Bone Miner Res 4:783–788. 25. Dodd JS, Raleigh JA, Gross TS 1999 Osteocyte hypoxia: A novel mechanotransduction pathway. Am J Physiol 277:C598 –C602. 26. Cartmell S, Porter B, Garcia A, Guldberg R 2003 Effects of medium perfusion rate on cell–seeded three– dimensional bone constructs in vitro. Tissue Eng 9:1197–1204. 27. Mueller SM, Mizuno S, Gerstenfeld LC, Glowacki J 1999 Medium perfusion enhances osteogenesis by murine osteosarcoma cells in three-dimensional collagen sponges. J Bone Miner Res 14:2118 – 2126. 28. Noble BS, Peet N, Stevens HY, Brabbs A, Mosley JR, Reilly GC, Reeve J, Skerry TM, Lanyon LE 2003 Mechanical loading: Biphasic osteocyte survival and targeting of osteoclasts for bone destruction in rat cortical bone. Am J Physiol Cell Physiol 284: C934 –C943. 29. Lozupone E, Palumbo C, Favia A, Ferretti M, Palazzini S, Cantatore FP 1996 Intermittent compressive load stimulates osteogenesis and improves osteocyte viability in bones cultured “in vitro.” Clin Rheumatol 15:563–572. 30. Petrov N, Pollack SR 2003 Comparative analysis of diffusive and stress induced nutrient transport efficiency in the lacunarcanalicular system of osteons. Biorheology 40:347–353. 31. Brighton CT, Fisher JR Jr, Levine SE, Corsetti JR, Reilly T, Landsman AS, Williams JL, Thibault LE 1996 The biochemical pathway mediating the proliferative response of bone cells to a mechanical stimulus. J Bone Joint Surg Am 78:1337–1347. 32. Nagatomi J, Arulanandam BP, Metzger DW, Meunier A, Bizios R 2001 Frequency- and duration-dependent effects of cyclic pressure on select bone cell functions. Tissue Eng 7:717–728. 33. Klein-Nulend J, Roelofsen J, Semeins CM, Bronckers AL, Burger EH 1997 Mechanical stimulation of osteopontin mRNA expression and synthesis in bone cell cultures. J Cell Physiol 170:174 –181. 34. Glanstchnig H, Varga F, Rumpler M, Klaushofer K 1996 Prostacyclin (PGI2): A potential mediator of c-fos expression induced by hydrostatic pressure in osteoblastic cells. Eur J Clin Invest 26: 544 –548. 35. Roelofsen J, Klein-Nulend J, Burger EH 1995 Mechanical stimulation by intermittent hydrostatic compression promotes bonespecific gene expression in vitro. J Biomech 28:1493–1503. 36. Jones DB, Broeckmann E, Pohl T, Smith EL 2003 Development of a mechanical testing and loading system for trabecular bone studies for long term culture. Eur Cells Mater 5:48 – 60.

Address reprint requests to: X Edward Guo, PhD 351 Engineering Terrace, MC 8904 Columbia University New York, NY 10027, USA E-mail: [email protected] Received in original form February 6, 2004; in revised form April 2, 2004; accepted May 5, 2004.