Research ArticleCELL BIOLOGY

Dysregulation of ectonucleotidase-mediated extracellular adenosine during postmenopausal bone loss

See allHide authors and affiliations

Science Advances  21 Aug 2019:
Vol. 5, no. 8, eaax1387
DOI: 10.1126/sciadv.aax1387

Abstract

Adenosine and its receptors play a key role in bone homeostasis and regeneration. Extracellular adenosine is generated from CD39 and CD73 activity in the cell membrane, through conversion of adenosine triphosphate to adenosine monophosphate (AMP) and AMP to adenosine, respectively. Despite the relevance of CD39/CD73 to bone health, the roles of these enzymes in bona fide skeletal disorders remain unknown. We demonstrate that CD39/CD73 expression and extracellular adenosine levels in the bone marrow are substantially decreased in animals with osteoporotic bone loss. Knockdown of estrogen receptors ESR1 and ESR2 in primary osteoprogenitors and osteoclasts undergoing differentiation showed decreased coexpression of membrane-bound CD39 and CD73 and lower extracellular adenosine. Targeting the adenosine A2B receptor using an agonist attenuated bone loss in ovariectomized mice. Together, these findings suggest a pathological association of purine metabolism with estrogen deficiency and highlight the potential of A2B receptor as a target to treat osteoporosis.

INTRODUCTION

Emerging studies suggest the pivotal role played by naturally occurring purinergic nucleoside adenosine and its signaling in bone tissue formation, function, and repair (13). There has been a surge in research activity to understand the role of adenosine receptors, including the A1 receptor (A1R), A2AR, and A2BR, in bone tissue formation and maintenance (15). Studies have shown that A2AR and A2BR activation promotes osteogenic differentiation of osteoprogenitors (2, 3, 6, 7) and inhibits osteoclastogenesis (6, 810). Recently, we have shown that extracellular phosphate uptake by the SLC20a1 phosphate transporter on the cell membrane supports osteogenesis of mesenchymal stem cells (MSCs) via adenosine, which acts as an autocrine/paracrine signaling molecule through the A2BR (7). Activation of the A2BR also inhibited adipogenesis of human MSCs (11). These findings suggest the possibility of dysregulation of extracellular adenosine and its signaling during bone disorders.

CD39 (ectonucleoside triphosphate diphosphohydrolase-1) and CD73 (ecto-5′-nucleotidase) are membrane-bound ectonucleotidases that regulate extracellular adenosine by hydrolyzing extracellular adenosine triphosphate to adenosine diphosphate and adenosine monophosphate (AMP) and AMP to adenosine, respectively (12, 13). These ectonucleotidases are well known for their immunosuppressive functions (14) and affect a variety of pathophysiological events, including but not limited to autoimmune diseases (15, 16), infections (17, 18), atherosclerosis (19, 20), ischemia-reperfusion injury (21, 22), cancer (23), and transplant tolerance (14). Recent studies have demonstrated the importance of these enzymes in bone health: Mice lacking CD73 develop osteopenia with impaired osteoblast function (24) and the activity of CD73 in osteoblasts is essential to bone repair in aged mice (25). Despite these studies implying the role of ectoenzymes on bone tissue formation and osteogenic commitment of progenitor cells (24, 26), their role in bona fide skeletal disorders such as postmenopausal osteoporosis remains unknown.

Osteoporosis is a condition of severe bone loss affecting 10 million individuals above 50 years of age and inflicting 2 million fractures each year (27, 28). Dynamic bone remodeling, dictated by the balance between osteoblast and osteoclast functions that contribute to bone formation and bone resorption, respectively, is crucial to maintain bone health (29). During postmenopausal osteoporosis, reduced production of estradiol (E2) disturbs bone homeostasis due to altered estrogen receptor (ER) signaling, resulting in decreased osteoblast (30, 31) and increased osteoclast activities (32). Several mechanisms have been proposed for the regulation of ERs on osteoblast and osteoclast functions, such as cross-talk and synergy of ER signaling with osteogenic signaling, including Wnt/β-catenin (33) and transforming growth factor–β (34). During osteoclast differentiation, estrogen blocks RANKL/M-CSF (receptor activator of nuclear factor κB ligand/macrophage colony-stimulating factor)–induced activator protein-1–dependent transcription, likely through direct regulation of c-Jun activity (35), and ESR1 promotes apoptosis of osteoclasts via the induction of the Fas/Fas ligand system (36).

In this study, we examine the expression levels of ectonucleotidases CD73 and CD39 during osteoporosis and the effects of altered activity of these enzymes on extracellular adenosine levels. Our findings demonstrate a direct correlation between impaired ectonucleotidase expression and extracellular adenosine levels in a mouse model of postmenopausal bone loss. Given the importance of estrogen in postmenopausal bone loss, we also established the role of ER signaling on ectonucleotidase expression and extracellular adenosine levels with a focus on A2BR signaling. In an experiment yielding clinical implications, we used the A2BR agonist BAY 60-6583 to compensate for the decrease in adenosine signaling and show that this intervention attenuates bone loss.

RESULTS

Surface membrane ectonucleotidases and extracellular adenosine are decreased in bone marrow of osteoporotic animals

In this study, we used ovariectomized (OVX) mice, which is widely recognized as a model of postmenopausal osteoporosis (37). Estradiol measurements in the peripheral blood and microcomputed tomography (microCT) analyses of the vertebrae after 4 weeks of ovariectomy were used to ensure bone loss (fig. S1, A and B). Analyses of CD73 and CD39 on OVX bone surfaces revealed a decreased expression compared to sham control (Fig. 1, A and B). We next examined the levels of these ectonucleotidases in the hematopoietic and nonhematopoietic fraction of bone marrow (BM) cells (Fig. 1, C to H). Quantification of the flow cytometric analyses of hematopoietic cells (Fig. 1, C to E) demonstrated a significantly lower fraction of cells expressing CD73 and CD39, as well as decreased median fluorescence for these ectonucleotidase compared to the control. Analyses of the nonhematopoietic cells showed a similar observation that a significantly lower fraction of the cells express CD73 and CD39 compared to the control (Fig. 1, F to H). The lower values of ectonucleotidase expression in nonhematopoietic population are possibly associated with their underdetection due to debris in as-isolated samples. We have also characterized the primary cells isolated from bone chips for transcription levels and found decreased expressions of CD73 (Fig. 1I) and CD39 (Fig. 1J) in OVX bone compared to corresponding healthy controls. Concomitant with the lower ectonucleotidase expressions, measurement of extracellular adenosine in the BM plasma showed a significant decrease in its concentration in OVX mice (Fig. 1K).

Fig. 1 Deficient CD73 and CD39 expressions and extracellular adenosine concentration in BM of OVX animals.

(A to K) Characterization of healthy (sham) and OVX animals 4 weeks after ovariectomy. (A) Immunofluorescence staining of CD73 (green) and (B) CD39 (red) in vertebrae of OVX animals. Nuclear staining (blue). Scale bars, 100 μm. Inset shows magnified image of bone surface. Yellow arrowheads indicate cells positive for CD73 or CD39 on bone surface. Scale bars, 50 μm. (C) Flow cytometric analysis of CD73 and CD39 membrane expression of hematopoietic cells from mouse BM cells 4 weeks after ovariectomy (OVX) and healthy controls. (D) Percentage and median fluorescence intensity of hematopoietic cells expressing CD73. (E) Percentage and median fluorescence intensity of hematopoietic cells expressing CD39. (F) Flow cytometric analysis of CD73 and CD39 membrane expression of nonhematopoietic cells from mouse BM cells 4 weeks after ovariectomy and healthy controls. (G) Percentage and median fluorescence intensity of nonhematopoietic cells expressing CD73. (H) Percentage and median fluorescence intensity of nonhematopoietic cells expressing CD39. (I) CD73 gene expression and (J) CD39 gene expression of cells from bone chips. (K) Extracellular adenosine concentration in BM plasma of sham and OVX animals. n = 5. *P < 0.05, **P < 0.01, ***P < 0.001.

ERs regulate ectonucleotidase expression and availability of adenosine in vitro

To explore whether estradiol (E2) is involved in maintaining CD73 and CD39 expression, we have withdrawn E2 during culture as described in Materials and Methods. Flow cytometric analyses of the osteoprogenitor cells (Fig. 2A) revealed that the ratio of double-positive CD73- and CD39-expressing cells was decreased in the absence of E2 (Fig. 2B). Since ERs are the main receptors of E2, we further examined whether ERs regulate the expression of CD73 and CD39 and subsequently the derivation of extracellular adenosine in osteoprogenitor cells. Small interfering RNA (siRNA) oligonucleotides against Esr1 and Esr2 were used. ER expression of osteoprogenitor cells was decreased in the knockdown of ESR1 (fig. S2A), ESR2 (fig. S2B), or dual knockdown of ESR1/ESR2 (fig. S2C). Flow cytometric analyses of the osteoprogenitor cells (Fig. 2C) revealed that the ratio of double-positive CD73- and CD39-expressing cells was decreased in dual knockdown groups (Fig. 2D). We also observed a similar trend in single ESR1 and ESR2 knockdown groups (Fig. 2D). Immunofluorescence staining of CD73 and CD39 in osteoprogenitor cells also demonstrated decreased double-positive cells in dual knockdown groups compared to control (fig. S2, D and E). Concomitant with the decrease in ectonucleotidase expressions, the concentration of extracellular adenosine decreased in all groups (Fig. 2E). We also examined the expression levels of individual ectonucleotidase (CD73 or CD39). Flow cytometric analyses of CD73 alone (fig. S3A) showed a decrease in the percentage of CD73-expressing cells and median fluorescence in osteoprogenitors with both single and dual ER knockdown (fig. S3, B and C). Contrary to CD73, expression level of CD39 was found to increase in all groups (fig. S3, D to F).

Fig. 2 Regulation of CD73 and CD39 cell membrane expressions and extracellular adenosine levels by ERs in osteoprogenitor cells.

(A) Flow cytometric analyses and (B) quantification of CD73 and CD39 in osteoprogenitors in the absence or presence of E2 (100 nM) for 3 days. (C to E) Single (ESR1 or ESR2) or dual (ESR1 and ESR2) ER knockdown (KD) by siRNA in primary mouse osteoprogenitors and analyzed after 3 days. (C) Flow cytometric analyses of CD73 and CD39 after single knockdown (ESR1 or ESR2) and dual knockdown (ESR1 and ESR2). (D) Percentage of double-positive (CD73/CD39) cells in single knockdown and dual knockdown cells. (E) In vitro adenosine levels normalized by cell number in single knockdown and dual knockdown cells. Control (scrambled) siRNA concentration for single knockdown and dual knockdown are 5 and 10 nM, respectively. n = 5. *P < 0.05, **P < 0.01, ***P < 0.001.

We also carried out a similar analysis for mononuclear cells undergoing osteoclast differentiation. Flow cytometric analyses of the osteoclasts (Fig. 3A) revealed that the ratio of double-positive CD73- and CD39-expressing cells was decreased in the absence of E2 (Fig. 3B). ESR1 and ESR2 expressions of mononuclear cells undergoing differentiation were decreased in single knockdown of ESR1 (fig. S4A), ESR2 (fig. S4B), or dual knockdown (fig. S4C). Flow cytometric analyses of CD73 and CD39 in osteoclasts (Fig. 3C) revealed that the ratios of double-positive CD73- and CD39-expressing cells were decreased in ESR1 or ESR2 knockdown or dual knockdown (Fig. 3D). Immunofluorescence staining of CD73 and CD39 in osteoclasts also demonstrated decreased double-positive cells in dual knockdown groups compared to control (fig. S4, D and E). Extracellular adenosine levels showed a significant decrease in single and dual knockdown groups that correlated with the decreased ectonucleotidase expression (Fig. 3E). Flow cytometric analyses of CD73 alone (fig. S5A) showed a decrease in the percentage of cells expressing CD73 in single and dual knockdown groups (fig. S5, B and C). Analyses of CD39 showed a decrease in the percentage of cells expressing the marker in all the groups (fig. S5, D to F).

Fig. 3 Regulation of CD73 and CD39 cell membrane expression and extracellular adenosine levels by ERs in osteoclasts.

(A) Flow cytometric analyses and (B) quantification of CD73 and CD39 in primary mouse mononuclear cells undergoing osteoclast differentiation in the absence or presence of E2 (100 nM) for 3 days. (C to E) Single (ESR1 or ESR2) or dual (ESR1 and ESR2) ER knockdown by siRNA during macrophage differentiation for 3 days and subsequent osteoclast differentiation for 6 days. (C) Flow cytometric analyses of CD73 and CD39 after single knockdown (ESR1 or ESR2) and dual knockdown (ESR1 and ESR2). (D) Percentage of double-positive (CD73/CD39) cells in single knockdown and dual knockdown cells. (E) In vitro adenosine levels normalized by cell number in single knockdown and dual knockdown cells. Control (scrambled) siRNA concentration for single knockdown and dual knockdown are 5 and 10 nM, respectively. n = 4. *P < 0.05, **P < 0.01, ***P < 0.001.

Since macrophages are precursors to osteoclasts, we have also analyzed this cell population. Knockdown of ERs resulted in a decreased ratio of CD73-expressing cells (fig. S6A) in the dual, but not in single (ESR1 or ESR2), knockdown groups (fig. S6B), along with a decreased median fluorescence of positive cells (fig. S6C). Similarly, analysis of CD39 (fig. S6D) showed a significant decrease in CD39-expressing macrophages in ESR1 and dual knockdown groups but not in the ESR2 knockdown group (fig. S6E). The median fluorescence of CD39-expressing cells decreased in dual knockdown but not in the case of single knockdowns (fig. S6F).

Adenosine regulates osteoblastogenesis and osteoclastogenesis through adenosine A2BR in vitro

Mouse primary osteoprogenitor cells cultured in medium supplemented with adenosine devoid of osteogenic-inducing factors showed up-regulation of osteoblast-specific transcription factors osterix (OSX) and osteopontin (OPN) (Fig. 4, A and B). To determine the involvement of A2BR signaling during adenosine-induced differentiation, we treated osteoprogenitors and osteoclast precursors with A2BR siRNA to perturb its expression (fig. S7, A and B, respectively). Knockdown of A2BR expression abrogated the increased OSX and OPN expressions compared to groups with no siRNA and control (scrambled) siRNA (Fig. 4, A and B). Contrary to osteoblastogenesis, extracellular adenosine diminished osteoclast differentiation as demonstrated by reduced expression of osteoclast transcription factor nuclear factor of activated T cells 1 (Nfatc1) and osteoclast-associated markers acid phosphatase type 5 (ACP5) and cathepsin K (CTSK) (Fig. 4, C to E). This exogenous adenosine-mediated down-regulation of Nfatc1, ACP5, and CTSK gene expressions were reversed upon knockdown of A2BR expression (Fig. 4, C to E). The diminished osteoclastogenesis was also verified by tartrate-resistant acid phosphatase (TRAP) staining, which showed an inhibitory effect of A2BR signaling during osteoclast differentiation (Fig. 4F). Together, the results suggest that A2BR activation promotes osteoblastogenesis and reduces osteoclastogenesis.

Fig. 4 Adenosine A2BR signaling promote osteogenic and inhibit osteoclast differentiation in vitro.

(A and B) In vitro knockdown of adenosine A2BR using siRNA in primary mouse osteoprogenitor cells isolated from the BM for 2 days, followed by adenosine treatment (ADO; 30 μg/ml) for 7 or 14 days. Gene expression of (A) osteoblast-specific marker and (B) Opn. (C to F) In vitro knockdown of adenosine A2BR by siRNA in mouse mononuclear cells isolated from BM undergoing macrophage differentiation for 3 days, followed by osteoclast differentiation along with treatment of small-molecule adenosine (30 μg/ml) for 6 days. Gene expressions of (C) osteoclast transcription factor Nfatc1, (D) ACP5, and (E) CTSK. (F) TRAP staining. Second group from the left is no siRNA control, while the third group from the left involves control (scrambled) siRNA. Scale bar, 200 μm. n = 3. *P < 0.05, **P < 0.01, ***P < 0.001.

Treatment of osteoporotic mice with adenosine A2BR agonist prevents bone loss

In vitro results demonstrating the dual action of A2BR signaling suggest that targeting this receptor could prevent bone loss. Immunohistochemical staining of A2BRs on the bone surface of OVX animal confirmed their presence on osteoporotic bone (fig. S8). We treated OVX mice displaying bone loss with the A2BR agonist BAY 60-6583 for 8 weeks and examined the changes in bone tissue. Hematoxylin and eosin (H&E) staining of the bone tissues displayed an attenuation of bone loss with BAY 60-6583 treatment (fig. S9A). TRAP staining revealed less staining and a decreased number of TRAP-positive cells on the bone surface of mice treated with BAY 60-6583 compared to the vehicle control (Fig. 5, A and B). Double-fluorescence bone labeling with calcein and alizarin complexone showed mineral deposition in all groups (Fig. 5C). Quantification of the images for mineral apposition rate (Fig. 5D) and bone formation rate (Fig. 5E) showed an abrogation of the bone loss in mice treated with BAY 60-6583. microCT analysis of vertebra (Fig. 5F) demonstrated the decrease in bone mineral density (BMD), bone volume per total volume (BV/TV), and trabecular number (Tb.N), as well as an increase in trabecular spacing (Tb.Sp) in OVX animals. We attenuated these changes after treatment with BAY 60-6583 (Fig. 5, G to J). We also observed a similar finding in femur again, showing that the bone loss was diminished upon treatment with BAY 60-6583 (fig. S9, B and C). Mechanical measurements of the bone tissue further supported these findings. Cohorts treated with BAY 60-6583 showed improved maximum load and stiffness of bone tissues compared to the corresponding controls (Fig. 5, K and L, respectively). Although we do not anticipate the treatment involving A2B agonist to increase the CD73/CD39 expression levels, we carried out immunofluorescence staining of CD73 (fig. S10A) and CD39 (fig. S10B), and the results revealed no significant differences in their expression levels between the vehicle- and BAY 60-6583–treated groups.

Fig. 5 Adenosine A2BR agonist BAY 60-6583 attenuates bone loss in OVX animals.

(A to L) Administration of BAY 60-6583 and vehicle for 8 weeks in OVX animals (4 weeks after ovariectomy). Groups are compared to healthy control with no surgery and no treatment (CTL). (A) TRAP staining (purple). Scale bar, 50 μm. (B) Quantification of TRAP-positive cells on bone surface. n = 4. (C) Double-fluorescence bone labeling by calcein (green) and alizarin complexone (red). Scale bar, 100 μm. (D) Quantification of mineral apposition rate (MAR) from bone labeling images. (E) Quantification of bone formation rate (BFR/BS) from bone labeling images. n = 5. (F) Reconstructed microCT images of L4 vertebra. Scale bar, 500 μm. (G to J) Quantification of microCT images. (G) BMD. (H) BV/TV. (I) Tb.N. (J) Tb.Sp. n = 5. Mechanical measurement for (K) maximum load and (L) stiffness of tibia. n = 12. OV, ovariectomy, vehicle [dimethyl sulfoxide (DMSO)]; OB, ovariectomy, BAY 60-6583 (1 mg/kg). *P < 0.05, **P < 0.01, ***P < 0.001.

DISCUSSION

This study establishes the pathophysiological correlation of altered ectonucleotidase CD39 and CD73 expressions and the extracellular adenosine availability in a postmenopausal osteoporotic model. Our results demonstrated that OVX mice deficient in estrogen have lower expression of CD73 and CD39 in hematopoietic and nonhematopoietic cells of the BM that correlated with a decrease in extracellular adenosine. This is consistent with the in vitro studies where ER signaling maintained the coexpression of ectonucleotidases CD73 and CD39 and extracellular adenosine during the differentiation of osteoblasts and osteoclasts. However, results showed some differences in ER regulation of ectonucleotidases CD73 and CD39 between precursors undergoing osteoblastogenesis and osteoclastogenesis. Specifically, analyses following the ER perturbation showed that the percentage of cells expressing CD39 increased in osteoblasts and decreased in osteoclasts. Contrary to CD39, the ER perturbation showed that the percentage of cells expressing CD73 decreased in both osteoblasts and osteoclasts. These findings suggest that CD39 could have a disparate role between osteoblasts and osteoclasts and that CD73 has a dominant role in regulating the extracellular adenosine level over CD39 during ER signaling in osteogenic differentiation of progenitor cells. This agrees with prior findings that osteogenic cells from CD39 knockout mice exhibit diminished osteogenic differentiation only when extracellular adenosine uptake was restricted (26). Furthermore, unlike CD73 knockout in male mice, CD39 knockout mice do not display an aberrant bone phenotype.

One of the cardinal reasons of osteoporosis is the compromised function of osteoblasts and excessive activity of osteoclasts. Studies have shown that A1 activation promotes osteoclast differentiation, while A2AR and A2BR exert inhibitory effects (4, 6, 8, 9, 26, 38, 39). Consistent with these reports, our in vitro studies showed that the supplementation of adenosine promotes osteoblastogenesis while decreasing osteoclastogenesis. Furthermore, our results showed that this extracellular adenosine–mediated osteoblastogenesis and osteoclastogenesis involved A2BR signaling. The dual ability of the adenosine molecule to promote osteoblastogenesis while preventing excessive osteoclastogenesis through A2BR signaling could be an ideal therapeutic strategy to treat bone loss. As a proof of concept, administration of A2BR agonist BAY 60-6583 in OVX animals showed attenuation of bone loss. Despite being a partial agonist (40), administration of BAY 60-6583 showed therapeutic potential, suggesting that the adenosine A2BR may serve as a therapeutic target in postmenopausal osteoporotic disease. Here, stimulation of A2BR was used to compensate for the low levels of extracellular adenosine in the bone milieu. While our results show that A2B stimulation can be used to compensate for the low levels of extracellular adenosine in the bone milieu, the treatment does not introduce a feedback signaling to increase CD73/CD39 expression.

Unlike current treatments involving bisphosphonates that inhibit osteoclastogenesis, targeting A2BR signaling to modulate the function of osteoblasts and osteoclasts offers an unexplored therapeutic strategy for treating osteoporosis. However, the ubiquitous nature of adenosine receptors in the human body warrants a more targeted therapeutic approach to move such an approach to clinic. Adenosine is well known for its immunomodulatory effect and anti-inflammatory properties (41, 42), implying that extracellular adenosine could also regulate the inflammatory-like environment present in osteoporosis. For example, proinflammatory cytokines have been implicated as primary mediators of accelerated bone loss during postmenopausal osteoporosis (4345). In addition to osteoclasts, multiple immune cells also participate in tissue resorption including destructive T cells and macrophages (44, 4648). In line with adenosine as an immune suppressor, the lack of adenosine signaling in CD73 knockout mice develops spontaneous arthritis associated with inflammatory symptoms (49, 50).

Besides ectonucleotidases, other factors contributing to extracellular adenosine availability, including soluble CD73 (51), metabolism by adenosine deaminase (52), cellular transport by equilibrative nucleoside transporter 1 (53), alternative CD38/CD203a/CD73, or CD203a/CD73 pathways (54, 55), were not investigated. Furthermore, we used OVX animals to study the pathophysiology of postmenopausal osteoporosis. Whether the same observations occur in osteoporosis during natural aging, secondary osteoporosis, human disease, and gender differences remain to be explored. Similar to females encountering postmenopausal osteoporosis, recent studies have shown that estrogen plays a crucial role in age-mediated male osteoporosis (56, 57).

In conclusion, we have demonstrated a direct correlation between expression of the ectonucleotidase CD73 and CD39 and extracellular adenosine levels in a mouse model of postmenopausal bone loss. We have established the active role of ER signaling in maintaining CD73 and CD39 expression and extracellular adenosine levels. As a proof of concept, we have shown that stimulation of A2BR using a small molecule can be used to compensate for low levels of extracellular adenosine and attenuate the associated bone loss. These results suggest that A2BR could be a potential therapeutic target for osteoporosis.

MATERIALS AND METHODS

Animals experiments

Female C57BL6/J mice were used (the Jackson laboratory, Bar Harbor, ME). All animal studies were performed with the approval of the Institutional Animal Care and Use Committee at Duke University and in accordance to guidelines of National Institutes of Health (NIH). All tools were sterile-autoclaved before use.

Ovariectomy surgery. Ovariectomy surgeries were performed at the Jackson laboratory or in-house at 12 weeks old as previously described (58). Before in-house ovariectomy surgery, animals were anesthetized with isoflurane (Henry Schein, Dublin, OH) by inhalation at 1 to 3% induction and 4% maintenance and administered with buprenorphine SR (Zoopharm, Windsor, CO) subcutaneously. A 3 cm by 3 cm of area cephalic from the iliac crest on left and right side of mice was shaved and wiped with 10% povidone-iodine (Purdue Products, Stanford, CT). A 2- to 3-cm midline incision was made, and the skin was bluntly dissected from the underlying fascia. Another incision was made through the fascia, 1 cm lateral of the midline, and bluntly dissected laterally until it reaches the abdominal cavity. The adipose tissue that surrounds the ovary in the abdominal cavity was gently pulled out by tweezers. The uterine horns and vessels were ligated 0.5 to 1 cm proximally, and the ovary was cut. The fascia wound was closed using a degradable vicryl 5-0 suture (Ethicon, Somerville, NJ), and the skin wound was closed with a 3-0 nylon suture (Ethicon) with a topical application of 0.5% bupivacaine (Hospira, Lake Forest, IL). Another incision in the contralateral fascia was performed, and the procedure was repeated. Animals were monitored for the duration of the surgery.

Administration of BAY 60-6583. BAY 60-6583 (Tocris Bioscience, Minneapolis, MN) was dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich, St. Louis, MO). For in vivo administration, fresh solutions of BAY 60-6583 (10%, v/v), 0.9% sodium chloride (NaCl; 50%, v/v; Hospira), and polyethylene glycol 400 (PEG 400; 40%, v/v; Thermo Fisher Scientific, Hampton, NH) were mixed and sterile-filtered through 0.22-μm-pore filter disk, and 100 μl of solution was injected intraperitoneally at mouse weight (1 mg/kg), after 8 weeks of ovariectomy surgery, once every 2 days. A solution comprising DMSO (10%, v/v), 0.9% NaCl (50%, v/v), and PEG 400 (40%, v/v) was injected in animals as vehicle control.

Cell isolation, culture, and differentiation

Three- to 4-week-old female C57BL/6 J mice were euthanized by carbon dioxide and bilateral thoracotomy.

Isolation of osteoprogenitor cells and culture. Osteoprogenitor cells were isolated as previously described with modifications. All buffers were ice cold, unless otherwise indicated. Briefly, the femur, tibia, humerus, radius, and ulna of mice were harvested. BM was flushed out and discarded. The bone was collected and cut into 1-mm3 chips in harvest buffer and then digested in digestion buffer containing growth media [α minimum essential medium (αMEM), fetal bovine serum (FBS) (10%, v/v), penicillin/streptomycin (10,000 U/ml; 1%, v/v)], and collagenase type 2 (1 mg/ml, w/v; Worthington Biochemical, Lakewood, NJ) while shaking at 60 rpm on orbital shaker (catalog no. 51700-13, Cole-Parmer, Vernon Hills, IL) in humidified incubator (37°C, 5% CO2) for 1.5 hours. Digested bone chips were rinsed three times with growth media and transferred to two wells of six-well plate for culture. Media were replaced after 3 days, and bone chips were further cultured for 3 days for cells to adhere and proliferate before passage. For the first passage, cells were treated with 0.25% trypsin-EDTA (Thermo Fisher Scientific) for 3 min, neutralized with growth media, detached with cell scraper, and subcultured along with bone chips. The procedure for subsequent passages were similar but without bone chips. All experiments were performed at three to four passages. For estradiol (E2) withdrawal experiments, cells were cultured in the absence or presence of E2 (Sigma-Aldrich) supplemented at 100 nM every 3 hours to charcoal-stripped growth media [αMEM, charcoal-stripped FBS (10%, v/v), penicillin/streptomycin (10,000 U/ml; 1%, v/v)]. For adenosine supplementation, cells were supplemented with adenosine (30 μg/ml; Sigma-Aldrich) in growth media with fresh changes of media every day.

Isolation of mononuclear cells and macrophage/osteoclast differentiation. All buffers used were ice cold, unless otherwise indicated. Briefly, the femur, tibia, humerus, radius, ulna, and vertebra were harvested and crushed with pestle and mortar in harvest buffer [phosphate-buffered solution (PBS) and FBS (2%, v/v)] to release BM tissue. BM was passed through a 70-μm cell strainer and centrifuged at 200g for 5 min. Cells were resuspended in harvest buffer, gently layered onto Ficoll-Paque PLUS (GE Healthcare, Marlborough, MA) at 1:1 ratio, and centrifuged without rotor acceleration and deceleration at 200g for 15 min. Afterward, the opaque middle layer with cells was collected, washed with harvest buffer, and centrifuged at 200g for 5 min to yield a cell pellet. Isolated mononuclear cells were cultured in macrophage induction media, containing growth media, prostaglandin E2 (PGE2; 10−7 M; Santa Cruz Biotechnology, Dallas, TX), and M-CSF (10 ng/ml; PeproTech, Rocky Hill, NJ) at 100,000 cells/cm2. For osteoclast differentiation, macrophages cultured for 3 days were further induced in osteoclast induction media containing growth media, PGE2 (10−7 M), M-CSF (10 ng/ml), and RANKL (10 ng/ml; PeproTech). For estradiol (E2; Sigma-Aldrich) withdrawal experiments, osteoclasts were cultured in the absence or presence of E2 supplemented at 100 nM to media containing charcoal-stripped growth media, PGE2 (10−7 M), M-CSF (10 ng/ml), and RANKL (10 ng/ml; PeproTech). For adenosine supplementation, cells were supplemented with adenosine (30 μg/ml; Sigma-Aldrich) in differentiation media with fresh media change every day.

siRNA knockdown

Cells were transfected with Silencer Select siRNA oligonucleotides (Thermo Fisher Scientific), according to the manufacturer’s instructions. Briefly, 5 nM siRNA was mixed with RNAiMAX transfection reagent (Thermo Fisher Scientific) in the presence of Opti-MEM (Thermo Fisher Scientific) for 5 min at room temperature (RT). The solution was transfected into osteoprogenitor cells in the presence of growth media or mononuclear cells in macrophage induction media for 2 days. The following siRNA oligonucleotides were used: ADORA2B (A2BR; catalog no. 4390771; ID, s62047), ESR1 (catalog no. 4390771; ID, s65686), ESR2 (catalog no. 4390771; ID, s65689), and negative control #1 (catalog no. 4390843). A concentration of 5 nM negative control siRNA was used for single ER knockdown, while a concentration of 10 nM negative control siRNA was used for dual ER knockdown.

Flow cytometry

Whole BM flush were collected from tibia and femur into buffer containing PBS [3% (v/v) bovine serum albumin (BSA)] and filtered through 40-μm nylon filter (BD Biosciences, San Jose, CA). Red blood cells (RBCs) were lysed with RBC lysis buffer (Thermo Fisher Scientific) for 5 min at RT. Cells were stained with CD39 phycoerythrin (PE)/cyanine-7 (8 μg/ml; 143805, BioLegend, San Diego, CA), CD73 PE (1.25 μg/ml; 12-0731-82, Thermo Fisher Scientific), and CD45 allophycocyanin (APC) (1.25 μg/ml; 17-0451-82, Thermo Fisher Scientific) antibodies for 30 min at RT. Stained cells were analyzed with BD Accuri C6 flow cytometer and CFlow software. Analyses were performed by comparing to unstained cells and gated for singlets.

RNA isolation, reverse transcription, and real-time polymerase chain reaction

Total RNA was extracted with TRIzol (Thermo Fisher Scientific), phase-separated with chloroform, and precipitated using isopropanol. One microgram of RNA was reverse-transcribed using iScript cDNA Synthesis Kit (Bio-Rad), according to the manufacturer’s instructions. iTaq Universal SYBR green reagent (Bio-Rad) was used to detect gene expression during amplification of complementary DNA after initial denaturation at 95°C for 30 s for one cycle and 95°C for 5 s and 60°C for 30 s for 40 cycles on a polymerase chain reaction (PCR) cycler (Bio-Rad). The mouse primer sequences are as follows:

OSX (forward, TGCCTGACTCCTTGGGACC; reverse, TAGTGAGCTTCTTCCTCAAGCA), OPN (forward, AAACCAGCCAAGGTAAGCCT; reverse, TCAGTCACTTTCACCGGGAG), NFATC1 (forward, GGTAACTCTGTCTTTCTAACCTTA; reverse, GTGATGACCCCAGCATGCACCAGTCACAG), CTSK (forward, GGGCTCAAGGTTCTGCTGC; reverse, TGGGTGTCCAGCATTTCCTC), ACP5 (forward, CAGCAGCCCAAAATGCCT; reverse, TTTTGAGCCAGGACAGCTGA), ESR1 (forward, CTTGAACCAGCAGGGTGGC; reverse, GAGGCTTTGGTGTGAAGGGT), ESR2 (forward, GACGAAGAGTGCTGTCCCAA; reverse, GCCAAGGGGTACATACTGGAG), ADORA2B (forward, ATCTTTAGCCTCTTGGCGGTG; reverse, GACCCAGAGGACAGCAATGAT), and 18S ribosomal RNA (forward, ACCAGAGCGAAAGCATTTGCCA; reverse, ATCGCCAGTCGGCATCGTTTAT).

Adenosine assay

Adenosine levels were measured from plasma of BM flush or cultured media using adenosine assay kit (Abcam), according to the manufacturer’s instructions. For BM measurements, femur and tibia containing marrow were centrifuged at 200g for 1 min at 4°C to collect whole-marrow flush. The flush was then centrifuged at 2000g for 5 min at 4°C to separate the cell and plasma fractions. For cell media measurements, media cultured with cells for 3 days were collected. The BM plasma or cell media samples were then diluted and mixed with reaction mix containing adenosine assay buffer, adenosine detector, adenosine converter, adenosine developer, and adenosine probe at 1:1 ratio in a well of 96-well white plate. Fluorescence intensity was detected with a plate reader (Tecan Infinite 200 PRO) using Ex535 and Em590 nm filters. Fluorescence was subtracted from background and quantified using a standard.

Enzyme-linked immunosorbent assay

Plasma estradiol from mouse peripheral blood was quantified using estradiol assay kit (R&D Systems, Minneapolis, MN), according to the manufacturer’s instructions. Briefly, murine peripheral blood was collected from tail vein in the presence of heparin and centrifuged at 1000g for 10 min at 4°C to separate the cell and plasma fractions. Samples were pretreated with pretreatment E solution and centrifuged, and supernatant was mixed with pretreatment F solution. To wells coated with estradiol antibody, samples and estradiol conjugate were added and incubated for 2 hours at RT on a shaker. Wells were washed, and substrate solution was added for 30 min at RT and protected from light. Then, stop solution was added, and measurements were performed with a plate reader (Tecan Infinite 200 PRO) using Ex450 and Em540 nm filters.

Histology and staining

Vertebral samples were fixed with 4% paraformaldehyde (PFA) at 4°C for 1 day and decalcified using 10% EDTA (pH 7.3) for 2 weeks at 4°C. The samples were gradually dehydrated using increasing concentrations of ethanol and incubated in CitriSolv (Decon Labs, King of Prussia, PA) until equilibrium was reached. Following dehydration, samples were immersed in a mixture of 50% (v/v) CitriSolv (Decon Labs) and 50% (w/w) paraffin (General Data Healthcare) for 30 min at 70°C. The samples were embedded in paraffin and sliced into sections of 10-μm thickness using a rotary microtome (RM2255, Leica). Before staining, the sections were deparaffinized using CitriSolv and subsequently rehydrated with decreasing concentration of ethanol until the samples were equilibrated with deionized (DI) water.

H&E staining. H&E staining was performed by first incubating the samples in hematoxylin solution (RICCA Chemical, Arlington, TX) for 1 min, followed by incubation with Eosin-Y solution (Richard-Allan Scientific, San Diego, CA) for 20 s. Stained sections were gradually dehydrated using increasing concentrations of ethanol until equilibrium was reached. Sections were mounted in glycerol and imaged using a Keyence BZ-X700 microscope.

TRAP staining. TRAP staining was performed using TRAP kit following the manufacturer’s instructions (Sigma-Aldrich). Briefly, the solution was prepared by first mixing 50 ml of Fast Garnet GBC base solution with 50 ml of sodium nitrite solution. This mixture was added to 4.5 ml of DI water prewarmed at 37°C. After mixing, 50 ml of Naphthol AS-B1 phosphate solution, 200 ml of acetate solution, and 100 ml of tartrate solution were added to the solution and mixed to generate a working solution. Rehydrated sections were immersed in the working solution, incubated at 37°C for 1 hour covered from light, and rinsed with ultrapure water. Sections were then gradually dehydrated using increasing concentrations of ethanol until equilibrium was reached. Slides were mounted with glycerol and imaged immediately. Images were quantified with ImageJ for TRAP+ cells and normalized to the length of bone surfaces.

Immunofluorescence staining. For immunofluorescence staining, rehydrated sections were immersed in a solution of proteinase K (20 mg/ml; Thermo Fisher Scientific) in 95% (v/v) TE buffer [50 mM tris-HCl, 1 mM EDTA, and 0.5% (v/v) Triton X-100 (pH 8)] with 5% (v/v) glycerol and incubated for 15 min at 37°C. Sections were washed with PBS and permeabilized using 0.1% Triton X-100 in PBS for 10 min at RT. The sections were immersed in a blocking solution [1% (w/v) BSA, glycine (0.25 M), normal donkey serum (5%, v/v), and normal goat serum (5%, v/v) in tris-buffered solution (TBS)] and incubated for 1 hour at RT. Sections were then incubated with primary antibody against CD39 (5 μg/ml; AF4398, R&D Systems), CD73 (5 μg/ml; AF4488, R&D Systems), and A2BR (1:200; MBS8207549, MyBioSource, San Diego, CA) in diluent solution (1%, w/v) and normal donkey serum (1%, v/v) in TBS overnight at 4°C. Sections were stained with secondary antibody using anti-donkey or anti-goat Alexa Fluor 647 (1:250; Jackson ImmunoResearch, West Grove, PA). For immunocytochemical costaining, cells were incubated with primary antibody against CD39 (1:100; ab227840, Abcam, Cambridge, UK) and CD73 (5 μg/ml; AF4488, R&D Systems) and stained with secondary antibody using anti-donkey Alexa Fluor 488 and anti-goat Alexa Fluor 647 (1:250; Jackson ImmunoResearch). Images were acquired and presented as pseudocolors.

Microcomputed tomography

Bone mineralization was analyzed as previously described (59). L3 to L5 vertebra and femur were collected, fixed in 4% PFA at 4°C for 1 day, and rinsed thoroughly with PBS. The fixed samples were placed in a 50-ml centrifuge tube with styrofoam spacers and loaded into a microCT scanner (vivaCT 80, Scanco Medical, Wayne, PA) and scanned at 55 keV at a pixel resolution of 10.4 μm. The reconstruction of scanned images was performed using microCT Evaluation Program V6.6 (Scanco Medical), followed by generation of radiographs and three-dimensional models using microCT Ray V4.0 (Scanco Medical). Mineral density of the scaffolds was quantified and presented as a percentage of BV/TV based on 100 contiguous slices.

Bone labeling

Animals were administered with calcein (10 mg/kg body weight; Sigma-Aldrich) at 14 days and alizarin complexone (30 mg/kg body weight; Sigma-Aldrich) at 5 days before euthanization. Collected cranium were fixed in 4% PFA at 4°C for 1 day and stored in 70% ethanol. Subsequently, the samples were dehydrated at 70% ethanol (2 days), 95% ethanol (2 days), 100% 2-propanol (twice for 1 day), and xylene (twice for 2 days). After dehydration, the samples were infiltrated with methyl methacrylate embedding mixture, sectioned, and imaged.

Mechanical measurement

Mechanical properties of 12 tibiae per group were measured as previously described. After removing soft tissues, tibia samples were wrapped in wet tissue and frozen at −20°C. Sixteen hours before testing, the samples were transferred to 4°C and then to RT an hour before testing. The four-point bending mode of ElectroForce 3220 (TA Instruments, New Castle, DE) instrument with 225-N load cell was used. Samples were aligned on the fixtures in a manner that the load was applied perpendicular to the principal axis of tibia. The span length of the bottom support was 9.2 mm, while the top span length was 2.8 mm. Bending test was performed in displacement control mode at a rate of 0.025 mm/s. Load-displacement data were recorded at a data acquisition rate of 10 Hz. Displacement was tared at the first data point at which the load equaled or exceeded 1 N. Maximum load is the highest load (newtons) before the sample fractures. Bending stiffness (newtons per millimeter) was calculated as the slope of load versus displacement between 30 and 70% of maximum load to failure in the linear region.

Statistical analyses

Statistical analyses were carried out using GraphPad Prism 5. Two-tailed Student’s t test was used to compare two groups. One-way analysis of variance (ANOVA) with Tukey post hoc test was used to compare three or more groups. The P values were obtained from each test.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/8/eaax1387/DC1

Fig. S1. Measurement of estradiol (E2) and microCT imaging in OVX mice.

Fig. S2. ER knockdown in osteoprogenitors and immunofluorescent staining of ectonucleotidase expression.

Fig. S3. ER knockdown in osteoprogenitors and flow cytometric analyses of ectonucleotidase expression.

Fig. S4. ER knockdown in osteoclasts and immunofluorescent staining of ectonucleotidase expression.

Fig. S5. ER knockdown in osteoclasts and flow cytometric analyses of ectonucleotidase expression.

Fig. S6. ER knockdown of BM cells undergoing macrophage differentiation and flow cytometric analyses of ectonucleotidase expression.

Fig. S7. siRNA knockdown of A2BR and reverse transcriptase quantitative PCR.

Fig. S8. Immunofluorescent staining of adenosine A2BR expression in vertebra of animals.

Fig. S9. H&E staining and microCT of BAY 60-6583–treated mice.

Fig. S10. Immunofluorescent staining of CD73 and CD39 in vertebra of BAY 60-6583–treated animals.

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: We thank H. C. Cutcliffe and L. E. DeFrate for help with mechanical measurements. Funding: We acknowledge the financial support from NIH, USA (R01 AR063184 and R01 AR071552). Author contributions: Y.-R.V.S. and S.V. designed the study, interpreted the data, and wrote the manuscript. M.L., S.K.K., M.I., Y.G., and N.S. performed different experiments. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Navigate This Article