|Home | About | Journals | Submit | Contact Us | Français|
Age-dependent changes in skeletal growth are important for regulating skeletal expansion and determining peak bone mass. However, how G protein–coupled receptors (GPCRs) regulate these changes is poorly understood. Previously, we described a mouse model expressing Rs1, an engineered receptor with high basal Gs activity. Rs1 expression in osteoblasts induced a dramatic age-dependent increase in trabecular bone with features resembling fibrous dysplasia. To further investigate how activation of the Gs-GPCR pathway affects bone formation at different ages, we used the tetracycline-inducible system in the ColI(2.3)+/Rs1+ mouse model to control the timing of Rs1 expression. We found that the Rs1 phenotype developed rapidly between postnatal days 4 and 6, that delayed Rs1 expression resulted in attenuation of the Rs1 phenotype, and that the Rs1-induced bone growth and deformities were markedly reversed when Rs1 expression was suppressed in adult mice. These findings suggest a distinct window of increased osteoblast responsiveness to Gs signaling during the early postnatal period. In addition, adult bones encode information about their normal shape and structure independently from mechanisms regulating bone expansion. Finally, our model provides a powerful tool for investigating the effects of continuous Gs-GPCR signaling on dynamic bone growth and remodeling. © 2010 American Society for Bone and Mineral Research.
Osteoporosis affects over 10 million people in the United States and contributes to 1.5 million fractures each year.(1) A crucial determinant of osteoporosis and fracture risk in adults is thought to be the acquisition of peak bone mass during childhood and puberty.(2–4) Osteoblasts and their precursors play key roles in regulating bone development and acquisition of peak bone mass.(5) Although certain G protein–coupled receptors (GPCRs) expressed in osteoblasts [e.g., parathyroid hormone receptor (PTHR1)(6) and HTR1b serotonin receptor(7)] are involved in regulating bone mass, the exact in vivo roles of GPCR signaling during skeletal growth have not been clearly elucidated.
GPCRs signal through a select number of pathways, including the Gs and Gi pathways that influence intracellular cyclic AMP (cAMP) levels.(8) We described a synthetic biology approach(9) for activating the Gs pathway that uses the strong basal Gs signaling activity of Rs1, an engineered receptor activated solely by a synthetic ligand (RASSL).(10,11) RASSLs are engineered receptors that no longer respond to endogenous hormones but can be activated by synthetic small-molecule drugs. They have proven useful for studying the roles of G protein signaling in complex systems.(12–15)
We found that when Rs1 was expressed continuously in mouse osteoblasts from gestation until adulthood, skeletal bone mass and trabecular bone formation were dramatically greater than in wild-type mice.(11) Cortical bone was lost, and serum markers of bone turnover were greatly increased. In addition, the normal bone marrow space was obliterated and replaced with small uniform cells morphologically resembling young osteoblasts. In contrast, when Rs1 was expressed beginning at 4 weeks of age, no grossly abnormal skeletal phenotype was evident through 30 weeks of age. These findings suggested that receptor-mediated Gs signals in osteoblasts have age-dependent effects on bone formation. In addition, evidence from human diseases and animal models suggest that osteoblast function is markedly different in children than in adults.(16–21)
In this study, we asked how the timing of Gs signaling affects the formation of skeletal bone at different ages. We used Rs1 to activate the Gs signaling pathway in osteoblasts in mice at different ages in four experiments (Fig. 1). First, we identified a window of osteoblast responsiveness to Rs1 expression by examining when the Rs1-induced trabecular bone phenotype begins. Second, we narrowed the window of Rs1 expression necessary for developing the trabecular overgrowth phenotype by delaying Rs1 expression until after birth. Third, we assessed whether expression of Rs1 after 4 weeks of age could lead to long-term attenuation of the Rs1 phenotype in adult mice. Finally, we determined whether the Rs1-induced bone phenotype could be reversed if Rs1 expression was discontinued after formation of the abnormal trabecular bone in adult mice. Our findings demonstrate the utility of the Rs1 mouse model for controlling the timing and tissue-specific activation of Gs signaling as a way to understand the age-dependent effects of continuous Gs signaling by GPCRs.
All transgenic mouse studies were approved by and performed in accord with the Institutional Animal Care and Use Committee and the Laboratory Animal Research Center at the University of California, San Francisco. The ColI(2.3)-tTA/TetO-Rs1 double transgenic mice were generated by heterozygote crosses of mice carrying the TetO-Rs1 transgene with mice carrying the ColI(2.3)-tTA transgene as described previously.(11) Transgene expression was suppressed by continuous administration of doxycycline-impregnated mouse chow (DoxDiet 200 mg/kg; BioServ, Frenchtown, NJ, USA) during the times indicated in Fig. 1. Transgene expression was activated by switching the mice to regular mouse chow without doxycycline (LabDiet 5053, PMI Nutrition, St. Louis, MO, USA) at the times indicated in Fig. 1. Full transgene expression was expected to occur within 1.5 to 2 weeks of doxycycline withdrawal.(12,22) Transgene expression was verified by quantitative polymerase chain reaction (qPCR) in representative animals from each cohort (see Figs. 1B, ,3C,3C, ,5C,5C, and and7F).7F). Neonatal pups analyzed in Fig. 3 were not manipulated until the day of tissue collection to minimize any risk of maternal distress or cannibalism. Genotyping and selection of the appropriate ColI(2.3)+/Rs1+ and control pups were done after tissue collection, or after 4 weeks of age for adult animals. No alterations in litter size, pup health, or maternal health were observed in our experiments. Both males and females were analyzed together in our experiments because no sex-dependent differences were observed.
Gene expression analysis was performed on RNA isolated from the right femur or right humerus of adult experimental animals, as indicated in each figure legend. For the experiments in Fig. 3, both humeri from the same 4- or 5-day-old animal were combined to maximize RNA yield. Briefly, whole bones were dissected away from surrounding soft tissue and frozen in liquid nitrogen. Bones for each experiment were batch processed by crushing (multisample Bio-Pulverizer, Research Products International, Mt. Prospect, IL, USA) followed by homogenization (4.5 mm Tissue Tearor, Research Products International) in RNAStat-60 (Iso-Tex Diagnostics, Friendswood, TX, USA). cDNA was generated using the SuperScript III First Strand Synthesis kit (Invitrogen, Carlsbad, CA, USA) as directed by the manufacturer. Expression was assayed using SybrGreen primers as described previously,(11) Sybr green primers for RANKL (forward = TTGCACACCTCACCATCAAT; reverse = TCCGTTGCTTAACGTCATGT) and OPG (forward = CAGAGACCAGGAAATGGTGAA; reverse = AAGCTGCTCTGTGGTGAGGT) or TaqMan primers (Mm99999913_g1 for GAPDH, Mm00484036_m1 for cathepsin K, Mm0081666_g1 for collagen 1a1, Mm00504574_m1 for osterix, and Mm00501578_m1 for runx2). All samples were assayed in technical triplicate, and expression levels were normalized to GAPDH. All qPCR reactions were run on an Applied Biosystems (Foster City, CA, USA) 7900HT real-time thermocycler.
Mice identified for dual-energy X-ray absorptiometry (DXA) to measure whole-body areal bone mineral density (aBMD) were anesthetized with inhaled isofluorane (1.5% to 2% in oxygen) and scanned on a GE Lunar Piximus2 (Madison, WI, USA) at predetermined times. Mice that underwent whole-mouse or ex vivo femur micro-computed tomographic (µCT) scans were sacrificed before scanning on a Scanco vivaCT-40 µCT scanner, (SCANCO, Waynes, PA, USA). Femora for ex vivo scanning were fixed in 10% neutral buffered formalin (Fisher Scientific, Fairlawn, NJ, USA) for 24 hours and stored in 70% ethanol. Images were obtained at an X-ray energy of 55 kV, with a voxel size of 76 or 10.5 µm and integration times of 200 or 1000 ms for whole-body images or ex vivo femur images, respectively. Because of age-dependent variations in bone mineralization, segmentation values for whole-animal µCT were determined empirically as 0.1/1/100 (gauss sigma/gauss support/threshold in mg HA/ccm) for newborns and 1.5-week-old animals. Segmentation values of 0.7/1/250 were used for the ex vivo femur CT scans.
Mice identified for histomorphometry were injected with 15 mg/kg calcein (Sigma-Aldrich, St. Louis, MO, USA) 7 days before harvesting and with 90 mg/kg xylenol orange (Sigma-Aldrich) 2 days before harvesting to label areas of active bone mineralization. Harvested bones were processed for von Kossa staining, tartrate-resistant acid phosphatase (TRAP) immunostaining, and fluorescence microscopy as described previously.(11) Bones for hematoxylin and eosin staining were decalcified for 12 to 48 hours in 10% EDTA/PBS before paraffin embedding, sectioning, and staining by the Gladstone Histology Core.
Blood was collected from euthanized mice by cardiac puncture and processed in MicroTainer serum separator tubes according to the manufacturer's instructions (BD Biosciences, San Jose, CA, USA). Routine blood analysis was carried out by Antech Diagnostics (Irvine, CA, USA). Serum osteocalcin and pyridinoline measurements were carried out using the mouse osteocalcin EIA Kit BT-470 from BTI (Stoughton, MA, USA) and the MetraBiosystems SerumPYD Kit 8019 (Palo Alto, CA, USA) according to manufacturers' directions.
Two-tailed Student's t tests with unequal variances were performed using Microsoft Excel 2003. ANOVA analysis was performed using JMP8 (Version 8.0, SAS Institute, Cary, NC, USA).
We previously reported that Rs1-induced basal Gs signaling in osteoblasts of ColI(2.3)+/Rs1+ mice increased trabecular bone and decreased cortical bone in lesions reminiscent of fibrous dysplasia of the bone.(11) In addition, increases in the whole-body BMD of ColI(2.3)+/Rs1+ mice expressing Rs1 were detectable only after 3 weeks of age.(11) To determine if Rs1-induced trabecular bone formation occurs earlier in postnatal growth and if Rs1 affects the patterning of the normal bone structures, we used µCT, histology, and gene expression to analyze bone formation in young mice expressing Rs1 from gestation.
µCT imaging showed no gross differences in skeletal patterning or overall bone shape before 1.5 weeks of age (see Fig. 2). At 1.5 weeks of age, small changes in bone structure were observed, primarily in the long bones of the mice. Since µCT analysis in prepubertal mice is complicated by rapidly changing bone mineral content during postnatal growth, we assessed bone structure by performing histology on femora from ColI(2.3)+/Rs1+ mice at 4 to 6 days after birth.
Histology revealed that femora of the ColI(2.3)+/Rs1+ mice had normal periosteum, bone collar, and bone marrow cells through 4 days of postnatal growth (see Fig. 3A) despite detectable Rs1 expression (see Fig. 3C). However, by 5 days of postnatal growth, femora from the ColI(2.3)+/Rs1+ mice showed increased trabecular bone, decreased cortical bone, and narrowed bone marrow space (see 3A, B). By 6 days, the femora fully recapitulated the trabecular bone replacement of the entire bone as described in 3-week-old ColI(2.3)+/Rs1+ mice(11) (see (see3A3A).
Gene expression analysis from 4 day old ColI(2.3)+/Rs1+ and control mice was used to examine the expression of genes marking mature (osteocalcin and collagen I) and immature osteoblasts (osterix and runx2) (see Fig. 3D–G). Osteocalcin levels were unchanged in the 4-day-old ColI(2.3)+/Rs1+ mice. However, collagen I, osterix, and runx2 all showed increased levels in the ColI(2.3)+/Rs1+ mice despite the lack of an obvious histologic phenotype. By day 5, expression of all these genes was greater in the bones of ColI(2.3)+/Rs1+ mice than in those of wild-type mice. In addition, expression levels of cathepsin K (a marker of osteoclasts) as well as osteoprotegerin (OPG) and RANKL (both regulators of osteoclast activity) were increased in the ColI(2.3)+/Rs1+ mice at both 4 and 5 days of age. The RANKL/OPG ratios, used as a measure of osteolytic balance,(23) were not significantly different in the ColI(2.3)+/Rs1+ mice between days 4 and 5 (1.25 versus 1.11, respectively).
These results show that continuous Gs signaling induced by Rs1 in osteoblasts leads to increased trabecular bone formation starting after day 4 of postnatal growth and increased expression of early osteoblast markers. In contrast, the increased bone formation was not accompanied by a change in the RANKL/OPG ratios.
The prior results show that Rs1 expression occurs before development of the trabecular bone phenotype at day 5 of postnatal growth. In addition, Rs1 expression after 4 weeks of age (ColI/Rs1-late mice, mice,1D)1D) led to no grossly abnormal bone phenotype in adult mice,(11) suggesting a window of sensitivity to Rs1 signaling between postnatal day 5 and 4 weeks of age. We used the tetracycline transactivator system to limit Rs1 expression to narrow the window of maximal osteoblast sensitivity to Rs1 signaling.
To determine if Rs1 expression during the late prenatal to early postnatal period is required to increased trabecular bone formation in adult mice, we suppressed Rs1 expression through birth (ColI/Rs1-early mice, mice,1E)1E) by feeding pregnant mice doxycycline-containing chow. Pups and their nursing mothers were removed from doxycycline on the day of delivery to allow Rs1 expression. The pups were followed by serial whole-body BMD scans through the first 9 weeks of postnatal life. The ColI/Rs1-early mice developed the trabecular overgrowth phenotype, but to a milder degree than in the ColI(2.3)+/Rs1+ mice expressing Rs1 from gestation (see (see4A),4A), as assessed by whole-body BMD.
To further define the timing of Rs1 expression necessary to develop and maintain the bone-overgrowth phenotype, we examined ColI/Rs1–1 week mice in which Rs1 expression was limited to gestation through the first week of postnatal life (see (see1F).1F). These mice did not develop the increased trabecular bone as adults or show whole-body BMDs different from wild-type littermate controls (see (see4B4B).
These results indicate that maximal Rs1-induced bone formation in an adult mouse is strongly influenced by early Rs1 expression during prenatal gestation and early postnatal development. Furthermore, discontinuing Rs1 expression early in postnatal growth but after the trabecularization has been initiated resulted in normal BMD in adult animals, suggesting that growing bone could recover from the Rs1-induced abnormal bone formation if Rs1 expression was discontinued early.
The results for the ColI/Rs1-early mice and our prior observation that ColI/Rs1-late mice (see (see1D)1D) did not develop the grossly abnormal bone through 30 weeks of age(11) suggested that the ability of Rs1 to induce trabecular bone formation decreased as the mice aged despite detectable Rs1 expression in bone. To identify whether the ColI/Rs1-late mice remained free of the bone overgrowth phenotype as the mice aged, we preformed long-term follow-up on these mice through 60 weeks of age.
The ColI/Rs1-late mice showed no significant increases in whole-body BMD until 40 weeks of age. Between 40 and 60 weeks of age, all the ColI/Rs1-late mice showed increased trabecular bone, but only 50% of the mice developed a whole-body BMD greater than 2 SD above the peak wild-type BMD (see (see5A).5A). Sites of abnormal bone formation often were adjacent to regions of normal bone and occurred as distinct nodules (see (see5B).5B). Serum osteocalcin (a marker for bone formation) was increased slightly at 20 and 55 weeks, but serum pyridinoline cross-links (a marker for bone resorption) did not change (see (see5D).5D). These results suggest that Rs1 activity in adult bone primarily increases bone formation over bone resorption. No significant differences in serum calcium or phosphate levels were detected at either 20 or 55 weeks (data not shown).
µCT analysis on femora from 20-week-old ColI/Rs1-late mice showed a subtle expansion of trabecular bone starting at the femoral metaphyses. The increase in trabecular bone was accompanied by a loss of endocortical bone and narrowing of the bone marrow space in the ColI/Rs1-late mice (Fig. 6). Double fluorochrome labeling to assess bone mineralization showed an increase in disorganized trabecular bone formation in the 20 and 55-week-old ColI/Rs1-late mice. In addition, TRAP staining showed a localized increase in osteoclast-like cells within the regions of disorganized trabecular bone in the ColI/Rs1-late mice at both ages despite the normal systemic levels of pyridinoline crosslinks (see (see5D).5D). Von Kossa staining showed that the regions of disordered trabecular bone in the ColI/Rs1-late mice shared similar histologic features with the abnormal bones in the ColI(2.3)+/Rs1+ mice expressing Rs1 from gestation, including a decrease in bone marrow cells and an increase in small osteoblast-like cells by morphology.(11)
These results indicate that delaying Rs1 expression until 4 weeks of age leads to long-term attenuation of the Rs1-induced phenotype. However, a subtle increase in bone formation could be detected as early as 20 weeks of age. Although the bone-overgrowth phenotype was clearly evident by 40 to 60 weeks of age, bone formation was significantly slower in the ColI/Rs1-late mice than in the ColI(2.3)+/Rs1+ mice expressing Rs1 from gestation. In addition, the new trabecular bone appeared to originate from sites of existing trabecular bone, such as at the femoral metaphyses, rather than as de novo formation of trabecular bone at a nontrabecular site, such as at the mid-diaphysis.
The absence of an increased BMD in the ColI/Rs1–1 week mice indicated that the Rs1-induced trabecular bone growth might be reversible if Rs1 expression were discontinued before weaning. To determine if established Rs1-induced bone lesions in adult mice also could be reversed and if continued activation of the Gs signaling pathway by Rs1 was required to maintain the abnormal increase in trabecular bone, we assessed mice in which Rs1 expression was suppressed after 4 weeks of age (ColI/Rs1–4 week mice, see see1G).1G). The ColI/Rs1–4 week mice developed the abnormal increase in trabecular bone, decrease in cortical bone, and loss of bone marrow space during the first 4 weeks of life. After this bone phenotype was established, we suppressed Rs1 expression by feeding the mice doxycycline-containing chow.
ColI/Rs1-4 week femora examined over the next 32 weeks showed a gradual decrease in trabecular bone and a restoration of cortical bone and the bone marrow space (see 7A, B). Double fluorochrome labeling showed that the normal linear bone mineralization pattern gradually returned as the trabecular bone disappeared (see (see7C).7C). TRAP staining also indicated a decrease in osteoclast lineage cells within the bones (see (see7D).7D). Although the overall shape of the bone did not fully return to normal by 36 weeks of age, the gross deformation originally present in the 4- to 6-week-old mice was significantly reversed. Histology on undecalcified bone sections (see (see7E)7E) indicated the return of a normal hematopoietic bone marrow space. In addition, we observed a rapid decrease in the number of morphologically small osteoblast-like cells previously identified in the ColI(2.3)+/Rs1+ trabecular overgrowth bones.
BMDs of individual mice at the start of this experiment at 4 weeks of age were variable (see (see7G),7G), but all BMDs declined or stabilized after Rs1 expression was suppressed. In mice with a starting 4 week BMD of greater than 3 SD above the mean wild-type BMD, whole-body BMDs gradually decreased over 26 weeks, correlating with the reversal of the trabecular bone phenotype observed by µCT imaging (see (see7H).7H). Serum calcium and phosphate levels at 20 and 36 weeks were similar between wild-type and ColI/Rs1–4 week mice (data not shown). Three-week-old ColI/Rs1–4 week mice (before starting doxycycline) showed higher levels of serum osteocalcin but not pyridinoline cross-links than age-matched control littermates (359.20 ± 45.91 versus 257.50 ± 50.15 ng/mL, p = .01, and 1.42 ± 0.31 versus 1.53 ± 0.42 nmol/L, p = .6, respectively; n = 5 mutants and n = 5 controls). Osteocalcin and pyridinoline cross-link levels were both normal in the 20- to 25-week-old ColI/Rs1–4 week mice compared with age-matched control littermates (78.36 ± 44.78 versus 62.91 ± 35.22 ng/mL, p = .5, and 1.53 ± 0.63 versus 1.81 ± 0.89 nmol/L, p = .6, respectively; n = 6 mutants and n = 5 controls).
These results indicate that the abnormal trabecular bone formation and loss of cortical bone observed in the ColI(2.3)+/Rs1+ mice expressing Rs1 from gestation can be reversed in adult mice by suppressing the expression of Rs1. Although the bone did not return fully to a normal structure at the end of the 36 week experiment, our results demonstrate that adult bone maintains a high degree of plasticity for bone remodeling.
In this study, we applied an engineered receptor approach and mouse model of temporally regulated osteoblast-specific Gs activation by the Rs1 RASSL to identify how Gs signals regulate skeletal bone formation in mice of different ages. We focused our experiments on the effects of Rs1 expression without ligand activation because Rs1 has a high degree of basal Gs signaling activity.(10,11) The double-transgenic “Tet-Off” strategy allowed doxycycline-dependent regulation of Rs1 expression and hence control of the basal Gs signaling activity of Rs1. Our studies are especially relevant to the long-term effects of continuous Gs signaling by GPCRs in osteoblasts. We found that the Rs1 phenotype developed rapidly between postnatal days 4 and 6, that progressively delayed Rs1 expression resulted in decreased Rs1-induced bone formation, and that the Rs1-induced bone growth and deformities were dramatically reversed when Rs1 expression was suppressed in adult mice.
Our findings with the ColI/Rs1-early and ColI/Rs1-late mice indicate that osteoblasts from young animals and osteoblasts from older animals both form trabecular bone in response to Gs activation by Rs1. However, delaying Rs1 expression leads to quantitatively different amounts of bone formation. Prior results suggested that Rs1 expression may be lower in adult animals,(11) possibly as a result of age-dependent changes in collagen I promoter methylation.(24) Age-dependent decreases in sensitivity to Gs signaling or cAMP also may contribute because osteoblast adenylyl cyclase activity decreases with age.(25)
A number of GPCRs are expressed in bone, including the parathyroid hormone/parathyroid hormone–releasing peptide (PTH/PTHrP) receptor,(26) endothelin receptor,(27) prostaglandin E2 receptor,(28) adrenaline receptor,(29,30) thyroid-stimulating hormone receptor,(31) and the HTR1b serotonin receptor.(7) Age-dependent activity of these receptors or their ligands may allow other GPCR signaling pathways to modulate the bone-formation effects of Rs1. Activation of the Gi signaling pathway by the Ro1 engineered RASSL(22) or other GPCRs(7,32) is associated with decreased trabecular bone formation, suggesting that the Gi pathway could dampen the bone-formation effects of Gs signaling in older animals. In addition, activation of the Gq pathway leads to osteopenia and may desensitize osteoblasts to Gs signaling.(33,34) Since the PTH receptor signals through Gs, Gi, and Gq, this suggests that the non-Gs pathways balance the Gs-mediated trabecular bone-formation process. The interactions between the different GPCRs expressed in an osteoblast also may account for the phenotypic variability we observed between individual mice.
Age-dependent increases in osteoclast formation(39) also may contribute to the decreased Rs1-induced trabecular bone formation we observed in adult mice. Although increased markers of osteoclasts are evident in the ColI(2.3)+/Rs1+ bone lesions, our results indicate that osteoblast-induced osteoclast formation does not occur during the initial formation of abnormal trabecular bone between days 4 and 5 of postnatal growth. Continued exposure to Rs1 activation eventually may result in increased osteoclastogenesis driven by more mature osteoblasts, similar to what has been reported after PTH treatment of cultured bone marrow stromal cells.(40) As a result, a net increase in osteoclast formation or activity may protect adult bones from the early effects of Rs1.
The bone-formation pattern in the ColI(2.3)+/Rs1+ mice expressing Rs1 from gestation resembles the lesions in fibrous dysplasia patients(35) but to a more severe degree. Clinically, fibrous dysplasia lesions are heterogeneous with abnormal and normal trabecular bone, often at distinct anatomic sites.(36) The ColI/Rs1-late mice in our study particularly resemble this type of lesion, suggesting that the heterogeneous nature of fibrous dysplasia may be a result of changes in osteoblast sensitivity to Gs signaling. This idea that osteoblasts in fibrous dysplasia lesions may have reacquired the characteristics of a perinatal osteoblast correlates with suggestions that fibrous dysplasia may be a stem cell disease.(37) Furthermore, a non-cell autonomous mechanism may contribute to the development of fibrous dysplasia lesions(38) and may similarly modify the osteoblast response to Rs1-mediated Gs activation in an age-dependent manner.
Our finding that Rs1-induced trabecular bone formation occurs during a discrete window of days 4 to 5 of early postnatal growth was surprising because the ColI(2.3) promoter is active starting before birth,(41,42) and Rs1 expression was detectable before any bone changes were evident on histology. This result suggests that the early developmental processes that establish bone patterning and structure prior to postnatal day 4(43) are not significantly affected by Gs signaling. Since the distinct transition at days 4 to 5 of postnatal growth is associated with early weight-bearing activity and the appearance of the secondary spongiosum in mice,(44,45) Gs signaling may act at this point to promote trabecular bone expansion. In contrast, Gi signaling appears to have roles in neonatal bone formation. Earlier results using the ColI(2.3) promoter to express a Gi coupled engineered receptor in osteoblasts before birth resulted in embryonic lethality.(22) In addition, cre-mediated inactivation of the osteoblast GNAS gene encoding the Gsα subunit resulted in severe neonatal skeletal malformations and decreased trabecular bone.(32) These findings suggest that the roles of Gs signaling may shift during early skeletal development from maintenance of the osteoblast niche to driving expansion of trabecular bone.
Our result that the Rs1-induced trabecular bone formation is largely reversible has several important implications. First, bone formation during prepubertal growth can have long-term effects on adult bone, but that bone remodeling may temper these effects. Recent animal studies(46) and clinical studies in children(2) suggest that exercise during childhood can lead to an increase in bone mass that persists with age. The unique nature of the prepubertal skeletal system is further emphasized in that a number of human diseases affecting G protein signaling in the bone have a tendency to present during childhood with long-term sequela in adults.(47–49) However, very early changes in bone formation, as indicated by the ColI/Rs1–1 week mice, may not have long-term effects on the skeleton. Second, continuous activation of the Gs signaling pathway is required to maintain the increase in trabecular bone induced by Rs1. Clinical studies show that bone formed by recombinant PTH (teriparatide) is lost rapidly unless an antiresorptive medication is used when PTH is discontinued.(50,51) Our results indicate that the changes in bone mass are more likely due to changes in Gs pathway activation rather than to the signaling properties of a specific receptor (e.g., PTHR1). Finally, our results show that normal bone has a dramatic ability to repair itself once an insult is discontinued. Since the ColI(2.3)+/Rs1+ mice share strong histologic resemblance to polyostotic fibrous dysplasia(35) and the bone lesions in McCune-Albright syndrome,(35) our results suggest that inhibiting the Gs signaling pathway may be an effective strategy for treating these patients. Our findings also suggest that bones retain information about their normal shape and structure independent of the capacity for skeletal expansion, which could be used during bone remodeling to restore normal bone structure.
In conclusion, bone formation is a highly plastic process with a complex array of signaling inputs that ultimately determine peak bone mass and normal bone architecture. Understanding how bone tissue can be grown and re-formed may reveal basic properties of bone that will be useful in bone regeneration and repair. Our model provides a dramatic example of how osteoblast-specific hormone signaling can completely reshape bone architecture in a reversible fashion. Insight into how hormone signals can alter morphology will help in the development of new ways to modulate the signaling of endogenous GPCRs to reshape bone in a therapeutic setting.
We thank Carlota Manalac, Alyssa Louie, Mark Scott, Gary Howard, Mimi Zeiger, and the Gladstone Histology Core (Jo-dee Fish and Caroline Miller) for valuable technical assistance and discussions. We also thank Scott Kogan and the Center for Genomic Pathology (http://ctrgenpath.net/) slide rounds for valuable histology input. This work was supported by National Institutes of Health (NIH) Grants R01 HL60664-07 (to BRC) and R01 DK072071 (to RAN). ECH received support from an NIH Training Grant (2T32DK07418-26), an NIH career development award (7 K08 AR056299-02), from the California Institute of Regenerative Medicine/Gladstone Institute CIRM Fellowship Program (Grant T2-00003), and from the San Francisco Veterans Affairs Research Enhancement Awards Program. RAN and BPH are senior research career scientists of the Department of Veterans Affairs and received support from the Veterans Affairs Merit Review Program. The J. David Gladstone Institutes received support from National Center for Research Resources Grant RR18928-01.
The authors state that they have no conflicts of interest.