Introduction

The vertebrate skeleton has evolved into a complex articulated organ that classically subserves multiple functions necessary for survival: locomotion, protection, and mineral homeostasis. The complexity of bone is appreciated through a morphology characterized by a light trabecular bone and strong cortical bone framework. However, only through the use of molecular probes and a systems approach will the intricate details of skeletal physiology and its subtle indirect effects on the body be elucidated.

The principle roles of osteoblasts and osteoclasts are bone formation and bone resorption, respectively. These two cell types are functionally coupled (Fig. 1). Osteoblast-derived receptor activator for nuclear factor-κB ligand (RANKL) binds to the osteoclast RANK receptor to facilitate osteoclast differentiation and resorptive activity (Fig. 1). This process is competitively antagonized by the osteoblast- and B-cell-derived decoy receptor osteoprotegerin (OPG), which binds RANKL and prevents it from stimulating osteoclast activity. Conversely, osteoclastic resorption functionally couples to osteoblastic bone formation through the formation of transforming growth factor (TGF) gradients (Fig. 1). A gradient of bone matrix-released TGF acts as a chemoattractant for osteoblast precursors to sites of bone resorption (Fig. 1) [1••].

Fig. 1
figure 1

Coupling between osteoclasts and osteoblasts. a Osteoclastic bone resorption releases TGF-β from the bone matrix, which diffuses out to form a chemoattractant gradient that recruits osteoblast precursors to sites of resorption; these cells differentiate into osteoblasts and subsequently fill in areas of resorbed bone. b Osteoblasts produce the cytokine RANKL to facilitate the differentiation of osteoclast precursors (round cells) into osteoclasts at sites needing resorption. Osteocytes (not shown) are believed to transmit a signal to osteoblasts to cause RANKL production. TGF-β–transforming growth factor-β; RANKL–receptor activator for nuclear factor-κB ligand

Full size image

This two-cell model of bone physiology with layers of regulatory complexity can be investigated as a component of a local skeletal subsystem or as a part of a complete whole-body system. Thus, an integrative physiology paradigm can be used to reconcile shortcomings and gaps in our understanding of how bone affects whole-body responses to various environmental stressors. In other words, for any given physiologic or pathologic state, this approach can reveal the role of the skeletal subsystem without the usual impediments from incomplete knowledge.

There are many spatiotemporally oriented control mechanisms that confer uninterrupted bone remodeling (Table 1). These mechanisms impart to bone the ability to repair microscopic skeletal damage, preserve mineral homeostasis, and enable skeletal participation in multiorgan-level adaptive processes. As a consequence of this integrative network that depends on strong direct nodal connections with bone, any disruption in bone physiology can be presumed to lead to whole-body consequences. Here we will review recent revelations on skeletal remodeling and how intrinsic and extrinsic molecular regulators as well as systemic physiologic regulators interact to prevent excessive bone loss and preserve skeletal structure and function.

Table 1 Multisystem control factors that govern bone remodeling over various time scales
Full size table

Molecular and Cellular Biology of Bone

The intricacy of multiorgan physiology reflects the innate complexity of transcription factors, gene expression, and cellular responses. From a nonteleological standpoint, this recursive pattern of complexity over many spatiotemporal scales may not be a coincidence. By methodically unraveling the molecular mechanisms involved with simple cellular responses, complex whole-body physiology may be parsed out in a way that uncovers vulnerable and potential therapeutic targets.

Osteoblasts

Osteoblasts are bone-forming cells that develop from the mesenchymal stem cell lineage. Wingless-Ints (Wnts) and bone morphogenetic proteins (BMPs) control the spatiotemporal regulation of early osteoblast differentiation. The canonical Wnt pathway and BMP signaling pathway represent important promoters of Runt-related transcription factor 2 (Runx2) gene expression, an obligatory transcription factor required for osteoblast differentiation and bone development [2, 3]. Runx2 -/- mice lack osteoblast differentiation resulting in a concomitant loss of endochondral and intramembranous ossification [4, 5]. Although Runx2 is regulated by many other pathways and molecules, its activation results in upregulation of osteocalcin (OCN), osteopontin, and type I collagen genes, all of which play a major role in the bone extracellular matrix. Wnts activate the canonical pathway by binding low-density lipoprotein receptor-related protein 5 (Lrp5), which interacts with the frizzled receptor complex to inhibit the phosphorylation of β-catenin by glycogen synthetase kinase-3β. The unphosphorylated β-catenin translocates to the nucleus, binds to lymphoid enhancer factor transcription factors, and stimulates osteoblast differentiation [6]. The Wnt inhibitors sclerostin, soluble frizzled-related proteins, Wnt inhibitory factors, and the Dickkoff family members, Dkk1 and Dkk2, all reduce osteoblastogenesis and bone formation through various mechanisms [7–10]. In addition, Twist-1, a basic helix-loop-helix transcription factor, negatively controls Runx2 expression resulting in inactivation of osteoblast precursors [11].

Conversely, the TGF-β superfamily of BMPs acts by binding serine/threonine protein kinase receptors that phosphorylate R-Smad proteins. The R-Smad proteins complex with C-Smad proteins and translocate to the nucleus where they function as positive transcriptional regulators [12]. BMP-2 and BMP-7 specifically activate the mandatory transcription factors Runx2 and osterix in mesenchymal stem cells, increasing osteoblast differentiation [13, 14]. The zinc finger-containing transcription factor, osterix, is similar to Runx2 in that they are both regulated by the anabolic signals BMP-2 and insulin-like growth factor-1 (IGF-1). In addition, both osterix -/- and Runx2 -/- mice lack the ability to form bone [15]. Concomitantly, nuclear factor for activated T cells (NFAT2) is activated by the calcium/calmodulin-regulated phosphatase, calcineurin. Osterix interacts with activated NFAT2 to increase osteoblastogenesis and bone formation [16].

Osteocytes

Osteoblasts become osteocytes when the matrix they secrete completely surrounds them. Osteocytes are terminally differentiated cells that assist in maintaining the bone architecture and communicate with each other through long dendritic processes inside canaliculi. Compartmentalized within the matrix, osteocytes in some unknown way sense and respond to stress, strain, or pressure through changes in fluid flow. However, at the junction of the aforementioned dendritic processes, connexins form membrane-spanning channels that allow intercellular exchange of small molecules between adjacent cells. This provides a possible route for osteocyte communication. For instance, connexin expression in osteocytes increases with mechanical stimulation [17]. In addition, nitric oxide and eicosanoids, products potentially involved in cell-to-cell communication, are released from osteocytes when shear stress is administered, with the possible underlying mechanism being activation of nitric oxide synthase [18].

Osteoclasts

Deregulation of the osteoclast results in excessive resorption and occurs in diseases such as inflammatory arthritis, postmenopausal osteoporosis, Paget’s disease, and malignancy with osseous metastases. The first step in osteoclastogenesis involves PU.1, an ETS domain-containing protein. Mice that are PU.1 deficient fail to generate both osteoclasts and macrophages, illustrating the control PU.1 has over the differentiation of these two cells [19]. The cell subsequently expresses the c-fms gene product macrophage colony-stimulating factor (M-CSF) receptor, and when M-CSF binds, the cell proliferates. The precursor fully commits to the osteoclast lineage upon appearance of c-fos and RANK. Osteoclast differentiation requires direct contact between the precursors and the surrounding osteoblasts, stromal cells, and immune cells. These cells express RANKL, a key requirement for osteoclastogenesis. The action of RANKL on its receptor, RANK, requires two costimulatory signals from the immunoreceptors TREM2 and OSCAR [20]. The activation of RANK results in TRAF-6-mediated activation of nuclear factor-κB and mitogen-activated protein (MAP) kinases. At the same time, the adaptor proteins DAP12 and FcRγ bind to TREM2 and OSCAR, respectively, and recruit Syk kinases, which prompt phospholipase Cγ to release calcium from intracellular stores [21]. The released calcium activates calcineurin, which in turn dephosphorylates the transcription factor NFAT2 [22]. NFAT2 translocates to the nucleus where it acts along with c-fos to amplify the necessary osteoclast genes [23]. The critical importance of NFAT2 in osteoclastogenesis is explained by the overexpression of NFAT2 in RANKL-deficient cells resulting in osteoclast formation [24] and NFAT2-deficient embryonic stem cells failing to differentiate into osteoclasts in response to RANKL stimulation [22].

Once a mature osteoclast forms, it expresses αvβ3 integrins, which bind to RGD-containing matrix proteins in the bone extracellular matrix. The binding results in osteoclast polarization and intracellular signals that produce free radicals. In addition, the osteoclast reorganizes itself creating a ruffled border over the site of resorption and inserting vacuolar, nongastric H+-ATPases into the cell membrane. The H+-ATPases pump acid across into the microenvironment allowing for acid-optimal enzymes, such as cathepsin K, to cleave collagen and release N-telopeptides (NTxs). After transcytosis and extrusion at the dorsolateral surface, NTxs are released into the blood and can be used clinically as markers of bone turnover [25].

Integrative Physiology

The osteoblast-osteoclast coupling model is maintained on the physiologic systems level. Rather than exploring bone integrative physiology using data from component biology and highly controlled scientific experimentation (“bottom–top research”), we flip this investigative paradigm to a “top–bottom” approach using systems biology. We construct an integrative physiologic network involving many, if not all recognized subsystems having direct and/or indirect effects on bone. How could this approach provide novel treatment strategies? The answer may be found in methodologies that incorporate mathematical tools, such as graph theory, network analysis, and analytical inference. These tools, as well as others used to solve problems in systems analysis, are not designed to determine the specific “cause-effect” aspects of binary connections, but rather can determine the overall behavior of complex nonlinear dynamic systems. If new ideas can be spawned by this relatively new and sophisticated approach, then innovative and promising therapies can subsequently be devised and properly investigated.

Endocrine Control of Bone

The endocrine system exerts control over skeletal physiology primarily through the effects of parathyroid hormone (PTH) and 1,25-dihydroxyvitamin D (1,25-D) with other hormones having lesser effects. However, a systems approach to the endocrine control of bone requires examination of each of these hormonal influences and their relevant interactions.

PTH increases bone resorption to maintain basal plasma calcium levels and has obvious evolutionary advantages by maintaining calcium homeostasis. PTH induces osteoclast formation indirectly via an increase in local cytokine expression (RANKL and OPG) from bone marrow stromal cells and/or osteoblasts [26]. PTH does not directly induce osteoclast formation, although some studies have found functional PTH receptors in osteoclasts [27]. Also, in contrast to chronic/tonic hyperparathyroidism, intermittent administration of human recombinant PTH (teriparatide) results in anabolic effects on osteoblasts [28], possibly via a PTH-induced transcriptional expression of osteoblast-derived interleukin (IL)-18 [29].

1,25-D has classically been understood to indirectly act upon bone through control of calcium and phosphate homeostasis. However, 1,25-D also directly acts through the 1,25-D receptors (VDRs) present on osteoblasts [30]. In the early stages of osteoblastogenesis, VDR-dependent signaling inhibits differentiation, but in terminally differentiated osteoblasts, an increase in osteoblast activity occurs from VDR signaling [31]. Therefore, the direct effects of 1,25-D function on osteoblasts are dependent on the stage of differentiation. The neuropeptide calcitonin functions in receptor-mediated inhibition of osteoclasts. Even though calcitonin has been used as an antiresorptive agent for many years, it is not an antiresorptive hormone because its genetic deletion results in a high bone mass rather than osteopenia [32].

Albright, Smith, and Richardson recognized as early as the 1940s that sex steroids regulate anabolic skeletal maturation. Estrogen’s antiresorptive activity is due to both genomic and nongenomic effects on bone marrow and bone cells, leading to decreased osteoclast formation, increased osteoclast apoptosis, and decreased osteoclast-mediated resorption [33]. Conversely, estrogen deficiency elicits slower bone formation resulting in multiple skeletal diseases such as osteoporosis and bone loss [34]. Recent evidence has shown that the osteoporosis and bone loss seen in estrogen deficiency may be due to activation of the immune system, in turn increasing osteoclast formation and bone resorption both directly and indirectly [35]. An early increase in bone resorption and a later decrease in bone formation characterize the hypogonadal state. The late-stage decrease in bone formation occurs primarily because of the estrogen deficiency. In addition, it is now known that the anterior pituitary-derived follicle-stimulating hormone (FSH) causes the early osteoclastic hyperresorption that accompanies hypogonadism [34].

In contrast, another anterior pituitary-derived hormone, thyroid-stimulating hormone (TSH), plays an important antiresorptive role in bone remodeling by reducing the formation, function, and survival of osteoclasts. Sun et al. [36••] show that TSH prevents bone loss and restores lost bone after ovariectomy. Both TSH and FSH act reciprocally through tumor necrosis factor-α (TNF-α) to elicit these effects, mechanisms that will be explored later. However, when suppressed TSH levels are coupled with increased thyroid hormone levels, elevated bone turnover and decreased bone mass result [37]. Although thyroid hormone is anabolic during periods of growth, excess thyroid hormones act in opposition to TSH in bone remodeling. In addition, thyrotoxicosis has been found to cause secondary osteoporosis and fractures [38].

Glucocorticoids have various functions on bone formation and bone resorption. The bone formation effects of glucocorticoids include decreased osteoblast function and number mainly due to decreased osteoblastogenesis, impaired differentiation, and activation of mature osteoblast apoptosis [39]. In addition, activation of osteoclastogenesis occurs through glucocorticoid-induced expression of RANKL and M-CSF as well as a decrease in the OPG decoy receptor. However, these effects seem to be stage dependent because eventually osteoclast numbers decline through decreased osteoblastic signaling and activation of osteoclast apoptosis [39]. Recently, it was shown that dexamethasone-treated mice lacking the glucocorticoid receptor in osteoclast lineage cells failed to develop the bone degradation seen in wild-type osteoclasts [40]. This furthered the idea that osteoblast-osteoclast coupling was important in moderating glucocorticoid action on osteoblasts.

Growth hormone increases the secretion of IGF-1 from the liver, which in turn is an active physiologic stimulator of bone formation. Both osteoblast progenitors and fully differentiated osteoblasts secrete IGF-1. Regardless of the source, IGF-1 is stored in the bone matrix and released during active bone resorption, providing a possible coupling mechanism between bone formation and resorption. Also, an epidemiologic and genetic correlation exists between bone mass and serum IGF-1 [41].

Recently, the duodenal enterochromaffin cell-produced bioamine, serotonin, was found to inhibit bone formation through stimulation of the osteoblast membrane receptor Htr1b and CREB [42]. Interestingly, gut-specific activation of Lrp5 or inactivation of the rate-limiting step of serotonin synthesis, tryptophan hydroxylase 1, results in increased bone mass and prevents ovariectomy-induced bone loss [42].

Bone and the Central Nervous System

Adipose tissue influences bone remodeling via the adipocyte-derived hormone leptin acting on the central nervous system [43, 44]. Mice lacking leptin (B6.V-Lepob/J or ob/ob mice) or its receptor (C57BL/6J-Leprdb or db/db mice) display an increase in bone formation resulting in a high bone mass phenotype [43]. As expected, stimulation of neurons expressing the leptin receptor with centrally administered leptin results in impaired bone formation and low bone mass [43]. These experiments firmly solidify the role for a central leptinergic relay in skeletal regulation. Of note is the fact that the decreased bone formation proves to be independent of a leptin anorexigenic effect [43]. Leptin regulates bone resorption via sympathetic signaling that induces RANKL gene expression in osteoblasts [45••]. Mice lacking adrenergic receptor β2 (Adrb2; the only adrenergic receptor expressed in osteoblasts) or dopamine β-hydroxylase display a high bone mass and lose central leptin inhibitory effects [44, 46]. Interestingly, Adrb2 -/- mice manifest not only enhanced bone formation, but also low bone resorption, indicating that in the absence of direct osteoclastic innervation, the neural effects on osteoclastic bone resorption are indirect and probably mediated through cocaine- and amphetamine-regulated transcript (CART) signaling [44]. Furthermore, osteoblasts control the development of hematopoietic stem cell precursors into the osteoclast lineage through adrenergic signaling [47]. This evidence provides strong support that central leptinergic neurons and peripheral sympathetic nerves regulate not only osteoblastic bone formation, but also the maturation and function of bone-resorbing osteoclasts.

Most intriguing, however, is the finding that this adrenergic stimulation in osteoblasts provides a circadian rhythmicity to bone remodeling by affecting the clock genes Per and Cry. These clock genes inhibit G1 cyclin expression and cause a decrease in osteoblast proliferation [48]. The circadian control of bone remodeling may explain the diurnal variation in human bone resorption markers [49]. Despite the support that bone mass is regulated by the leptinergic-sympathetic axis, it remains unclear how much regulation is conferred by additional neuronal systems.

Immune System, Cytokines, and Bone

For many years, the inflammatory response has been known to be associated with bone loss. However, only recently have aspects of the overall mechanism of osteoimmunology (interactions between bone and the immune system) been elicited through the discovery that T lymphocytes regulate osteoblast and osteoclast formation, life span, and activity [50] via the production of cytokines. TNF-α increases osteoclast formation by upregulating production of bone stromal cell-derived RANKL and M-CSF [51] and also increases the responsiveness of osteoclast precursors to RANKL [52]. In addition, TNF-α increases osteoclast activity and decreases osteoblastogenesis [53]. Thus, TNF-α proves to be an important proresorptive regulator of bone remodeling.

Taking a step backward, multiple regulators of TNF-α exist. As stated above, FSH (proresorptive) and TSH (antiresorptive) reciprocally regulate macrophage and T-cell production of TNF-α [54, 55•]. Mice that are both TSH receptor -/- and TNF-α -/- do not develop the increased bone loss that is seen in mice that are only TSH receptor -/- [55•]. Additionally, hypogonadism increases T-cell-produced TNF-α, the number of T cells in the bone marrow, the number of T cells produced in the thymus, as well as IL-7 production in B cells [52, 56, 57]. Furthermore, a decrease in estrogen leads to an increase in adaptive immunity, characterized by an increase in T-cell-secreted TNF-α. The increase in TNF-α increases osteoclast formation both directly and indirectly [58]. The increased bone resorption seen in estrogen deficiency is now mainly viewed as a cytokine-driven increase in osteoclast formation.

In addition to TNF-α, many other cytokines exist that aid in bone remodeling, complicating further the interactions between the immune system and bone. IL-1 stimulates osteoclastic bone resorption. The proinflammatory cytokine IL-6 induces RANKL on mesenchymal cells and T cells, thus contributing to osteoclast activation [59]. IL-7, a lymphocyte-produced cytokine, has a strong association with bone destruction in vivo by modifying osteoclastogenesis [60]. Interferon-γ has been shown in vitro to inhibit osteoclastogenesis [61], while in vivo it has been shown to foster bone resorption [62] through increased antigen presentation and T-cell activation. These disparate effects can be attributed to direct and indirect effects of interferon-γ on osteoclast modification [58]. Mainly produced by estrogen-stimulated osteoblasts, TGF-β’s role in bone remodeling had proven complex with conflicting in vivo and in vitro data as to whether bone resorption is increased or decreased [63]. However, recent evidence elucidates TGF-β as an important chemoattractant of osteoblasts to bone resorption sites [1••].

B lymphocytes and dendritic cells are also important cellular effectors of bone remodeling. The crosstalk between T cells and B cells generates a wide variety of cytokines that have diverse and complicated effects on bone as well as other immune cells. Although osteoblasts and stromal cells are generally thought to be the main sources of OPG, recent in vivo experiments discovered B lymphocytes to be the predominant manufacturer of this decoy receptor in bone [64]. By regulating B-cell production of OPG through CD40-CD40 ligand costimulation, T cells provide protection for bone homeostasis [65••]. In addition, crosstalk between T cells and dendritic cells occurs through a CD80-CD28 costimulatory mechanism that results in T-cell activation and subsequent TNF-α production [66].

Cardiovascular System and Bone

In the realm of component biology, a direct cardiovascular-bone relationship has been poorly explored. However, in the realm of systems biology, there are many potentially important indirect relationships between these two subsystems. In general terms, both subsystems are influenced by mineralization phenomena [67], steroid hormone [68], thyroid hormone [69], growth hormone [70], leptin [43, 71], catecholamines [72], PTH [73], cytokines, and chemokines [74], as well as environmentally relevant levels of endocrine disruptors involving homocysteine [75], lipids [76], and vitamins D [31, 73] and K [77]. Other studies have linked atherosclerosis to osteoporosis and vice versa [78]. Both of these diseases have many of the same risk factors including hypertension, diabetes mellitus, smoking, alcohol abuse, and a low level of physical activity. Moreover, many pathogenetic pathways link the two subsystems as well. Hydroxyapatite, a main component of the mineral phase of bone, is also present in the calcium deposits within atherosclerotic lesions. Preosteoblast cells, preosteoclast cells, and bone matrix proteins such as osteopontin and BMP-2 have also been found in arterial plaques [79]. A persistent inflammatory process characterizes both diseases, and the inflammatory cytokine IL-6 is associated with cardiovascular mortality [80] as well as osteoclast-stimulated bone resorption [81].

Surprisingly, medications traditionally used to treat cardiovascular pathologic conditions were also found to help prevent bone-related conditions. Beta-blockers, mainly used as antihypertensive drugs, reduce the risk of bone fractures in human studies [82] by blocking the sympathetic nervous system’s stimulation of bone resorption. In addition, statins, which are widely prescribed to lower low-density lipoprotein cholesterol levels and reduce the risk for atherogenic vascular disease, have anabolic effects in osteoporosis [83].

Energy Metabolism, β-cells, and Bone

Although adipose tissue regulates bone via leptinergic effects on the sympathetic nervous system (vide supra), recent studies show that bone regulates energy metabolism through OCN production by the osteoblast [84]. Although OCN is embedded in the bone extracellular matrix, past experiments failed to yield a function in this setting. What was found though was that the Esp gene-encoded tyrosine phosphatase, osteotesticular protein tyrosine phosphatase (OST-PTP), favors the carboxylation of OCN, modifying it into the specialized bone matrix gla protein [45••]. However, a small amount of OCN evades OST-PTP and subsequently is secreted into the general circulation [45••]. Using Esp-/- and OCN-/- genetically altered mice, it was determined that the noncarboxylated, secreted OCN favors pancreatic β-cell proliferation and the expression of genes encoding insulin and adiponectin, thus increasing insulin secretion and insulin sensitivity [45••].

Another example of the multisystem involvement of bone regulation is 1,25-D’s direct and indirect effects on β-cell function and insulin secretion. Both the VDR and 1,25-D-dependent calcium-binding proteins are expressed in β cells and pancreatic tissues, respectively [85]. Studies looking at 1,25-D supplementation have shown improvements in glucose tolerance in patients with 1,25-D deficiency [86] and improvements in insulin response in female patients with type 2 diabetes [87]. In addition, patients with osteomalacia undergoing long-term 1,25-D treatment develop improved insulin secretion and glucose tolerance [88]. One mechanism that might explain this effect is that 1,25-D raises the intracellular calcium concentration through nonselective voltage-dependent calcium channels [89]. Consequently, calcium-dependent endopeptidases then participate in the cleavage of proinsulin to insulin [90], effectively increasing insulin secretion. 1,25-D also has been suggested to increase insulin secretion through direct modulation of β-cell growth [91].

1,25-D acts directly to increase expression of the OCN gene [92], bone γ-carboxyglutamic acid-containing protein (BGLAP) gene. As stated above, uncarboxylated, secreted OCN increases insulin secretion from pancreatic β cells. However, despite previous efforts, a β-cell-associated OCN receptor remains elusive. Moreover, serum OCN levels are dramatically decreased in type 2 diabetic patients and normalize after tight glycemic control [93]. These ideas raise two questions that may shed light on the bone-β-cell connection: 1) What is the OCN receptor and mechanism of OCN-stimulated insulin secretion? and 2) Does 1,25-D deficiency aid in the onset of diabetes, and if so, can we prevent diabetes by preventing 1,25-D deficiency? These questions are more intuitive than one might initially think because it has long been suspected that 1,25-D insufficiency is a risk factor for type 1 diabetes [94].

Complicating the matter further, the connections between the immune system, 1,25-D, and pancreatic β cells appear to be multifold and complex. TNF-α and TNF-β, IL-6 and its receptor, C-reactive protein, and plasminogen activator inhibitor-1 have all been shown to be abnormal in type 2 diabetic patients. TNF-α and IL-6, in addition to being proresorptive cytokines for bone, may also directly interfere with insulin signaling. The finding that VDRs existed in activated T cells, macrophages, and thymic tissue proposed the idea that 1,25-D had in addition to its stage-dependent bone-remodeling role, an immunomodulatory role as well [90, 95]. This idea led to the discovery that 1,25-D elicits a wide range of effects on the immune system, some of which may be blunted during the onset of type 2 diabetes. These effects are generally anti-inflammatory in nature and include reducing the antigen-presenting activity of macrophages to lymphocytes, preventing dendritic cell maturation, and inhibiting T-cell-mediated immunoglobulin synthesis in B cells [96–98]. However, despite reports that 1,25-D downregulates the production of multiple cytokines including the pro-bone resorbing/anti-insulin-secreting IL-6 [99], the idea that 1,25-D and the immune system have an integrated role in the onset of type 2 diabetes still remains largely speculation.

Frequency Modulation and Amplitude Modulation of Bone Remodeling Signals

It has been hypothesized that bone remodeling is controlled through both frequency modulated (FM) and amplitude modulated (AM) signals that act upon the osteoblast and osteoclast (Fig. 2) [34]. Osteoblast-mediated bone formation appears to be regulated by both mechanisms. Negative fast FM signals likely arise from central leptinergic stimulation of the peripheral sympathetic nervous system that controls the clock genes Per and Cry through activation of β-adrenergic receptors. In contrast, positive fast AM signals likely exert their influence through various systemic regulators, principally IGF-1.

Fig. 2
figure 2

Systems biology representation of osteoblast and osteoclast function. Osteoclastic bone resorption is controlled through a combination of both fast and slow AM to ensure maintenance of skeletal homeostasis. With the exception of FSH as a direct, positive, fast amplitude modulator of osteoclastic bone resorption, most positive regulation of osteoclast resorption occurs through an osteoblast-mediated, two-step process. This slow AM combined with fast amplitude negative regulation of osteoclast function provides an evolutionarily designed mechanism to prevent against excessive bone loss. Osteoblast function is controlled through both slow AM and a neural-derived FM mechanism using SNS stimulation. 1,25-D—1,25-dihydroxyvitamin D3; Adrb2—adrenergic β2 receptor; AM—amplitude modulation; CT—calcitonin; E2—estrogen; FM—frequency modulation; FSH—follicle-stimulating hormone; IGF-1—insulin-like growth factor-1; OPG—osteoprotegerin; PTH—parathyroid hormone; RANKL—receptor activator for nuclear factor-κB ligand; SNS—sympathetic nervous system; T3/T4—triiodothyronine and thyroxine; TSH—thyroid-stimulating hormone

Full size image

In contrast, osteoclast-mediated bone resorption is controlled through slow and fast AM mechanisms that counterbalance osteoblastic bone formation and ensure skeletal homeostasis. The activation of osteoclastic bone resorption predominantly occurs through a slow AM “system” activation of osteoblasts. Most cytokine and hormone-mediated bone resorption results from this indirect type of osteoclast stimulation that requires osteoblast intermediation. Additional osteoclast stimulation can be provided directly by FSH as a fast AM signal. Similarly, cessation of osteoclast activity also results from fast AM signals. In this case, TSH, calcitonin, estrogen, and local calcium concentration act as direct inhibitors of osteoclast activity. Thus, a hybridization of indirect two-step activation of resorption coupled with a one-step direct activation/deactivation is evolutionarily advantageous. This integrative systems network prevents overstimulation of osteoclast action and subsequent loss of bone mass.

Conclusions

Through recent advances, it is now known that bone homeostasis is regulated by more than just PTH, 1,25-D, and calcitonin. Rather, a complex number of whole-body controls, including immune, neuronal, and metabolic signals, provide mechanisms that regulate skeletal and mineral balance. In this review, we summarize an integrative physiologic model of bone regulation that is controlled by both FM and AM signals. These regulatory signals enable a precise recursive spatiotemporal balance to be struck between bone formation and resorption, thus providing homeostatic control over bone remodeling. Further research and advancements are needed to provide explanations of how bone functions in physiologic systems. In addition to highly controlled scientific experimentation to elucidate mechanistic explanations, other methods of exploration include mathematical modeling and network analysis for ultimate explanations. Bidirectional research using both of these modalities will likely be the future of rewarding biomedical research.