1 Introduction

The deeply dynamic nature of bone may be missed by a less attentive eye. Bones are resilient and apparently quite rigid structures. They vary in shape, size and number and are divided in the axial and appendicular skeleton. Through life, they are subjected to loads and strains that temper their shape, with old matrices being replaced by newly formed ones, maintaining bone volume and strength. When trauma and fractures occur, bones are capable of healing if enough stability and proper alignment are guaranteed.

Osteogenesis, bone repair and remodelling are directed by the exchanges involving the environment, cell-to-cell interactions and cell–extracellular matrix.

Mechanical forces are crucial in early embryonic development. Morphogenesis is controlled through fluid flow mechanisms and by cellular contractility. Early embryo shaping depends on morula contraction determined by cohesivity. Multiple layers result in the development of endoderm, mesoderm and ectoderm in the blastula [1,2,3,4]. Mechanical forces, cell geometry, and oriented cell division together orchestrate normal airway tube morphogenesis [5] and may help determine the neocortical organization [6]. Cells generate tension through contraction of actin-myosin cytoskeleton filaments, which are transmitted through cadherin-mediated adhesion sites to surrounding structures, these being either cells or extracellular matrix. The cytoskeletal conformation and cell shape stress-dependent changes regulate cell phenotype; interactions with the extracellular matrix are of paramount importance for cell phenotype [2]. Organ lateralization and asymmetry depend on unidirectional fluid flow, generated by specialized motor protein complex dynein. The fluid flow induces differences in key molecules’ expression (such as the TGF-family signalling molecules) [7,8,9,10]. Lateralization may also depend on fluid shear, in the embryo, by acting on a group of non-motile cilia, coupled to calcium channels; fluid flow generated shear may cause the intracellular calcium concentrations to increase and initiate the cascade of events responsible for lateralization [11].

The mechanical environment is also a determining factor for vasculogenesis, angiogenesis [12, 13] and neuronal development [14,15,16,17].

The embryo mesoderm is constituted by spindle- or star-shaped cells called mesenchymal stem cells (MSCs). MSCs are the utmost pluripotential cells in the organism, originating different tissues such as the connective tissue, muscle, cardiovascular tissue and the whole skeletal system. Bone, cartilage, tendons and ligaments develop through mechanisms of proliferation, migration and differentiation, but also programmed cell death/apoptosis [18].

We now begin to address how complex bone is in its functions, how its macro-architecture, microarchitecture and arrangement at molecular level play together remarkably, ensuring its responsiveness to external and internal stimuli and close entwining with other organs.

2 The Complexity Behind Simplicity

2.1 Bone Functions

Osseous tissue is the most rigid and resilient tissue of the body. Bone is composed of dense connective tissue; it is the primary skeleton component, thus providing structure, support and protection to vital organs, like the brain (skull), the spinal cord (vertebrae) and the heart and lungs (ribs and sternum). Vertebrae also participate in the spine shock absorbance—providing adequate load cushioning for the fibrocartilaginous joints at the intervertebral discs. Long bones provide structure, stability and, along with the joints, enable body movement—providing levers for the muscles.

Moreover, bones act as the major source of blood, since haematopoiesis occurs in their medullary cavity. In infants, the bone marrow of all long bones is capable of blood synthesis. With ageing, part of the red marrow turns into yellow fatty marrow, no longer capable of haematopoiesis. Functional red marrow in adults is limited to the vertebrae and the extremities of femur and tibia.

Bones also partake a vital role as:

  • Mineral storage: mostly calcium, phosphate and magnesium; bone plays an important metabolic role, mediated by several hormones, regulating mineral homeostasis [19,20,21].

  • Acid-base balance, as bone can buffer blood against extreme pH changes by absorbing or releasing alkaline salts, through bone cells’ activity [22,23,24].

  • Osteoblasts have been shown to produce growth factors, with production regulated by systemic hormones and local mechanical stress [25]. The bone matrix holds several growth factors such as insulin-like growth factors I and II, transforming growth factor beta, acidic and basic fibroblast growth factor, platelet-derived growth factor and bone morphogenetic proteins, released when resorption occurs [26].

  • Adipose tissue storage (yellow bone marrow functions as a fatty acid/energy reserve) [27,28,29].

  • Heavy metals and other foreign elements, after detoxification from the blood, are stored in bone and later excreted [30, 31].

  • Bone functions as an endocrine organ, as it produces two known circulating hormones:

    1. a.

      Fibroblast Growth Factor 23 (FGF23): FGF23 was first described by Yamashita et al., and it is produced mainly by osteocytes [32, 33], but also by osteoblasts [34]. FG23 acts on the kidneys, inhibiting 1α-hydroxylation of vitamin D and promoting phosphorus excretion in urine [35,36,37]. FGF23 also decreases phosphorus absorption in the intestine, regulating inorganic phosphate metabolism and thus mineralization [35]. Serum calcium concentration regulates FGF23 production [38], thus making FGF23 into a calcium–phosphorus regulatory hormone [39]. Hence, FGF23 excess or deficiency results in anomalies of phosphate metabolism. Excess FGF23 hinders renal phosphate reabsorption and 1,25 dihydroxy vitamin D3 [1,25(OH)2D] production, causing hypophosphatemia and suppression of circulating 1,25(OH)2D levels and, eventually, rachitic changes in bone [40]. These changes occur in autosomal dominant hypophosphatemic rickets and osteomalacia [41] and, in tumour-induced osteomalacia (TIO), a paraneoplastic syndrome [42]. In contrast, reductions in FGF23 cause tumoral calcinosis syndrome, characterized by hyperphosphatemia, increased 1,25(OH)2D and soft tissue calcifications [40, 43]. An obligate FGF23 co-receptor was identified—Klotho [44]. Klotho is essential to activate FGF receptors and their signalling molecules. Secreted Klotho suppresses either by direct interaction or interference with receptors, the activity of several growth factors: insulin, insulin-like growth factor-1 (IGF-1) [45], Wnt [46] and TGF-β1 [47]. The FGF23-Klotho axis represents a specialized system responsible for the external and internal calcium and phosphorus balance in the bone, intestine and kidney. FGF23-Klotho axis works under parathormone regulation, with parathormone increasing serum FGF23 levels and directly promoting FGF23 expression by osteocytes [48, 49]; FGF23 exerts negative feedback by inhibiting the parathyroid glands [50, 51]. FGF23 production in the osteocyte may be inhibited by osteopontin [52].

    2. b.

      Osteocalcin is a protein produced by osteoblasts in bone, and it is a major regulator of insulin secretion by direct action over the pancreatic β-cell. Osteocalcin also increases insulin sensitivity of peripheral tissues, e.g. muscles and liver, up-regulating glucose uptake and energy expenditure, thus contributing to glycaemia regulation [53,54,55,56]; it also reduces fat deposition by inducing adiponectin secretion by adipocytes [57, 58]. Blood osteocalcin levels are significantly lower in diabetic patients when compared to non-diabetic controls, and osteocalcin levels are inversely related to fat mass and blood glucose [59, 60]. Lastly, osteocalcin influences male fertility, by enhancing testosterone production by Leydig cells in the testes [61].

2.2 Bone Structure and Mechanical Behaviour

Bone is a composite material; the inorganic portion of bone comprising 70% (of which 95% is hydroxyapatite and 5% are impurities impregnated in hydroxyapatite), whilst 22–25% are organic (of which 94–98% are mainly collagen type I and other non-collagen proteins and 2–5% are cells); 5–8% is water [62].

Bone mechanical properties depend on porosity, composition, mineralization degree and organization of solid matrix. Therefore, the mechanical behaviour of an entire bone is highly dependent on its properties at a microscale [63, 64].

Bone can be classified according to its structural features at a microscopic level in woven and lamellar bones.

Woven bone is immature or pathologic, primary bone, and it is present in growth, fracture healing and diseases such as Paget’s disease. Cells and matrix are laid randomly. Woven bone is formed during intramembranous, endochondral or rapid appositional bone growth. In large animals (whether reptiles, birds or mammals), woven bone with large vascular canals is rapidly deposited in the subperiosteal region. Canals are lined with osteoblasts that gradually deposit lamellae until the canal has a reduced diameter; the resulting structure is a primary Haversian system or osteon. The random distribution of its components explains woven bone’s isotropy.

Lamellar bone is organized, mature bone, and it is morphologically classified into two different types: cortical or compact bone and cancellous or trabecular bone. Cortical and cancellous bone types differ in both structure and functional properties, but both are highly anisotropic.

The typical structure of a long bone, such as the femur or the humerus, comprises the cylindric shaft, or diaphysis, and the extremities, or epiphyses (Fig. 1). The outer surface is covered by a layer of dense connective tissue called periosteum, except for the areas of mobile articulation, covered with hyaline cartilage. The periosteum is highly vascularized and responsible for appositional bone growth. The endosteum is a thin layer of connective tissue that lines the inner surface of the diaphysis, containing the medullary canal. The epiphysis consists of an outer layer of cortical bone surrounding the porous network formed by trabecular bone. Within the spaces between trabeculae lays red bone marrow [65]. Long bones, fundamental for load bearing and leverage, evolved as structures in which stiffness along the long axis was favoured.

Fig. 1
figure 1

Illustration of a long bone structure, showing the distribution of two different types of lamellar bone: cancellous and cortical compact bone

Full size image

Cortical bone (Fig. 2) accounts for approximately 80% of the skeletal mass. Cortical bone is vital to skeletal mechanical competence, both of long and flat bones. It is formed by tightly aligned collagen fibrils, making concentric lamellae. Each lamella is 2–3 μm thick and is arranged in distinct layers of parallel fibrils, each layer with a different fibril orientation [66]. Mineralization occurs by apatite crystals (mainly carbonated apatite) deposition within and around these fibrils. The lamellae form cylinders containing a hollow central canal where blood vessels and nerves run, composing the cortical bone microstructural unit, called Haversian system or osteon. From the centre of the osteon (Haversian canals), blood vessels form a three-dimensional network and penetrate the cortical bone layer perpendicularly (running within Volkman’s channels) [67]. In between, the osteons are incomplete osteons, known as interstitial systems or interstitial bone.

Fig. 2
figure 2

Microphotograph of cortical bone in proximal tibia (undecalcified bone section of sheep tibia, Giemsa-Eosin, 40× magnification; slide digitalized using NanoZoomer SQ, Hamamatsu Photonics, Portugal). Haversian systems are evident, as are the concentric lamellae. Osteocytes are visible in their lacunae, in between lamellae

Full size image

Cancellous (or trabecular) bone is highly porous and adapted to compressive loads. The lamellae are organized in a parallel manner, forming trabeculae. These rod- and plate-shaped struts are organized into a flexible lattice with variable degrees of interconnectivity (Fig. 3).

Fig. 3
figure 3

Image of trabecular bone (undecalcified bone section of the proximal epiphysis of sheep’s tibia, Giemsa-Eosin, 5× magnification; slide digitalized using NanoZoomer SQ, Hamamatsu Photonics, Portugal). The picture illustrates the sponge-like structure of cancellous bone

Full size image

The trabecular network is light and of utmost importance for load transfer in long bones, absorbing and distributing sudden stresses. In vertebrae, cancellous bone is the main load-bearing structure and essential for shock absorption. Trabeculae are approximately 200 µm thick and are orientated according to routine load-bearing direction [68]. This is evident in epiphyses and metaphyses of long bones, but also in the vertebrae and ribs. Trabeculae are covered by osteoblasts and bone-lining cells. Osteoblasts actively lay extracellular matrix (ECM), and bone-lining cells are in an inactive state. The metabolic rate of trabecular bone is higher than that of cortical bone and the remodelling phenomena more prominent. [18, 69].

Bone endures both compressive and tensile stresses. Bone is subjected to bending and torsion [62]. In humans, there is a large variation in strains, ranging from to 400 to 2000 μstrains or even as high as 4000 μstrains [62, 70, 71].

The bone exhibits a stress–strain response of sequential elastic and plastic responses. In its elastic region, no permanent damage is caused to the bone structure; if the stress increases, a gradual transition to a plastic response occurs. Post-yield deformations are permanent and cause trabecular fracture, cement lines and cracks. Crack formation and growth allow energy dissipation and are a powerful stimulus for bone remodelling in healthy bone.

The mineral component contributes to compression strength, while collagen fibrils are fundamental for tensile strength. The mineral phase is highly related to stiffness, whilst collagen is determinant for toughness [72]. A higher Young’s modulus corresponds to less ductility and higher brittleness [73].

Bone material properties reflect, therefore, high functional specialization and depend on architecture, composition and component spatial distribution.

2.2.1 The Bone Matrix

Structure and material properties of the bone depend on collagen. The collagen I molecule is composed of three long peptide sequences, arranged helicoidally. Collagen is produced by osteoblasts and goes through several enzymatic modifications whilst still within the cell [74]. After leaving the cell, collagen molecule undergoes further cross-linking within itself and with other collagen molecules. Collagen chain mutations lead to diseases such as osteogenesis imperfecta [74, 75]. The triple tropocollagen units are aligned in fibrils and display a permanent dipole moment. Consequently, collagen acts as a piezoelectric and pyroelectric material and as an electromechanical transducer [76, 77]. The native polarity and the piezoelectric properties of collagen are associated with the mineralization process. Under compression, negative charges on the collagen surface become uncovered and attract calcium cations, which are then tailed by phosphate anions [77, 78]. Collagen can actively control mineralization, functioning in synergy with other non-collagenous proteins, inhibitors of hydroxyapatite nucleation. The positive net charge close to the C-terminal end of the collagen molecules promotes the infiltration of the fibrils with amorphous calcium phosphate; at the gap and overlap regions of the collagen molecule, the clusters of charged amino acids form nucleation sites and the amorphous calcium phosphate are changed into parallel oriented apatite crystals [79].

Non-collagenous proteins such as osteopontin, fibronectin, osteonectin and bone sialoprotein are present in much smaller quantities but are, nonetheless, essential for normal bone function and properties.

Osteopontin (OPN) is a non-collagenous glycoprotein, present in the bone matrix, binding to the cell surface and hydroxyapatite. It is mostly produced by proliferating pre-osteoblasts, osteoblasts and osteocytes, but also by fibroblasts, osteoclasts and macrophages [80, 81]. OPN intervenes in cell migration, adhesion and survival in diverse cell types. OPN is a key player in bone remodelling processes. Its production is modulated by mechanical loading, being up-regulated both by loading and by loading deprivation [81,82,83]. OPN has been proved to inhibit mineralization, but its deficiency significantly lessens bone fracture toughness and causes anomalous mineral distribution, leading to increased FGF23 production [52, 84,85,86].

Fibronectin mediates many cellular interactions with the ECM, playing an important part in cell adhesion, migration, growth and differentiation. It is determinant for vertebrate development and is mostly synthesized by osteoblast precursors and mature bone cells; it can also be produced at distant sites (such as the liver) and enters the systemic circulation. Some studies suggested that only circulating fibronectin exerts effects on the bone matrix [74, 87]. Fibronectin binds to collagen and may act as an extracellular scaffold, facilitating interactions of BMP1 with substrates [88]. Fibronectin may also be vital for the osteogenic differentiation of mesenchymal cells [89, 90].

Osteonectin or secreted protein acidic and rich in cysteine (SPARC) is secreted by osteoblasts during bone formation, and it is one of the most abundant non-collagenous proteins in the bone matrix. Osteonectin is a regulator of bone mineralization; its attachment to collagen can inhibit or promote mineral formation. It interacts also with apatite through its N-terminal domain, inhibiting crystal growth [91]. Osteonectin knockout mice suffer from osteopenia due to osteoblasts and osteoclasts defective function and low bone turnover. Changes in the osteonectin encoding gene have also been linked to idiopathic osteoporosis and osteogenesis imperfecta [92].

Thrombospondin-2 (TSP-2) is another matricellular protein that also exerts its effects on osteoblast proliferation and function, being involved in MSCs adhesion and migration; it has also influence on angiogenesis and tumour growth and metastization [93,94,95,96]. TSP-2 likely participates in bone remodelling, since it promotes osteoclastogenesis through the RANKL-dependent pathway [95, 96].

Bone sialoprotein (BSP) is a highly glycosylated and sulphated phosphoprotein that is found almost exclusively in mineralized connective tissues [97]. BSP knockout mice have higher trabecular bone mass and reduced amounts of cortical bone; they also present a very low turnover. BSP defective mice maintain unloading bone response, as opposite to OPN knockout mice [98]. The absence of BSP also leads to changes in the growth plates, decreased bone length and delayed ossification [99]. BSP and OPN are part of the small integrin-binding ligand N-linked glycoproteins (SIBLING) family, and the recent studies suggest the interplay in between these proteins is determinant in bone biology [100].

Proteoglycan (PG) encoding genes are expressed in skeletal and non-skeletal tissues but with stronger expression in bone, joints and liver. PGs are a large family of molecules and perform many biological functions. PGs help to structure bone by mediating collagen secretion and fibril organization; they also act as mineralization inhibitors. PGs also modulate cytokines and growth factors’ biological activity in bone [101]. In bone, PrG4 gene expression is under the control of PTH [102].

From the reviewed above, it becomes clear that non-collagenous and collagen matrix proteins are fundamental for bone morphology and material properties, interacting with each other and with cells and responding to stimuli generated locally or systemically. It is also evident that matrix components have a multiplicity of functions. The role of a molecule is modulated by changes in its structure and by interactions with other substances.

2.2.2 Bone Cell Population

Mature bone contains three core cell populations: osteoblasts, osteocytes and osteoclasts.

2.2.2.1 Osteoblasts

Osteoblasts arise from MSCs, sharing a common background with chondrocytes, myoblasts and fibroblasts. Osteoblasts differentiate under the influence of a variety of hormones, cytokines and the local mechanical environment [103]. These cells, when active, are cuboidal/round (Fig. 4), with specific features consistent with their secretory functions, such as prominent Golgi complexes and endoplasmic reticulum (with multiple vesicles and vacuoles); these are even more evident during matrix secretion and early stages of mineralization [104].

Fig. 4
figure 4

Osteoblasts are cuboidal cells that when actively deposing matrix on bone surfaces (microphotograph, decalcified sheep bone section, Mason trichrome, magnification 40×; black arrows point line of active osteoblasts). When quiescent, osteoblasts appear as flat bone-lining cells

Full size image

Osteoblasts can also remain on bone surfaces as flat bone-lining cells, in a quiescent state, with few apparent cell organelles. During the osteoblast maturation process, there are increased levels of expression of pro-collagen, osteopontin and osteocalcin; bone sialoprotein seems to be more strongly expressed at intermediate phases of differentiation [105, 106]. Osteoblast differentiation is impaired when gap junctions are inhibited, suggesting communication to neighbouring cells is essential for differentiation [107]. Osteoblasts produce non-mineralized matrix—osteoid—that becomes gradually mineralized, wherein they become trapped and some differentiate into osteocytes. Runx2 induces the expression of major bone matrix protein genes in vitro. Runx2 expression is up-regulated in pre-osteoblasts, being maximal in immature osteoblasts and down-regulated in mature osteoblasts. Although Runx2 is weakly expressed in undifferentiated mesenchymal cells, it induces their osteogenic commitment [108]. Once Runx2 is activated, cells undergo the three stages of differentiation, with the synthetization of different molecules: in stage 1, the cells proliferate and express fibronectin, collagen, TGFβ receptor 1, and osteopontin; during stage, osteoblast will differentiate and act on the extracellular matrix through alkaline phosphatase and collagen; at stage 3, the osteoblast will assume its characteristic cuboidal shape and secrete significant amounts of osteocalcin. Osteocalcin will promote matrix mineralization [109]. Osteoblast differentiation is influenced by 1,25(OH)2D3 and mechanical stimuli, amongst other factors [110].

2.2.2.2 Osteocytes

Osteocytes are the most abundant cells of bone, comprising more than 90% of the osteoblast lineage and contributing to bone formation and resorption [111, 112]. They are fully differentiated osteoblasts embedded in the mineralized matrix, inside the osteocytic lacunae. Lacunae are located between the lamellae and connected with surrounding lacunae by a canalicular system (Fig. 5). Osteocytes have long dendritic cell processes (50–60 by cell) that lay within the canaliculi. The extremities of the cell processes connect osteocytes amongst themselves and allow contact with osteoblasts and bone-lining cells [18, 113, 114]. The resulting functional syncytium shares a common environment [115].

Fig. 5
figure 5

Detail of microphotograph of an undecalcified bone section of sheep tibia, Giemsa-Eosin, on the left, showing osteocytes (Giemsa-Eosin, 40× magnification and 200% zoom; slide digitalized using NanoZoomer SQ, Hamamatsu Photonics, Portugal). The canaliculi where cell processes run are observable. The image on the right illustrates a simplified version of the resulting three-dimensional syncytium

Full size image

Osteocytes have no matrix secretion functions; however, they are responsible for sensing changes in the bone structure and commanding bone remodelling.

Pre-osteoblasts and osteoblasts are less responsive to fluid shear stress than osteocytes. Mechanosensitivity seems to increase during differentiation. However, osteoblasts are able to modulate their response according to the mechanical stimuli intensity [62]. Osteocyte functions include mechanosensing and maintaining bone matrix [116, 117]. The sensation of electrical signals may be one of the functions of osteocytes, and electrical signals mediated by osteocytes may regulate the cell behaviour in bone tissue [118]. Flexoelectric fields are generated by fractures in the bone mineral and may be large enough to induce osteocyte apoptosis and initiate bone remodelling [119].

The same mechanical stimulus may cause a different response in osteocytes according to their cell body shape [120]. Recently, a study reports that osteocyte plasma membrane disruptions, caused by mechanical loading, act as triggering mechanosensing events, both in vitro and in vivo [121].

Osteocytes’ early response to mechanical loading results in vesicular ATP release by exocytosis, tuned according to the magnitude of the stimulus [122]. Mechanical stimulation of osteocytes also causes fluctuations in intracellular calcium levels; these are responsible for calcium-dependent actin contraction and release of extracellular vesicles containing bone regulatory proteins [123]. In fact, osteocytes respond to mechanical stimuli by producing various messenger molecules, such as nitric oxide and prostaglandins, namely prostaglandin E2 (PGE2) [117, 124, 125]. This response is dependent on the function of stretch-activated calcium channels [126], although reserves of intracellular calcium also contribute [123]. PGE2 has anabolic effects, stimulating osteoblast activity and new bone formation [127]. Nitric oxide inhibits bone resorption, by suppressing osteoclast formation and increasing the expression of osteoprotegerin [128, 129].

The lifespan of osteocytes is highly variable and likely associated with the rate of bone remodelling, depending on mechanical and environmental factors such as hormones; osteocytes apoptosis may be inhibited or induced by a variety of physiological and pathological conditions. Osteocyte apoptosis may be induced by biological effectors such as hormones, without being accompanied by increased osteoclastogenesis [130,131,132,133,134].

Young osteocytes are polarized towards the mineralization front, just like osteoblasts are, with the nucleus remaining close to vessels [104]. As lamellar bone matures, the osteocytes tend to spread their processes perpendicular to the longitudinal axis of trabeculae and long bones and appear as flattened cells. In immature bone, plump osteocytes with randomly distributed processes predominate [130]. Osteocyte density is closely related to bone architecture and thus to its mechanical behaviour [135].

Ageing has been correlated with smaller canaliculi, in lower numbers per lacuna, leading ultimately to reduced mechanosensitivity in the aged individual [136, 137].

2.2.2.3 Osteoclasts

Osteoclasts are multinucleated cells and belong to the same lineage as macrophages and monocytes (Fig. 6). Like macrophages, osteoclasts are able to merge and form multinucleated cells and to phagocytize [138]. The cell precursor may differentiate into either an osteoclast or a macrophage. The differentiation path depends on the progenitor cell being exposed to a receptor activator of several ligands (receptor activator of nuclear factor-κB ligand—RANKL, osteoprotegerin and osteoclast differentiation factor—ODF) or to colony-stimulating factors related to immune system [139,140,141].

Fig. 6
figure 6

A microphotograph of TRAP positive osteoclasts firmly attached to the bone surface. The ruffled border membrane is visible in direct contact with bone. This is the resorbing organelle; along its enlarged ruffled contact surface, proton pumps lower the local pH, dissolving hydroxyapatite

Full size image

The osteoclast presents distinctive functional features:

  • osteoclasts can attach firmly to the bone surface, isolating the area under the cell membrane from its surroundings; the membrane domain responsible for the isolation of the resorption site is called sealing zone [142, 143];

  • osteoclasts acidify the mineral matrix by the action of protons pumps at the ruffled border membrane, a resorbing organelle; the lowering of the pH causes the dissolution of the hydroxyapatite crystals [144,145,146];

  • osteoclasts are capable of synthesizing and secreting enzymes such as tartrate-resistant acid phosphatase (TRAP) and cathepsins in a directional manner; the proteases secreted by osteoclasts cleave the organic matrix; through the combined action of lysosome enzymes, matrix metalloproteinases and the pH reduction, bone is resorbed [147, 148];

  • osteoclasts can phagocytize the resultant organic debris and minerals, removing them from the resorption lacunae, through a transcytosis process [149, 150].

The bone resorption process begins with differentiation and recruitment of osteoclast precursors, which merge and originate matured multinucleated bone-resorbing osteoclasts. Bone resorption begins when the osteoclast attaches to the mineralized bone matrix through the interaction of integrins with matrix proteins, like osteopontin and bone sialoprotein, previously laid down by osteoblasts [143].

2.3 Regulation of Bone Metabolism (Modelling/Remodelling)

The bone cell populations are responsible for bone remodelling and repair. These processes are regulated systemically by hormones, neuropeptides and other mediators and locally by cytokines and growth factors [151, 152].

2.3.1 Parathormone (PTH), Vitamin D and Calcitonin

The bone mineral metabolism (calcium and phosphorus) is regulated by parathormone (PTH), calcitonin, FGF23 and vitamin D.

PTH is a peptide hormone produced by the parathyroid glands in response to low levels of extracellular ionized calcium, detected by specific cell-surface calcium-sensing receptors located in the parathyroid glandular tissue. High levels of PTH increase the number of osteoclasts and trigger resorption of bone matrix, with consequent release of calcium phosphate and increasing calcemia. This mechanism has developed as a protection against acute hypocalcemia. Inversely, low levels of PTH cause the elevation of osteoblast numbers. PTH also acts on osteoblasts’ receptors, stimulating proliferation and differentiation and inhibiting apoptosis [153]. PTH also regulates kidney function by impairing phosphate reabsorption and promoting its excretion, by stimulating calcium reabsorption and up-regulating a hydroxylase enzyme (CYP27B1), thus promoting 1,25(OH)2 vitamin D3 synthesis [154].

Circulating hormonal metabolite, 1α,25-dihydroxy vitamin D3 (1,25(OH)2D3), enhances several physiological functions, including intestinal calcium and phosphate absorption, bone phosphate and calcium resorption, and renal calcium and phosphate reabsorption, which results in a rise in the blood calcium and phosphate, required for bone passive mineralization of unmineralized bone matrix to occur [155, 156]. Additionally, 1,25(OH)2D3 stimulates differentiation of osteoblasts and the expression of several bone proteins, like bone-specific alkaline phosphatase, osteocalcin, osteonectin, osteoprotegerin and other cytokines, and influences the proliferation and apoptosis of other bone cells, including hypertrophic chondrocytes [157]. This may help explain why endogenous PTH levels can have anabolic and catabolic effects and are associated with differential skeletal effects on cortical and trabecular bones [158].

Calcitonin is produced by parafollicular cells of the thyroid, in direct response to extracellular calcium, through the same sensor that regulates the production of PTH. It inhibits matrix resorption, promotes calcium and phosphate excretion, thus reducing calcium and phosphate serum levels; calcitonin inhibits osteoclast mobility and the secretion of proteolytic enzymes [159, 160].

2.3.2 Growth Hormone (GH)

Growth hormone or somatotropin is secreted in pulses by the anterior pituitary gland, inducing bone longitudinal growth [161]. It also induces organs such as the liver and the skeleton to synthesize somatomedins that influence growth, such as insulin-like growth factor 1 (IGF-1) and 2 (IGF-2) [162]. The chondrocytes in the epiphyseal plate are stimulated not only by IGF1 and IGF2 but also directly by GH; proliferative and hypertrophic chondrocytes also secrete IGFs; IGF-1 acts inhibiting further GH secretion [163, 164].

According to Ohlsson et al. [162], GH action in bone remodelling follows a “biphasic model”: initially, it increases bone resorption, causing bone loss, followed by a phase of increased bone formation. When bone formation is more stimulated than bone resorption (transition point), the bone mass increases. A net increase of bone mass will be seen after 12–18 months of GH treatment in GH-deficient adults [165]. GH increases bone growth, by increasing both periosteal and endocortical bone formation, bone mineral content (BMC) and bone mineral density (BMD). GH acts synergistically with PTH to increase bone growth and bone formation, bone density and mass and to decrease bone resorption [166].

2.3.3 Insulin and Insulin-like Growth Factors (IGF-1 and IGF-2)

IGF-1 stimulates chondrocyte proliferation in the growth plate, thus playing a crucial role in longitudinal bone growth [167]. It is also involved in the formation of trabecular bone [168]. Insulin and IGF-1 have anabolic effects over the osteoblast and promote bone development, mainly through the activation of Akt and ERK signalling pathways; also, IGF-1 is capable of inducing osteoblasts in vivo proliferation whilst inhibiting the gene expression of osteocalcin, a marker for differentiating osteoblasts; insulin enhances osteocalcin expression but has no effect on osteoblast proliferation [169]. Additionally, insulin indirectly enhances Runx2 expression, a regulator of osteoblast differentiation [55, 169]. A study with insulin-deficient type I diabetic mice showed that these mice presented a decreased expression of Runx2 and the Runx2-regulated genes, like osteocalcin and collagen type I, and a secondary decrease in bone formation. Bone loss was restored after insulin treatment, which increased Runx2 expression and the expression of related genes [170].

Likewise, IGF-2 potentiates BMP-9-induced osteogenic differentiation and bone formation [171] through PI3K/AKT signalling. Moreover, a recent study in mice aortas showed that IGF-2 induces the expression of miR-30e, in a feedback loop. miR-30e is a major down-regulator of osteogenic differentiation of MSCs and smooth muscle cells [172].

2.3.4 Sex Steroids (Oestrogen and Testosterone)

Bone metabolism is strongly influenced by sex steroids. Oestrogen is an important regulator of skeletal development and homeostasis, both in men and women, exerting direct and indirect effects on the skeleton [173,174,175]. Indirectly, it influences, for example, the calcium intestinal absorption [176, 177] and secretion [178], and the calcium renal excretion; oestrogen also influences the secretion of PTH [179, 180]. Oestrogen maintains bone homeostasis by inhibiting osteoblast and osteocyte apoptosis [134, 181, 182], and it inhibits osteoclast formation and activity, inducing osteoclast apoptosis [183,184,185,186,187]. Oestrogen deficiency causes bone loss and osteoporosis [188].

Androgens are also important to bone homeostasis. However, their role is likely more important during growth and contributes, via the GH/IGF system, to bone formation at the periosteum [189]. Androgens contribute to the maintenance of cancellous bone mass and integrity, regardless of age or gender [190, 191]. Androgen-deprivation therapy has negative effects on bone mineral density; these effects can be partially delayed by exercise, in the lumbar vertebrae but not in the hip [192].

2.3.5 Thyroid Hormones

The skeleton is a target tissue for thyroid hormones, namely for thyroid hormone 3,5,3′-L-triiodothyronine (T3). Thyroid hormones influence bone growth during early development and adult bone turnover and maintenance. They act both directly, by stimulating bone resorption and formation, and indirectly, by enhancing the effects of growth hormone over tissues. Hypothyroidism causes impaired bone formation and growth delay; thyrotoxicosis is a recognized cause of secondary osteoporosis, and abnormal thyroid hormone signalling has been recognized as an osteoarthritis’ risk factor [193]. T3 stimulates osteoblast proliferation and differentiation, with bone matrix secretion, modification and mineralization. Thus, bone turnover is increased by thyroid hormones, which is confirmed by increased biochemical markers of bone turnover, such as osteocalcin and bone-specific alkaline phosphatase [194,195,196], and therefore, bone loss can occur [160, 197]. Thyroid-stimulating hormone (TSH), produced by the hypophysis, has direct effects on bone turnover [198], and TSH receptors have been found on osteoblasts and osteoclasts, although available data does not allow conclusions on whether TSH inhibits, increases or does not affect osteoblast differentiation and function [193]. Still, recombinant TSH showed antiresorptive effects in ovariectomized rats [198, 199] and lower TSH levels—with no apparent association with free T4 levels—have been related to hip fracture risk, supporting the idea that TSH effect on the skeleton may be independent of free T4, though its action on dedicated membrane receptors can be up-regulated by modulators [196, 200].

2.3.6 Leptin (“Satiety” Hormone)

Leptin is produced mainly in adipose tissue, and it is a regulator of food intake and energy expenditure through its effects on the central nervous system (CNS). Its influence in bone metabolism probably follows two pathways: a central pathway, activating the sympathetic nervous system that inhibits bone formation, and a peripheral pathway, promoting bone formation through leptin receptors on osteoblastic cells [201, 202]. Leptin inhibits osteoclast generation [203], promotes the decrease in cancellous bone and increases in cortical bone, thus enhancing bone enlargement [204, 205]; it also increases osteoblast number and activity, acting primarily through the peripheral pathways [206]. Another study showed that leptin increases bone mineral content and density, especially at the lumbar spine [207]. However, in the ovine foetus, leptin infusion caused increased femur porosity and connectivity density, and vertebral trabecular thickness whilst leptin receptor antagonist infusion decreased trabecular spacing and increased trabecular number, degree of anisotrophy and connectivity density in the lumbar vertebrae; effects differed in females and males [208]. Leptin also increases the expression of IGF-1 receptor and IGF-1 receptor messenger [209]. During infancy and childhood, leptin and IGF-1 were associated with body composition in preterm-born children. The same study also describes leptin association with bone parameters in early infancy, but not in childhood [210]. These results suggest leptin role on bone metabolism and architecture may vary with gender, age and interaction with other hormones and factors. Leptin is also a key up-regulator of FGF23 secretion [211], and it has been described as a direct enhancer of parathormone secretion [212].

2.3.7 Bone Morphogenetic Proteins (BMPs)

BMPs are a group of 15 growth factors also known by cytokines, which belong to the transforming growth factor β (TGF-β) superfamily, with the ability to induce the formation of bone [213] and cartilage [214]. BMPs play a major role in the regulation of osteoblast lineage-specific differentiation and later bone formation [215]. Alterations in BMPs activity are often associated with a great variety of clinical pathologies, like skeletal and extra-skeletal anomalies, autoimmune, cancer and cardiovascular diseases [216]. BMPs crosstalk with several other major signalling pathways, e.g. Wnt, Akt/mTOR, miRNA, among others, having Runx2 as a key integrator [216, 217]. Among all BMPs, BMP9 has stronger osteogenic inductive activity over MSCs [215, 218, 219]; BMP9 also acts synergistically with TGF-β and GH to enhance bone formation [216, 220, 221]. In addition to BMP9, other BMPs also have shown the ability to induce osteogenesis in vivo, such as BMP2, BMP6 and BMP7 [222,223,224], with recombinant human BMP2 and BMP7 already being commercialized with the purpose of enhancing bone healing [225]. However, recent studies indicate the existence of age-related differences in BMP2-mediated bone regeneration, including relative dose sensitivity [226]. Contrariwise, BMP3 is known to be a negative regulator of bone formation and BMP4 has been shown to decrease trabecular bone formation in a murine model [219, 227].

2.4 Bone Remodelling and Cell Interchange

Healthy bone, both cortical and trabecular, is continuously remodelling, a dynamic process with bone resorption and formation. Bone remodelling is modulated by mechanical loading, blood calcium levels and a wide range of paracrine and endocrine factors.

The bone remodelling process depends on the coordinate actions of osteoblasts, osteoclasts, osteocytes and osteoblast-derived bone-lining cells, along with other cells, such as macrophages and immune cells. The ensemble constitutes the “basic multicellular unit” (BMU) or “bone remodelling unit” (BRU). In the BMU, the amount of bone lysis achieved by osteoclasts is equal to the amount of bone produced by osteoblasts. The balance between osteoblastic and osteoclastic activity is known as coupling. Frost proposed that bone longitudinal growth, modelling and BMU-based remodelling activities were modulated by a “mechanostat”, a mechanism modulating bone mass in the function of mechanical use, in which BMUs would play a central role, along with bone longitudinal growth and modelling (bone formation). Bone modelling was thus considered as an adaptative response to overloading and remodelling as a response to underloading, with given strain set points for each process [228].

Osteoclasts and osteoblasts within the BMU may function under the control of other cell types, since osteoblasts and osteoclasts may perform their functions in the absence of each other [229, 230]. Cells from the osteoblast lineage express receptors for cytokines and other local secreted factors that stimulate osteoclast formation [231]. The BMU can be inhibited by old age, drugs, endocrine, metabolic or inflammatory diseases.

Regardless of the triggering stimulus, osteoclast formation depends on RANKL. Osteoblasts express membrane-bond RANKL, and this regulatory molecule interacts with a receptor (receptor activator of nuclear factor-κB—RANK), expressed on the surface of osteoclast precursors. The RANK activation by RANKL is essential for the fusion of the osteoclast precursor cells and osteoclast formation [232].

Both down-regulation and up-regulation of RANKL expression by osteoblasts under similar mechanical stimulation have been described [233, 234]. Osteoblasts subjected to different mechanical stimuli respond by an increase in RANKL-bound and a decrease in soluble RANKL secretion [235].

The RANKL/RANK coordinated effects can only be understood by adding osteoprotegerin to the axis. Cells from osteoblastic lineage produce osteoprotegerin (OPG). OPG is soluble and blocks the interaction between RANKL and RANK, acting as a decoy receptor for RANKL. OPG thus inhibits osteoclast formation and induces osteoclast apoptosis [236]. Osteoblasts, in addition, secrete macrophage colony-stimulating factor-1 (M-CSF-1); M-CSF-1 promotes osteoclast precursor proliferation and RANK expression [237, 238]. Osteoblast-like cells’ cultures mechanically stimulated may respond by a decrease in the production of OPG, without a change in the RANKL production, with a consequent increase in the ratio of RANKL/OPG. This could translate into increased bone remodelling. However, subjecting osteoclast-like cells to the same mechanical stimuli regimen, decreased TRAP and a period of stimulation of one minute at 0.3 Hz frequency, a decrease in cell fusion and resorption activity was observed [239].

RANKL expression by osteoblast lineage cells is enhanced when microdamage within the bone matrix occurs. Microdamage may occur under physiological bone loading and in pathological conditions. The presence of microcracks is sensed by osteocytes and may induce osteocyte apoptosis; osteocyte apoptosis may also be induced by disuse and is closely correlated with higher bone remodelling levels [131, 240,241,242,243,244,245].

Pulsating fluid flow (PFF)-treated osteocyte cultures conditioned the culture medium, inhibiting osteoclast formation and decreasing in vitro bone resorption. These effects have not been detected in the medium from PFF-treated fibroblast cultures [246]. In osteocytes subjected to PFF, nitric oxide is involved in the up- and down-regulation of at least two apoptosis-related genes (Bcl-2 and caspase-3, with antiapoptotic protective and pro-apoptotic functions, respectively) [247]. Nitric oxide (NO) is a second messenger molecule produced in response to mechanical stimulation of osteoblasts and osteocytes, and other cell types such as endothelial cells, with a large variety of biological functions [248,249,250,251].

Osteocytes, thus, regulate osteoclastogenesis and osteoclast activity through soluble factors and messenger molecules.

Other pathways are relevant for osteoblasts, osteocytes and osteoclasts interweaved regulation, such as the Notch signalling pathway. In osteocytes, the Notch receptor’s activation induces OPG and Wnt signalling, decreasing cancellous bone remodelling and inducing cortical bone formation [252]. Wnt/Lrp5 signalling in osteocytes has been considered as a key pathway for bone response to loading [253].

2.5 Bone Mechanotransduction

Bone mechanotransduction, essential in health and disease states, is not yet fully understood. The elements involved in transduction include the ECM, cell–cell adhesions, cell–ECM adhesions, cell membrane components, specialized surface processes, nuclear structures and cytoskeleton.

2.5.1 The Cell Membrane Elements, Cell–Cell and ECM–Cell Adhesions

Cell membrane-associated mechanotransduction mechanisms depend on the integrity of the phospholipid bilayer. Mechanotransduction pathways are disrupted if membrane cholesterol is depleted, inhibiting the response to hydrostatic and fluid shear stress [254, 255]. Cytoskeleton actin polymerization and assembly are influenced by membrane cholesterol levels [256, 257]. However, it has been proposed that actin polymerization during synaptic vesicle recycling is influenced by vesicular cholesterol, but not plasma membrane cholesterol, as suggested by a study wherein the inhibition of actin polymerization by the extraction of vesicular cholesterol resulted in the dispersal of synaptic vesicle proteins [258]. But even with a functional cell membrane, if integrin binding is impaired, actin cytoskeleton will not re-organize in response to shear stress [259]. However, nanometre- to micron-sized tears, reparable defects in the cell plasma membrane promote particle flux across the cell membrane, namely Ca2+ influx [121].

Integrins are cell adhesion receptors, heterodimers of non-covalently associated 18α and 8β subunits, in mammals, that can combine to generate 24 different receptors with different binding properties and different tissue distributions [260, 261]. These subunits possess an extracellular portion with several domains, able to bind to large multi-adhesive ECM molecules, which in turn bind to other ECM molecules, growth factors, cytokines and matrix-degrading proteases [260]. Integrins were first acknowledged as bridging the ECM and the cell cytoskeleton, including the actin cytoskeleton but also the intermediate filament network, essentially vimentin and laminin [262]. Cells use multiple mechanisms to sense and respond to mechanical stress applied to integrins [263]. Recruitment of vimentin has been shown to depend on integrin β3 subunits, underpinning the relationship between the various cytoskeletal elements and integrins [264]. The cytoplasmatic portions of integrin β subunit bind to talin, which can also directly bind to vinculin and actin filaments [265]. On the other hand, integrin α4 subunit binds to paxillin [266], a protein that integrates sites of cell adhesion to the ECM.

Integrins allow communication between structures in the interior and outside of the cell, in a bidirectional way. The inside-out signalling turns the integrin extracellular domains into the active conformation. In the outside-in pathway, when an integrin binds to the extracellular ligand, it clusters with other bound integrins, forming focal adhesions, highly organized intracellular complexes; these are connected to the cytoskeleton. The focal adhesions integrate a range of different molecules, including the cytoplasmic portions of the clustered integrins, proteins of the cytoskeleton and signalling molecules [265, 267]. Initial adhesions to substrates are characterized by punctuating areas at the limits of lamellipodia, usually known as focal complexes. Focal adhesions are the mature form of cell–matrix adhesion, with an elongated shape, and are associated with bundles of actin and myosin (stress fibres). There is a specialized form of focal contact, in which integrin binds to fibronectin fibrils and tensin but with low levels of tyrosine kinases [268, 269]. Most focal adhesions also contain several types of signalling molecules like tyrosine phosphatases and tyrosine kinases and adaptor proteins [267, 270,271,272].

Matrix proteins may also modulate cell adhesion; connective tissue growth factor (CTGF), which is a matrix protein, enhances osteoblast adhesion (via αVβ1 integrin) and cell proliferation, by inducing cytoskeletal reorganization and Rac1 activation [273]. Another matrix protein—osteoactivin—also modulates osteoblast adhesion, differentiation and function, stimulating alkaline phosphatase (ALP) activity, osteocalcin production, nodule formation, and matrix mineralization [274]. α5β1 integrin interacts with its high-affinity ligand CRRETAWAC, enhancing the Wnt/β-catenin signalling mechanism to promote osteoblast differentiation independently of cell adhesion [275].

Cell adhesion and mechanical stimulation depend on integrin mediation [276]. Forces applied to integrin receptors cause local adhesion proteins to be recruited, and the cell adapts by making the integrin–cytoskeleton linkages more rigid; myosin II contraction makes the cell apply tension to the substrate [277]. Different signalling pathways are triggered by sensed stress through integrin receptors. Sequential expression of integrin ligands (osteopontin, fibronectin and bone sialoprotein) in response to mechanical stimulation of osteoblasts has been described [278]. Bonds between integrin and ligands become stronger in the presence of cell tension [279].

Osteocytes are highly specialized in their interaction with ECM; osteocyte cell bodies express β1 integrins while cell processes express β3 integrins, the latter in a punctuate distribution [280,281,282,283]. Thi et al. [284] identified the cell processes as the mechanosensory organs in osteocytes. It has been demonstrated that integrin αVβ3 is essential for the maintenance of osteocyte cell processes and also for mechanosensation and mechanotransduction by osteocytes, by ATP release that triggers calcium signalling [285, 286]. β1 integrins have been shown to regulate specific aspects of mechanotransduction, namely the cortical osteocyte response to disuse [283]. In osteoblasts, a mechanical load applied to β1 integrin subunit results in calcium influx [287], independently from gap junctions [288]. Another study showed that ERK1/2 activation by strain prevented osteocyte apoptosis but required the integrin/cytoskeleton/Src/ERK signalling pathway activation [133].

Apart from integrin, other membrane proteins are responsible for the conduction of mechanical stimuli. Cadherins, which connect to the cytoskeleton, also mediate force-induced calcium influx [289, 290] and participate in the Wnt/β-catenin pathway [291]. In osteoblasts, it has been suggested that GPI-anchored proteins may play an important role in mechanosensing, by demonstrating that the overexpression of GPI-PLD, an enzyme that can specifically cleave GPI-anchored proteins from cell membranes, inhibits flow-induced intracellular calcium mobilization and ERK1/2 activation in MC3T3-E1 cells [255]. Ephrins (ligands) and Ephs (receptors) contribute to cell–cell interactions between osteoclasts and osteoblasts, helping to regulate bone resorption and formation, and appear to be necessary for hMSC differentiation [292, 293]. Lastly, another family of proteins—galectins—is also involved in regulating osteogenesis; for example, Gal-3, which is expressed both by osteocytes and osteoblasts, plays a significant role as a modulator of major signalling pathways, such as Wnt signalling, MAPK and PI3K/AKT pathways [294]; Gal-8 induces RANKL expression by osteoblasts and osteocytes, osteoclastogenesis and bone mass reduction in mice [295]; and Gal-9 induces osteoblast differentiation through the CD44/Smad signalling pathway in the absence of bone morphogenetic proteins (BMPs) [296].

Gap junctions are transmembrane channels that connect the cytoplasm of adjacent cells. Only small metabolites, ions and signalling molecules like calcium and cAMP pass through these channels since the molecular weight must be lower than 1 kDa [297, 298]. Gap junctions are essential for bone mechanosensation since in osteoblastic cells, the PGE2 production induced by fluid flow is dependent on intact gap junctions; if these are disturbed, PGE2 production does not occur [288, 299]. Mice lacking Cx43 gap junctions in osteoblasts and/or osteocytes exhibit increased osteocyte apoptosis, endocortical resorption and periosteal bone formation [300].

2.5.2 Primary Cilia

In different cell types, different structures ensure recognition of mechanical stimuli; kidney epithelial cells possess a single microvillar projection on their apical surface (primary cilia). A similar structure was described in osteoblasts and osteoblast-like cells [301, 302]. Primary cilia originate in the centrosome and project from the surface of bone cells; its deflection during flow indicates that they have the potential to sense fluid flow. These cilia deflect upon application of 0.03 Pa steady fluid flow and recoil after cessation of flow [303, 304]. In bone, primary cilia translate fluid flow into cellular responses, independently of Ca2+ flux and stretch-activated ion channels [303]. It has been demonstrated in vitro that, apart from mediating the up-regulation of specific osteogenic genes, primary cilia are also chief mediators of oscillatory fluid flow-induced extracellular calcium deposition, thereby playing an essential role in load-induced mineral matrix deposition [301]. A study using knockout mice of Kif3a, which results in defective primary cilia, showed that primary cilia are essential for the ability of pre-osteoblasts to sense strain-related mechanical stimuli at a healing bone–implant interface, inducing osteoblast further differentiation [305]; using the same animal model, another study shown primary cilia were paramount for MSCs to sense mechanical signals and enhance osteogenic lineage commitment in vivo [306]. Primary cilia must also be present in osteocytes for pulsed electromagnetic fields to inhibit osteocyte-mediated osteoclastogenesis and inhibit osteocyte apoptosis, modulate cytoskeletal distribution and decrease RANKL/OPG expression [95, 96].

Concerning osteocytes, there is still conflicting information regarding in vivo expression of cilia. Their role as mechanosensors depends on the type and number of cells with cilia and on the local mechanical environment. The incidence of primary cilia in osteocytes has been described as of 4%; this may indicate that cilia function as mechanosensors on a selected number of cells or that cilia function in concert with other mechanosensing mechanisms [307].

2.5.3 The Cytoskeleton

The cell cytoskeleton network is coupled to the ECM through specific transmembrane receptors. Integrins connect to the cytoskeleton through focal adhesions that gather actin-associated proteins such as talin, vinculin, paxillin and zyxin. Both paxillin and zyxin belong to a group of LIM domain structural proteins, which have been suggested as mechanoresponders responsible for regulating stress fibres assembly, repair and remodelling in response to changing forces [308]. Focused stresses applied to the surface of the cellular membrane are transferred across the network of cell adhesions, microfilaments and microtubules and affect distant cellular sites such as the mitochondria and nucleus or the cell membrane on the opposite side. The transmission of strain towards the ECM stimulates structural changes at a higher organization level, making it stronger [309, 310].

The cell deformation in consequence of applied stress does not correspond to the predicted behaviour of an isotropic viscoelastic material; the interior of the cell, the cytoskeleton, is anisotropic. The intricate network of microtubules and microfilaments, how it spreads and is connected to the point of applied force, may result in structures away from the load application point to be further displaced than closer ones; displacements towards the origin of the compressive stimulus are also possible. Behaving in an anisotropic way, cells can respond to an external force according to its magnitude and direction [311,312,313]. An intact cytoskeleton is necessary for the rendering of applied forces into mitochondria movements. Since mitochondria are semi-autonomous organelles, highly dynamic, the distress caused by mechanical stimuli exerts biological effects on their function [313], both in health and disease [314].

It is, therefore, logical that mechanical properties of the ECM affect the behaviour of cells from osteoblastic lineage, with mature focal adhesions and a more organized actin cytoskeleton associated with more rigid substrates, suggesting that controlling substrate compliance enables control over differentiation [315] and that this influence on differentiation is independent of protein tethering and substrate porosity [316].

Other factors are determinant for cell fate. A recent in vitro study showed similar patterns in cell growth, differentiation and gene expression in human osteoblasts and endothelial cells when implanted in two different ceramic scaffolds—β-tricalciumphosphate and calcium-deficient hydroxyapatite. These scaffolds had different chemical and physical characteristics, with results suggesting that the interaction between different cell types and scaffold materials is crucial for growth, differentiation and long-term outcomes of tissue-engineered constructs [317]. It has also been highlighted the importance of surface roughness of the biomaterials in osteogenic differentiation and the contribution of specific integrin subunits in mediating cell response to different materials [318]; additionally, the application of synthetic integrin-binding peptidomimetic ligands (αVβ3- or α5β1-selective) to a titanium graft enhanced cell adhesion, proliferation, differentiation and ALP expression in vitro osteoblast-like cells, resulting in a higher mineralization on the surfaces coated with the ligands [319].

The biochemical nature of the substrate, its rigidity and spatial organization is recognized by cells through signalling from molecular complexes that are integrin-based.

In most anchorage-dependent cells, cell spreading on ECM is required for cell progression and growth; increasing cytoskeletal tension results in cell flattening, a rise in actin bundling and bucking of microtubules. Spread cells can transfer most of the load to the ECM.

The cell shape is influenced by how the cytoskeleton organizes its elements and it is determinant for cell function. For example, osteocyte morphology and alignment differ in two types of bone, fibula and calvaria, probably due to different mechanical loading patterns, which influence the cytoskeletal structure and thus cell shape [320]. Also, osteocyte and lacunae morphology may vary in pathological bone conditions, and these morphological variations may be an adaptation to the differences in matrix properties and, thus, different bone strain levels under similar stimuli [321]. Osteocyte morphology is characterized by long dendritic-like processes, cell shape also assumed by osteoblast MC3T3 cells cultured in 3D; however, differences in cytoskeleton elements in the processes of these two cell types may indicate differences in function; microtubules are predominant on osteoblasts’ processes while actin ensures integrity of osteocytes’ cytoplasmatic projections [322]. Osteocyte sensitivity to mechanical load applied to the microparticles varies between those attached to the cell bodies and the ones attached to the cell processes: a much smaller displacement of the second ones is needed to cause an intracellular calcium influx that rapidly propagates to the cell body; if local stimulus is applied to the cell body, the reaction is slower and a higher displacement is needed to cause the calcium transient [323].

Osteoblasts, osteoid–osteocytes and mature osteocytes have different mechanical properties. The elastic modulus is higher in the cell periphery than in the perinuclear region; the elastic modulus in both regions decreases as bone cells mature. These differences in elastic modulus probably depend on the number of actin filaments, as it has been shown in other cell types. Furthermore, focal adhesion area is smaller in mature osteocytes, when comparing to osteoblasts. If peptides containing RGD sequence are added to culture medium, both the focal adhesion area and the elastic modulus of osteoblasts decreases whilst remaining unaffected in osteocytes [324].

2.6 Mechanotransduction Mechanisms

The multitude of cellular structures, messenger substances, environmental factors and levels of organization of the organs involved in the mechanotransduction mechanisms in distinct cells and tissues makes it extremely complex to understand, predict and replicate how responses are composed at cellular, organ and living organism levels.

2.6.1 Strain, Frequency and Loading Duration

Bone remodelling is influenced by strain magnitude, frequency and loading duration. Wolff developed mathematical equations for trabeculae orientation and thickness prediction according to load [325]. Later, Turner enunciated three essential rules critical for bone remodelling [326]:

  1. 1.

    Remodelling is determined by dynamic loading, not by static loading;

  2. 2.

    Short periods of loading quickly trigger a response; prolonging loading times any further diminishes the magnitude of bone cell response;

  3. 3.

    Bone cells have memory and accommodate to routine loading, diminishing the amplitude of the response triggered by a same repeated stimulus.

Increasing loading frequency increased strain-related bone deposition in vivo, whilst decreasing the threshold for osteogenesis and bone formation [327]. Human osteoblasts subjected to strains varying from 0.8 to 3.2% respond to higher strain with increased expression of osteocalcin, type I collagen and Cbfa1/Runx2, and to lower strain magnitudes with an increase of alkaline phosphatase activity [328].

Bone formation depends on strain magnitude [329], along with the number of loading cycles at low frequencies [330]. Frost theorized that a minimum effective strain level was necessary to trigger bone formation, above 3000 micro-strain [228]. Strain distribution is also paramount for skeletal adaptation. Unusual strain distribution will rapidly trigger an osteogenic response, as suggested by the extensive periosteal and endosteal bone proliferation described by Rubin and Lanyon [331] in a study conducted in poultry. Rest periods between loading cycles also intensify osteogenic response [332] and maximize cell response [333]. During active exercise, peak strains in long bones may be high, but strains as low as 0.15% are enough to ensure osteoblast recruitment in vivo [331]. Human bone marrow stem cells show variable early osteogenic differentiation and gene expression accordingly with load and frequency regimen of cyclic hydrostatic pressure; osteogenic differentiation on the long term occurred under mechanical stimulation, independently of load magnitude and frequency, within the tested physiological ranges [334].

The adaptation of cortical bone is correlated with frequency, although not linearly; the changes in geometry are more significant with higher frequency, with a plateau for frequencies past 10 Hz [335].

Other mechanisms apart from direct deformation of cells are involved in bone mechanical stimulation. Bone’s canalicular system is filled with fluid. Simulation of osteogenic load levels has produced higher shear stresses due to fluid displacement in the canaliculi. The fluid flows within the canalicular system, wherein the osteocytes extend their cell processes, reinforcing osteocyte processes as the main mechanosensing organ in mature bone cells [336]. Multiple canaliculi intersect at points (canalicular joints); these occur with a density similar to that of lacunae and represent areas of enlarged space, with consequences on fluid flow variables such as fluid mass and velocity [337]. Microstructural changes associated with osteoporosis reduce interstitial fluid flow around osteocytes in the lacunar–canalicular system of cortical bone, impairing mechanosensation [338].

As reviewed previously, shear stresses resulting from fluid flow cause calcium influx through mechanosensitive channels [339, 340]. Calcium influx occurs is osteoblasts in response to oscillatory fluid flow [288].

Fluid also carries electrically charged particles. The resulting fluid flow phenomena are common to other biological tissues but not limited to living structures. The fact that fluid flow changes interfacial chemistry has been recognized; the flow of fresh water along the surfaces disturbs the equilibrium of dissolved ions, changing the surface charge and the molecular orientation of the water at the interface [341]. Likewise, when bone is deformed, a thin sheet of fluid with particles with charge opposite to that of the matrix and bone cells is formed [342]; when a non-uniform mechanical load is applied to the bone structure, the ions in the fluid move away from the matrix. Therefore, the displacement of the electrically charged fluid creates an electrical field aligned with the fluid flow. This causes an electrical potential, and the phenomenon is known as strain-generated bone streaming potential and has been described in bone [342,343,344,345]. The density of matrix fixed charges influences the magnitude of the generated streaming potential [346], so the mechanosensory ability along bone may vary and, ultimately, influence dynamic stiffness.

2.6.2 Bone Piezoelectricity and Flexoeletricity

Fukada and Yasuda first described bone piezoelectrical properties in 1957. In dry bone samples submitted to a compressive load, an electrical potential was generated, an occurrence explained by the direct piezoelectric effect [347]. In connective tissues, such as bone, skin, tendon and dentine, the dipole moments are related to the collagen fibres, composed by strongly polar protein molecules aligned [76, 348, 349].

Recently, it has been suggested that hydroxyapatite flexoelectricity is the main source of bending-induced polarization in cortical bone [119].

The architecture of bone itself, with its aligned lamellae, contributes to the existence of potentials through the bone structure [348].

Bone piezoelectric constants, i.e. the polarization generated per unit of mechanical stress, change with moisture content, maturation state (immature bone has lower piezoelectric constants when comparing to mature bone) and architectural organization (altered areas, such as bone neoplasia osteosarcoma, show lower values) [350]. In dentin, piezoelectric constants are higher when moisture contents increase, also behaving anisotropically; tubule orientation determined piezoelectricity, stronger parallel to the tubules [351]. Wet bone also behaves as a piezoelectric material [347, 350, 352].

Bone piezoelectrical properties have risen interest, in the context of bone physiology and electromechanics. It has been related to bone remodelling mechanisms and to streaming potential mechanisms [353, 354]. Using a piezoelectric substrate and the piezoelectric converse effect were tested in vitro and in vivo with promising results, mechanically stimulating osteoblastic cells and bone, suggesting the potential for clinical application [355, 356]. The development of new synthetic scaffolds is a new emergent field for the bone tissue engineering industry. Hydroxyapatite/barium titanate [357] or polycaprolactone/barium titanate composites [358] with piezoelectric coefficients dependent on distribution and density of barium titanate particles aim to improve cell adhesion and differentiation. A wide range of biomaterials with piezoelectric properties, with potential application for bone regeneration, is available [359]. Mesenchymal cell differentiation in 3Dpiezoelectric scaffolds can be modulated by the voltage output (or streaming potential); lower voltage output scaffolds promoted chondrogenic differentiation while high voltage output promoted osteogenic differentiation [360]. Electromechanical stimulation also promoted improved differentiation when compared to mechanical load alone [360].

Due to the potential impact on therapeutic approaches to bone remodelling and healing, more and more research is being conducted on bioinspired approaches that consider piezoelectric bone properties.