Both bone remodeling and modeling involve the activity of osteoclasts and osteoblasts. Osteoclasts secrete enzymes (e.g., cathepsin K) and acid that dissolve the bone mineral and degrade the bone matrix. This releases collagen split products and growth factors (such as transforming growth factor-β) embedded in the bone matrix that stimulate osteoblast recruitment to the resorption site. Osteoclasts are cells from the macrophage/monocyte (myeloid) lineage, which secrete and are regulated by cytokines [23]. Osteoclasts are formed primarily by stimulation of hematopoietic precursors with RANKL in the presence of macrophage stimulating factor (M-CSF) (Fig. 13.2). RANKL, a member of the tumor necrosis factor receptor superfamily, is produced by osteoblasts, stromal cells, fibroblasts, and activated T cells [24, 25] and stimulates osteoclast differentiation, activation, and survival. A complex network of cytokines and immune receptors regulate osteoclast formation and activity either directly or indirectly via RANKL [26–31]. RANKL-deficient mice have hyperdense bones secondary to lack of osteoclasts [25]. RANKL has been implicated as a key factor in the pathogenesis of bone loss associated with increased resorption, including postmenopausal osteoporosis and rheumatoid arthritis. A monoclonal antibody to RANKL, denosumab, is used to treat postmenopausal osteoporosis and bone loss associated with rheumatoid arthritis [32, 33].

Osteoprotegerin (OPG) is a soluble decoy receptor for RANKL produced by osteoblasts and stromal cells [34]. OPG is a potent inhibitor of osteoclast development. OPG transgenic mice have hyperdense bones, a phenotype that can be replicated by systemic administration of OPG to normal mice. OPG-null mice on the other hand are profoundly osteopenic due to unopposed osteoclast activity [35, 36]. Besides OPG, another control switch in osteoclast development is interferon (IFN)-β, which is induced by RANKL binding to its receptor RANK on osteoclast precursors [37]. IFN-β interferes with the activity of c-fos, a transcription factor that is essential for osteoclast formation. Other factors, such as transforming growth factor-β (TGF-β) and Wnt, inhibit osteoclastogenesis by upregulating the production of OPG by osteoblasts [38]. Wnt also directly represses RANKL expression via the Wnt canonical pathway [39, 40], while Wnt5a, a typical non-canonical Wnt ligand, enhances the expression of RANK in osteoclast precursors [41]. In addition, several cytokines relevant to the pathogenesis of IBD inhibit osteoclast differentiation, including interferon-γ (IFN-γ) [42], IL-10 [43, 44], IL-12 [45, 46], and IL-17 [47, 48]. Moreover, osteoclast differentiation involves transcription factors such as NFκВ, AP-1, and NFATc1, as well as co-stimulation via immunoglobulin-like receptors and activation of the phosphatase calcineurin, which can be regulated by inflammation [47, 49–52]. Therefore, osteoclast formation is subject to multiple regulatory controls by cytokines, transcription factors, and enzymes that play key roles in IBD. In addition, osteoclasts interact closely with the hematopoietic stem cells niche [53, 54]. These pathways are an example of the close physiological ties between the immune system and bone cells. However, it is not yet known whether these mechanisms are engaged in regulating bone mass in children with IBD.

The RANKL/OPG system also plays important roles outside of bone. This is evidenced by the lack of peripheral lymph nodes and impaired development of lactating mammary glands in RANKL or RANK-null mice [55]. In addition, RANKL/OPG may be involved in the formation of calcified atherosclerotic plaques, and serum OPG is emerging as a marker of cardiovascular mortality [56, 57]. The balance between RANKL and OPG may affect the severity of bone metastases of several cancers [58]. RANKL contributes to normal dendritic cell function and survival and the early development of B and T cells [25, 59, 60]. In addition, RANKL/RANK may play a role in intestinal mucosal tolerance [60]. OPG also has a role in the regulation of the immune response. Both B cells and dendritic cells secrete OPG, and this secretion is regulated by the CD40 receptor [61]. Also, dendritic cells isolated from OPG^−/− mice more efficiently present antigen in vitro and secrete more inflammatory cytokines when stimulated with bacterial products or soluble RANKL in vitro [62]. Collectively, this evidence suggests that RANKL/RANK/OPG plays an important role in the regulation of the immune response and in pathways involving the mobilization of calcium [63].

A role for the RANKL/OPG system is emerging in IBD. Circulating OPG levels are elevated in patients with IBD [64, 65], and expression of OPG and RANKL is increased in colonic macrophages, dendritic cells, and epithelial cells [66, 67]. High fecal OPG (which probably comes from the inflamed colonic epithelium) predicts resistance to corticosteroids and to infliximab in patients with IBD [68, 69]. In addition, fecal OPG decreases in children with IBD in remission [65]. Currently it is not clear whether circulating OPG in patients with IBD represents spillover from intestinal inflammatory activity or it comes from bone or other tissues (e.g., the endothelium or the liver) or from lipopolysaccharide that permeates into the systemic circulation from the inflamed gut. The function of RANKL/OPG in the pathogenesis of intestinal inflammation deserves further study.

Osteoblasts are cells from mesenchymal origin that lay down bone matrix that is rich in type 1 collagen. Several factors and hormones regulate osteoblast formation, both systemic and in the bone microenvironment [70]. Insulin-like growth factor-1 (IGF-1) is secreted by the liver in response to stimulation by growth hormone and enhances the expression of the mature osteoblast phenotype [71]. Serum IGF-1 is frequently reduced in children with active IBD due to growth hormone insensitivity in the liver and malnutrition [72]. Consequently, relative IGF-1 deficiency in children with IBD may negatively affect osteoblast differentiation and function. Tumor necrosis factor-α (TNF-α), an important cytokine in the pathogenesis of IBD, inhibits osteoblast development by inducing the degradation of Runx2 [73], a critical transcription factor in osteoblast development [73] and suppression of osteogenic factor signaling including Wnt [29] and bone morphogenetic protein-2 [74, 75]. TNF-α also regulates a number of inflammatory chemokines and cytokines, inflammatory genes, transcriptional regulators, bone-remodeling genes, signal transducers, cytoskeletal genes, and genes involved in apoptosis in pre-osteoblasts [76]. TNF-α and colitis decrease the expression of Phex in osteoblasts, which affects their mineralization function [77, 78]. TNF-α induces cAMP response element-binding protein H (CREBH), which blocks the anabolic effects of bone morphogenetic protein-2 on osteoblast precursors by inducing the Smurf1-mediated degradation of Smad1 [79]. In children, TNF-α blockade leads to a brisk increase in biomarkers of bone formation and significant linear growth, suggesting an activation of bone modeling [80, 81]. However, the effects of infliximab may be a product of improved disease control and not specific effects of this drug on bone metabolism in these patients. Besides their bone-forming activity, osteoblasts also interact with hematopoietic cells in the bone marrow [82].

T cells are emerging as important regulators of bone cell function [83]. Activated T cells can regulate osteoclast formation and activity by several mechanisms, both RANKL dependent and independent. Activated T cells secrete RANKL and consequently can promote osteoclast differentiation and survival. Both soluble and membrane-bound RANKL are produced by activated CD4⁺ and CD8⁺ T cells [24]. T cell-induced bone resorption has been implicated in tissue injury in animal models of arthritis and periodontal disease [84]. CD4⁺ Th17 T cells may be the most pro-resorptive T cell in the bone marrow [47] probably due to their ability to secrete cytokines that stimulate osteoclast formation and activity [85], upregulation of RANK in osteoclast precursors [86], and increased expression of RANKL in osteocytes [87]. This is significant given the importance of Th17 cells in the pathogenesis of IBD [88]. T cells may also play an important role in bone loss associated with estrogen deficiency, where osteoclast activity is upregulated. This is suggested by experiments performed in ovariectomized mice, where absence of T cells prevents bone loss [89]. In this model, the expansion of a TNF-α-producing T-cell pool appears to be essential and may occur as a result of upregulation of antigen presentation. The nature of the activating antigen(s) is not yet known, but it is possible that both self and foreign epitopes (including intestinal bacterial products) may play a role [90]. The concept that T cells activated by bacterial antigens may regulate bone cell function is intriguing in the setting of IBD, due to the defects in microbial recognition and processing that have been identified in this condition [91]. In IBD, it is possible that activated T cells may serve as “inflammatory shuttles” between the intestine and bone, since circulating T cells produce cytokines that can regulate both osteoblasts and osteoclasts. Ciucci et al. have shown that bone marrow IL-17/TNF-α-producing CD4⁺ T cells from IL-10^−/− mice with colitis (but not from IL-10^−/− without colitis or wild-type mice) induce osteoclast formation in vitro without addition of RANKL/M-CSF. These cells express membrane-bound RANKL and secrete M-CSF [16]. In addition, it is possible that circulating antigens may trigger immune responses via T-cell memory cells in the bone marrow that affect bone cell function. The activation state of T cells may also be important in their interaction with osteoclasts, since resting T cells inhibit osteoclastogenesis [92]. T regulatory cells (Treg) are present in the bone marrow and are potent inhibitors of bone resorption [93] probably due to their secretion of IL-4, IL-10, and TGF-β. In addition, T cells may also regulate bone formation by osteoblasts. For example, bone marrow CD8⁺ T cells stimulated by intermittent parathyroid hormone administration activate anabolic canonical Wnt signaling in pre-osteoblasts by CD8⁺ T cells [94]. Moreover, bone cells can influence T-cell differentiation and activity. Osteoclasts affect the differentiation and activity of γδ T cells from peripheral blood in vitro via soluble factors and cell-to-cell contact [95]. Osteoclasts can function as antigen-presenting cells and direct the formation of effector CD4⁺ and CD8⁺ cells [96] and induce FoxP3 expression in CD8⁺ cells [97]. Osteoclasts can also induce the formation of anti-resorptive CD8⁺ Treg [98], in a process that involves permissive levels of RANKL [99]. Examining these complex mechanisms in the context of IBD awaits additional research.

Genome-wide association studies have identified a number of unsuspected pathogenic pathways in IBD. Among them are endoplasmic reticulum (ER) stress, the unfolded protein response (UPR), and autophagy [100]. These pathways regulate the function of highly secretory cells such as Paneth cells and goblet cells in the intestinal lining and innate immune cells in the intestinal lamina propria [101]. When there is an overabundance of unfolded and misfolded proteins in the ER, the ER becomes stressed. The UPR is triggered, involving the activation of inositol-requiring kinase 1 α (IRE1α), pancreatic ER eIF2α kinase (PERK), and activating transcription factor 6 α (ATF6α) [102]. Each pathway leads to separate transcriptional events. The UPR aims to restore homeostasis to the ER by decreasing transcription and protein synthesis, degradation of proteins inside the ER, and shuttling of proteins away from the ER with chaperones. When ER stress is chronic and homeostasis cannot be achieved by the UPR, the cell goes into apoptosis. Osteoblasts secrete large amounts of collagen (osteoid, the bone matrix) and other factors and might be affected by defects in ER stress and the UPR found in IBD [103]. Bone morphogenetic protein-2 (BMP-2, a stimulator of osteoblast development and activity) induces the expression of ER stress transducers, such as old astrocyte specifically induced substance [104] and ATF [105]. The inositol-requiring protein 1α (IRE1α) and its target transcription factor X-box binding protein 1 (XBP1) are essential for BMP-2 -induced osteoblast differentiation [106]. The BMP-2 signaling pathway also activates the UPR during osteogenesis [105, 106], which induces the synthesis of RANKL and osteoclastogenesis [107]. To date, it is not known whether defects in the UPR that occur in IBD affect osteoblast function. Mature osteoclasts actively secrete acid and proteolytic enzymes such as cathepsin K to degrade the bone matrix and are also sensitive to ER stress. The IRE1α/XBP1-mediated branch is important in osteoclast development [108] and is involved in parathyroid hormone-induced osteoclast formation [107]. Therefore, defects in the UPR and ER stress present in IBD may affect the development and activity of both osteoblasts and osteoclasts.

Autophagy is a process by which cells recycle old proteins, damaged organelles, and other cellular debris. These elements are encircled by double-membrane vesicles called the autophagosomes, which fuse with lysosomes to become autolysosomes. Their content is recycled and returned back to the cell. Autophagy also plays a role in bacterial digestion after phagocytosis. The mammalian target of rapamycin (mTOR) is an important regulator of autophagy. In addition to controlling cell growth and metabolism, mTOR negatively regulates autophagy when nutrients and growth factors are abundant [109]. In IBD autophagy can be deficient, leading to persistence of bacteria inside of cells. It is possible that defects in autophagy in IBD may affect bone cell function. For example, induction of autophagy in osteoclasts decreases bone resorption [110]. On the other hand, autophagy induces osteoclast formation during hypoxia [111] and microgravity [112]. Autophagy is important for osteoblast differentiation [113, 114] and bone mineralization [115]. Therefore, it is possible that altered autophagy in IBD impairs normal osteoid mineralization by osteoblasts. Moreover, GWAS suggests that genes involved in autophagy regulate bone mineral density in humans [116]. In summary, pathogenic pathways involved in IBD may establish novel osteoimmune connections, an area that deserves additional study.

Cytokines produced by the inflamed intestine can regulate bone cell activity. IL-6, IL-17, and TNF-α induce osteoclast formation in vitro [16, 117, 118]. However, indirect effects of cytokines mediated through osteoblasts may affect their ultimate influence on osteoclasts. For example, IL-17 stimulates osteoblasts to secrete GM-CSF in the presence of vitamin D, resulting in inhibition of osteoclast formation in vitro [119]. IL-17 can also induce mesenchymal stromal cells and osteoblasts to secrete RANKL, which would stimulate osteoclastogenesis and bone resorption [47, 120]. In a mouse model of colitis, Th17 cells in the bone marrow that produce TNF-α and RANKL increase osteoclast formation; this effect can be blocked by an anti-IL-17 antibody, suggesting that IL-17 is an important pathogenic factor that reduces bone mass in this model [16]. Oncostatin M (OSM), a cytokine of the IL-6 family, is a major coupling factor produced by activated circulating CD14+ or bone marrow CD11b+ monocytes/macrophages upon activation of toll-like receptors (TLRs) by lipopolysaccharide or endogenous ligands that induce osteoblast differentiation and matrix mineralization from human mesenchymal stem cells [121].

Innate immune responses can be activated by toll-like receptors (TLRs). The mechanism of pathogen-induced bone disease includes activation of TLRs in immune cells by pathogen-derived molecules [122]. This activation results in synthesis and release of inflammatory cytokines that are capable of stimulating osteoclastic bone resorption, thus causing bone loss. Osteoclasts express functional TLRs. TLR ligands (CpG-ODN, LPS, Poly(I:C)) exert dual effect on osteoclast precursors. They inhibit the activity of the physiological osteoclast differentiation factor, RANKL, in early precursors, but strongly increase osteoclastogenesis in RANKL-pretreated osteoclast precursors [123–126].

The gut microbiome probably plays an important role in the pathogenesis of IBD. The study of the effects of the intestinal microbiota on bone development is in its early stages however. A report by Sjogren et al. suggests that gut bacteria are essential for normal postnatal bone remodeling [127]. Britton et al. showed that a strain of Lactobacillus reuteri can reverse osteoporosis caused by ovariectomy in mice [128]. The same group has treated bone loss associated by experimental type 1 diabetes in mice [129]. Collectively this evidence offers proof of principle that enteric organisms have the potential to regulate bone cell activity. More research is needed in this important area that is very relevant to human IBD.

IBD is a complex clinical entity, where multiple disease and treatment factors contribute to affect bone cell biology and ultimately skeletal health. In an effort to study mechanistic questions, animal models of intestinal inflammation have been used by several groups. A brief description of their observations follows.

Studies in both rat and mouse models suggest that intestinal inflammation can decrease bone mass by impairing bone formation. Lin et al. induced colitis in rats by rectal instillation of TNBS [130] to study its effects on bone mass, assessed by quantitative histomorphometry. After 3 weeks, rats with colitis had a 33% loss of trabecular bone loss in the tibia compared with age-matched, pair-fed control animals. This was associated with a marked suppression of the trabecular bone formation rate. As the colitis healed, bone formation became more active and bone mass normalized after 12 weeks. In IL-10^−/− mice with colitis, Dresner-Pollak et al. performed bone densitometry, ash weight, histomorphometry analysis, and mechanical fragility testing [131]. They observed that bone mass decreased secondary to decreased bone formation in 8- and 12-week-old mice; bone resorption was not increased in mice with colitis compared to wild-type controls. Long bones were more fragile in IL-10^−/− with colitis, and ash weight was reduced. However, since these studies did not include IL-10^−/− mice without colitis, it was not clear if at least some of the observations in the skeleton of IL-10^−/− mice were due to the IL-10 deficiency. More recently, Ciucci et al. addressed this gap and reported significant decreases in trabecular thickness, trabecular number, trabecular bone surface density, and trabecular bone volume per tissue volume in IL-10^−/− mice with colitis, but not in IL-10^−/− mice without colitis [16], suggesting that in this model bone effects are due to colitis and not IL-10 deficiency. IL-10^−/− mice with colitis harbor in their bone marrow IL-17/TNF-α-producing CD4⁺ T cells that attract osteoclast precursors. In addition, bone marrow mesenchymal stromal cells produce chemokines that may attract additional osteoclast precursors in this model [16]. Harris et al. have demonstrated that the inhibition of bone formation and bone modeling is reversible with healing of colitis in mice [132].

Three reports using adoptive transfer models of colitis suggest that bone mass decreases secondary to increased bone resorption. In the first paper, Ashcroft et al. studied IL-2^−/− mice with colitis at 4, 7, and 9 weeks of age and compared X-ray and histomorphometry with IL-2^−/+ and wild-type mice. IL-2^−/− mice develop colitis and also have splenomegaly, anemia, and other signs of systemic inflammation [18]. They observed a decrease in trabecular bone volume in IL-2^−/− with colitis compared with the other two groups of mice at 7 and 9 weeks of age. C57BL/6-Rag1^−/− mice transplanted with CD3⁺ cells from IL-2^−/− had significantly lower femoral BMD and % trabecular volume 6–8 weeks post-grafting. Serum OPG and osteoclast number were significantly higher in mice engrafted with T cells from IL-2^−/− mice compared to IL-2^+/+. In this model, treatment with OPG was associated with both improved bone mass and decreased intestinal inflammation. These results point to a possible role of T cells in bone loss in the context of intestinal inflammation and suggest a possible anti-inflammatory role for OPG. In the second study, Byrne et al. transferred CD4⁺CD45RB^Hi or CD4⁺CD45RB^Lo from CB6F1 mice to C.B.17 scid/scid mice [133]. CD4⁺CD45RB^Hi, but not CD4⁺CD45RB^Lo, caused colitis in recipient mice, and mice with colitis had lower bone mineral density in the femur/tibia. To treat bone loss, mice received Fc-OPG 3.4–5 mg/kg SC three times weekly for 34 days. OPG had no effect on the severity of colitis but significantly improved BMD. In the third study, Ciucci et al. reported CD4⁺T cells in the bone marrow of mice with colitis that produce IL-17 and TNF-α, capable of stimulating osteoclastogenesis in vitro [16].

Collectively, these observations suggest that intestinal inflammation can directly affect bone mass in rodents. Mechanisms may include decreased bone formation or increased bone resorption, depending on the model. It appears that intestinal inflammation present at an early age is associated with decreased bone formation. Administration of exogenous OPG increases bone mass in older mice with certain forms of colitis. However, this may be a nonspecific effect of OPG on normally active osteoclasts and by itself does not establish that increased bone resorption is responsible for bone loss in rodent models of colitis. In addition, in adoptive transfer models of colitis, it is not possible to distinguish whether intestinal inflammation in the recipient animals caused bone loss or if colitogenic T cells directly migrated into the bone marrow, influencing bone cell function. However, it is interesting that in the CD4⁺CD45RB^Hi model there is an inflammatory infiltrate in the bone marrow containing TNF-α-producing cells [133]. This provides proof of principle that intestinal inflammation is associated with the presence of activated T cells in the bone marrow that secrete pro-inflammatory cytokines which may influence the function of bone cells. The presence of these cells awaits studies in other animal models of intestinal inflammation and in humans.

Several studies have measured bone mineral density in children with IBD, both in incident and in prevalent cohorts (Table 13.1). The studies, which have been either longitudinal or cross-sectional and have used primarily DXA or pQCT to image bone, suggest that decreased bone mineral density is common in children with Crohn disease at the time of diagnosis, especially in patients with delays in growth and sexual maturation, active disease, and those with decreased lean tissue mass [4, 142, 148, 149]. Studies performed in incident cohorts of treatment-naïve patients suggest that disease factors can affect bone mass in children with IBD prior to the initiation of treatment. Collectively, this work suggests that children with Crohn disease are at greater risk for decreased bone mass than children with ulcerative colitis, probably because Crohn disease is more likely to affect linear growth. Patients with low body mass index, low serum albumin, and active severe IBD appear to be at particular risk for decreased BMD. The role of corticosteroids on BMD in pediatric Crohn disease, however, is not clear. The attainment of peak bone mass in Crohn disease is at risk, which may affect fracture risk later in life [150].

According to recent guidelines by the International Society for Clinical Densitometry, children with IBD should have DXA scanning if in the clinician’s judgment the measurement may influence the patients’ management [22]. In addition to measuring bone mass, body composition data provided by DXA may be helpful in guiding the nutritional rehabilitation of these patients. It is important to DXA BMD measurements to patient size, gender, and sexual maturation, because in any given patient with IBD, the challenge is to distinguish between small, normally mineralized bones and abnormally thin and weak bones [151]. Taken together, these studies indicate that the observed reduction of BMD in children with IBD can be attributed in part to decreased bone size due to growth delay. However, it is important to note that smaller bones may be weaker, and their physical properties may not be normal. It is not yet known whether smaller bone size leads to increased fracture risk in children with IBD. Conversely, increases in height track with significant improvements in BMD, especially in trabecular bone (Table 13.1).