Statistics from Altmetric.com
Osteoclast differentiation and activation
Differentiation, activation, and survival of osteoclasts are regulated by the balance between RANKL (RANK ligand) and osteoprotegerin (OPG). OPG is also known as osteoclast inhibitory factor (OCIT) and is a tumour necrosis factor (TNF) receptor family member, which is found as a “decoy” receptor only because of the lack of a membrane spanning domain.1 2 RANKL is a TNF family member and is also known as osteoprotegerin ligand (OPG-L), or osteoclast differentiation factor (ODF), or TNF related, activation-induced cytokine (TRANCE).3-5 RANKL exhibits its activity on cells through RANK (receptor activator of NF-κB), another membrane bound member of the TNF receptor family.6-11 In bone marrow, RANKL is produced by osteoblasts and both osteoblastic and fibroblastic stromal cells. The production of RANKL is induced by most (all?) known inducers of bone resorption and hypercalcaemia.12-16 Precursor cells in bone marrow exposed to M-CSF (monocyte/macrophage colony stimulating factor or CSF-1) and RANKL will differentiate within one week to mature osteoclasts.3 Both M-CSF and RANKL arenecessary and sufficient for this process,17-19 whereas RANKL but not M-CSF will activate mature osteoclasts to resorb bone.20 OPG added to such bone marrow cultures will inhibit the process of osteoclast generation reversibly.17 18 21 This can be shown by either removing the OPG after a few days or by adding more RANKL to the cultures, both of which will result in renewed osteoclast differentiation. Mature osteoclasts are characterised by markers such as tartrate resistant alkaline phosphatase (TRAP), cathepsin K, calcitonin, and vitronectin receptors (for review see Chambers22). A mature osteoclast forms a ruffled border, an actin ring, and a sealed or clear zone and will form a resorption pit on dentin slices.12 20 23 24However, there are differences between in vitro generated osteoclasts and harvested osteoclasts ex vivo. One such difference is the ability of interleukin 1 (IL1) to induce an actin ring in in vitro generated but not in ex vivo osteoclasts.25
Survival of mature osteoclasts is dependent on the presence of RANKL,16 though other factors such as IL1 can promote osteoclast survival, too.25 If osteoclast survival is not sufficiently supported osteoclasts will undergo apoptosis.17 26 In fact, apoptotic osteoclasts in low numbers can be detected after one week in bone marrow cultures driven by M-CSF and RANKL.17 However, if survival factors are withdrawn, all osteoclasts will undergo apoptosis quickly.17 These effects can also be demonstrated in vivo.17 26 For example, in mice treated with 10 mg/kg OPG intravenously all osteoclasts disappear within 48 hours.17Interestingly, within 7–10 days osteoclasts return and can be found in normal numbers and typical locations in these mice.17
Lessons from transgenic and knockout mice
Osteoclasts and osteoblasts drive bone remodelling, a continuous process of osteoclasts resorbing bone and osteoblasts laying down new bone. Bone resorption is necessary for skeletal growth as well as tooth eruption.27
Excess RANKL activity as in OPG−/− mice results in spontaneous fractures and vertebral deformities.28 Studies using knockout and transgenic mice have shown that skeletal growth is normal in OPG knockout mice (which have high net RANKL levels and show marked osteoporosis)28 29 and OPG transgenic mice (which have low net RANKL levels and show osteopetrosis).1 30 In contrast, in RANKL−/− and RANK−/− mice skeletal growth is impaired.9 31 This illustrates that biologically there is quite a difference between low RANKL levels or no RANKL at all. Without RANKL (or RANK) tooth eruption cannot occur (RANKL−/− or RANK−/− mice), whereas with a little RANKL tooth eruption is normal (OPG transgenic mice).
Although cells of the macrophage lineage also express RANK, it is interesting to note that macrophage differentiation is normal in OPG transgenic and RANKL−/− mice and thus seems to be independent of RANKL.1 31
RANKL treatment of mice results in induction, differentiation, and activation of osteoclasts which can be demonstrated and quantified histologically17 but which show functional sequelae, such as hypercalcaemia.3 12Known inducers of bone resorption and hypercalcaemia, such as IL1, TNFα, parathyroid hormone (PTH), PTH related peptide (PTHrp), vitamin D3 (1,25(OH)2-D3), etc, have been tested alone and jointly with OPG. In all cases OPG prevented the induction of hypercalcaemia, indicating that all those inducers act indirectly through the induction of RANKL.15 This conclusion is supported by the fact that injection of RANKL, IL1β, TNFα, PTHrp, or 1,25(OH)2-D3 into RANK−/− mice fails to induce hypercalcaemia.9 These studies and others not reviewed here led to the view, depicted in fig 1, of RANKL as a unique cytokine in osteoclast differentiation, activation, and survival.
Many tumour types metastasise into bone. These effects can be mediated through PTHrp or IL6, other inducers of bone resorption. Subjects with a tumour often show hypercalcaemia. In mouse tumour models, such as the C26 (originally derived from an adenocarcinoma) model, it has been shown that OPG treatment dose dependently prevents tumour-induced hypercalcaemia.32 Furthermore, OPG can also reverse tumour-induced hypercalcaemia in these models.32Interestingly, OPG treatment can prevent osteolysis and tumour metastasis into bone.33 The data demonstrate that tumour cell derived inducers of bone resorption, such as PTHrp, IL6, or TNFα, act indirectly through increases in RANKL.
Activated T cells produce RANKL
Other cells besides osteoblasts and osteoblastic stromal cells can also produce RANKL. Interestingly, RANKL is produced during, and is necessary for, mammary gland development (for details see Fataet al 34). However, in the following the discussion will be restricted to T cells as producers of RANKL.35 During T cell activation a first signal occurs through the interaction of peptide in the groove of the major histocompatibility complex class II on the surface of antigen presenting cells and the T cell receptor, and a second signal or costimulatory signal occurs through the interaction of B7 on antigen presenting cells and CD28 on T cells. Activation of T cells leads to induction of CTLA-4, a member of the CD28 family. In contrast with the activatory properties of CD28, CTLA-4 is a down regulatory element. CTLA-4 binds 50-fold better to B7 on antigen presenting cells, resulting in down regulation of T cell activity. Samson recently reviewed and discussed the relative roles of CD28 and CTLA-4.36 Not unexpectedly, in CTLA-4−/− mice a hyperactivated T cell compartment is seen.37 38 The activated T cells in CTLA-4−/− mice produce RANKL and, consequently, CTLA-4−/− exhibit severe osteoporosis.35OPG treatment of CTLA-4−/− mice results in increased bone mineral density and a significantly reduced number of osteoclasts.35
RANKL is induced in T cells upon activation through the T cell receptor. Several studies have considered the function of RANKL/RANK in the interaction of T cells and dendritic cells.5 6Supernatants of activated T cell cultures added in place of RANKL in the above described bone marrow culture system allowed for the differentiation of osteoclasts only if M-CSF was also present.35 39 This effect was inhibited by addition of OPG to the bone marrow cultures. Furthermore, formalin fixed, activated T cells could substitute for RANKL, which indicates that RANKL is active both in soluble and in the cell surface bound form. Both activated CD4 and CD8 T cells can support osteoclast differentiation.40 As many other cytokines (IL1, TNFα, IL6, interferon γ, IL3, etc) which induce RANKL in osteoblasts and osteoblastic stromal cells are generated in such T cell cultures, formalin fixed, activated T cells from RANKL−/− mice were added to bone marrow cells and found not to induce osteoclasts.35This indicates that it is the surface bound RANKL (and not other factors) on T cells which drives osteoclast differentiation in this cell culture system.35
OPG protects bone in arthritis
The above findings prompted us to look into T cell mediated autoimmune diseases and bone effects. Mycobacteria-induced adjuvant arthritis in Lewis rats was chosen as a model. Severe polyarthritis with extensive joint destruction is induced in these rats by an injection of a suspension of heat killed mycobacteria in oil at the base of the tail.41 Nine days later clinical symptoms of disease, paw swelling, and weight loss begin, and disease severity is monitored by their daily measurement. Bone and cartilage destruction in this model is very rapid. By day 16 significant loss of bone has occurred as shown by erosions and by significant loss of bone mineral density in the remaining bone.41 We routinely measure bone mineral density in the calcaneus on day 16 by dual enhancedx ray absorptiometry(DEXAscan).41 Treatment of arthritic rats with OPG daily subcutaneously from day 9 to 16 resulted in a dose dependent inhibition of the loss of bone mineral density and bone erosive processes.35 In fact, osteoclast numbers, which were high in the arthritis group (80 ± 20) osteoclasts/mm2), were dose dependently reduced to levels found in normal rats (9 ± 2) osteoclasts/mm2). Interestingly, inflammatory parameters, paw swelling and body weight loss, were not affected by OPG treatment at all. OPG inhibits these bone erosive processes leaving less room for infiltrating inflammatory cells (fig 2). The results indicate that at least in this animal model of arthritis all effects which lead to bone erosive processes go through (the induction of) RANKL and, furthermore, the effects of IL1, TNFα, and other mediators of bone resorption seem to be indirect through the induction of RANKL.35
(All) known stimulators of bone resorption seem to work indirectly through the induction of RANKL.
Activated T cells produce RANKL, which is active both when cell surface bound and when soluble.
The balance of RANKL and OPG determines osteoclast activity (bone resorption).
OPG can inhibit bone metastasis of tumours by inhibiting bone resorption.
OPG prevents bone loss in adjuvant arthritis without effect on inflammation.
OPG may provide a pharmacological tool for osteoporotic and erosive bone disorders.42
I thank Brad Bolon and Giuseppe Campagnuolo for their contributions and stimulating discussions.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.