Non-Canonical (RANKL-Independent) Pathways of Osteoclast Differentiation and Their Role in Musculoskeletal Diseases

Osteoclasts are multinucleated cells derived from mononuclear phagocyte precursors (monocytes, macrophages); in the canonical pathway of osteoclastogenesis, these cells fuse and differentiate to form specialised bone-resorbing osteoclasts in the presence of receptor activator for nuclear factor kappa B ligand (RANKL). Non-canonical pathways of osteoclastogenesis have been described in which several cytokines and growth factors are able to substitute for RANKL. These humoral factors can generally be divided into those which, like RANKL, are tumour necrosis family (TNF) superfamily members and those which are not; the former include TNFα lymphotoxin exhibiting inducible expression and competing with herpes simplex virus glycoprotein D for herpesvirus entry mediator, a receptor expressed by T lymphocytes (LIGHT), a proliferation inducing ligand (APRIL) and B cell activating factor (BAFF); the latter include transforming growth factor beta (TGF-β), interleukin-6 (IL-6), IL-8, IL-11, nerve growth factor (NGF), insulin-like growth factor-I (IGF-I) and IGF-II. This review summarises the evidence for these RANKL substitutes in inducing osteoclast differentiation from tissue-derived and circulating mononuclear phagocytes. It also assesses the role these factors are likely to play in promoting the pathological bone resorption seen in many inflammatory and neoplastic lesions of bone and joint including rheumatoid arthritis, aseptic implant loosening and primary and secondary tumours of bone.

Investigation of osteoclastogenic signalling of the RANKL substitute LIGHT

LIGHT (TNFSF14) is a member of the TNF superfamily and is known to substitute for RANKL to induce osteoclast differentiation. LIGHT binds HVEM and LTβR, but it is not known whether these receptors play a role in osteoclast formation or whether LIGHT acts via RANKL signalling pathways. We found that both RANKL and LIGHT strongly induced phosphorylation of Akt and NFκB but not JNK in mouse osteoclast precursor cells. The addition of an Akt inhibitor showed decreased osteoclast differentiation and resorption mediated by both RANKL and LIGHT. RT-PCR and FACS analysis showed that CD14(+) human osteoclast precursors expressed HVEM and LTβR; expression levels of HVEM increased in the course of osteoclastogenesis and a decrease in LIGHT expression was associated with an increase in HVEM suggesting that there is a feedback loop related to this receptor. Our findings show that LIGHT is not inhibited by the soluble RANKL receptor OPG and that LIGHT is a potent osteoclastogenesis factor that activates the Akt, NFκB and JNK pathways.

Smooth muscle actin expression in primary bone tumours

Alpha isoform of smooth muscle actin (SMA) expression has been reported in giant cell tumour of bone (GCTB) and other benign and malignant bone tumours, but the pattern of SMA expression and the precise nature of SMA-expressing cells in these lesions is uncertain. We determined by immunohistochemistry the expression of SMA and other muscle and vascular markers in normal bone, GCTB and a wide range of primary benign and malignant bone tumours. Cultured stromal cells of GCTB, chondroblastoma (CB), and aneurysmal bone cyst (ABC) were also analysed for SMA expression. SMA was only noted in blood vessels in normal bone. SMA was expressed by mononuclear stromal cells (MSC) cultured from GCTB, ABC and CB. SMA was strongly and diffusely expressed by MSC in non-ossifying fibroma, fibrous dysplasia, and "brown tumour" of hyperparathyroidism. SMA expression was also noted in GCTB, ABC, CB, chondromyxoid fibroma, malignant fibrous histiocytoma of bone and osteosarcoma. Little or no SMA was noted in Langerhans cell histiocytosis, simple bone cyst, Ewing's sarcoma, osteoblastoma, osteoid osteoma, enchondroma, osteochondroma, chondrosarcoma, myeloma, lymphoma, chordoma and adamantinoma. Our findings show that there is differential SMA expression in primary bone tumours and that identifying the presence or absence of SMA is useful in the differential diagnosis of these lesions. The nature of SMA-expressing cells in bone tumours is uncertain but they are negative for desmin and caldesmon and could represent either myofibroblasts or perivascular cells, such as pericytes.

In vitro generation of mature human osteoclasts

Mononuclear precursors of human osteoclasts are found in the CD14(+) monocyte fraction of circulating peripheral blood mononuclear cells (PBMCs). It is possible to generate osteoclasts in vitro from PBMCs cultured with macrophage colony-stimulating factor and receptor activator for nuclear factor κB ligand. In these cultures, however, it is not possible to distinguish the effect of a specific agent on osteoclast resorption activity as opposed to osteoclast differentiation. To produce a population of mature human osteoclasts to study osteoclast lacunar resorption specifically, we cultured CD14(+) human monocytes on hydrophobic dishes in order to generate and maintain osteoclasts in suspension prior to culturing them on coverslips and dentine slices. Multinucleated cells formed in these cultures expressed vitronectin receptor, tartrate-resistant acid phosphatase, and cathepsin K. These cells also produced F-actin rings and were capable of extensive lacunar resorption on dentine slices after 24 h in culture. Lacunar resorption was inhibited by calcitonin and zoledronate but not by osteoprotegerin. This method of generating a highly enriched population of mature human osteoclasts should provide a valuable means of specifically assessing the effect of molecular factors (e.g., cytokines, growth factors, hormones) and therapeutic agents on osteoclast resorption activity.

RANKL-independent human osteoclast formation with APRIL, BAFF, NGF, IGF I and IGF II

Non-canonical pathways of osteoclastogenesis have been described in which several cytokines are able to substitute for RANKL. These cytokines are few in number and their role(s) in pathological bone resorption has not been ascertained. We have identified five additional cytokines, APRIL, BAFF, NGF, IGF I and IGF II, that can induce RANKL-independent osteoclastogenesis. All five cytokines induced both osteoclast differentiation and activation with respect to the formation of significant numbers of TRAP(+) and VNR(+) multinucleated cells that were capable of resorbing bone. The number of TRAP(+) multinucleated cells that formed was in the range of 40-75% of that supported by MCSF plus RANKL. Resorption was at a similar level to that induced by the other known RANKL substitutes TNFα, IL-6 and TGF-β. The addition of osteoprotegrin, the endogenous decoy receptor of RANKL, revealed that this resorption was independent of RANKL. APRIL, BAFF, IGF I and IGF II were found to be expressed in giant cell tumour of bone. IGF I and IGF II demonstrated very strong expression in the stromal cell population of all tumour samples. This data suggests that non-canonical osteoclastogenesis plays a role in both normal and pathological bone resorption.