What Hope Is There for Myeloma Bone Disease? 2011

What hope is there for myeloma bone disease? 2011

Myeloma bone disease causes severe morbidity in patients with myeloma characteristically including pain, loss of mobility and disfigurement. The pathological features of myeloma bone disease are osteoporosis, focal lytic lesions often leading to pathological fractures, vertebral collapse and hypercalcaemia (Figure 1). Neurological sequelae secondary to bone disease are common caused by compression of nerves by damaged and displaced bone and most dramatically include spinal cord compression which often presents as a neuro-surgical emergency occurring in up to 5% of patients with myeloma (Figure 2) (Kyle, Gertz et al. 2003). Indeed, consequences of osteolytic bone disease are often the presenting features of myeloma. Approximately 67% of patients with myeloma present with bone pain and up to 90% of patients with myeloma exhibit features of myeloma bone disease at some stage of the disease course (Coleman 1997; Kariyawasan, Hughes et al. 2007).

Figure 1 – Typical features of myeloma bone disease including from left to right – vertebral wedge fracture, ‘pepper-pot’ skull, pathological fracture arising from a lytic lesion of the humerus, large lytic lesion of the proximal tibia and lytic lesion causing extensive destruction of the mandible.

Last month we reviewed the exciting news from the MRC Myeloma IX trial demonstrating that not only is zolendronic acid superior to clodronate in the prevention of skeletal events events (27% vs 35%) but it is also associated with a survival advantage. Even after a relatively short median follow-up of 3.7 years, patients receiving zoledronic acid survive 5.5 months longer than those receiving clodronate. Progression free survival is also increased by 2 months (19.5 months vs 17.5 months) (Morgan, Davies et al. 2010). However, bisphosphonates are only partially successful in the prevention of skeletal events. New lesions appear and crucially, existing lesions are not repaired. There is therefore an urgent need for additional agents and strategies. So what else is under development?

Figure 2 – MRI images of spinal cord compression involving the cervical spine (left) and the thoracic spine (right).

Pathophysiology of myeloma bone disease

A brief review of the pathophysiology of myeloma bone disease provides the rationale for new therapeutic targets and strategies. Myeloma bone disease is caused by disruption of physiological bone remodelling specifically the uncoupling of the normal balance between bone resorption mediated by osteoclasts and bone formation mediated by osteoblasts resulting in net osteolysis (Mundy 1974; Bataille, Chappard et al. 1991; Roodman 1992; Bataille, Chappard et al. 1995; Bataille, Chappard et al. 1996; Croucher and Apperley 1998). Over thirty years ago, in studies of myeloma, Mundy et al. (1974) suggested that increased osteoclastogenesis was mediated by secreted osteoclast activating factors (OAFs). Histological and biochemical analysis has confirmed increased bone resorption associated with tumour infiltration (Taube, Beneton et al. 1992). Initially, the identity of the OAFs was uncertain. Since then many factors have been confirmed to have osteoclastogenic effects including interleukin 1β (Il-1β) (Cozzolino, Torcia et al. 1989; Kawano, Yamamoto et al. 1989; Yamamoto, Kawano et al. 1989), interleukin-6 (IL-6) (Lowik, van der Pluijm et al. 1989; Ishimi, Miyaura et al. 1990), interleukin-11 (Il-11) (Paul 1990), tumour necrosis factor-α (TNF-α) (Lichtenstein, Berenson et al. 1989), tumour necrosis factor-β (TNF-β) (Garrett, Durie et al. 1987; Bataille, Klein et al. 1989), macrophage inflammatory protein 1 α (MIP 1- α) (Wolpe, Davatelis et al. 1988), and hepatocyte growth factor (HGF) (Kukita, Nomiyama et al. 1997). It has been proposed that many of these osteoclast activating factors mediate their effects via the receptor activator nuclear factor kappa B ligand/osteoprotogerin (RANKL/OPG) system, which acts as a final common pathway (Simonet, Lacey et al. 1997; Lacey, Timms et al. 1998; Hsu, Lacey et al. 1999; Choi, Cruz et al. 2000).

Anti resorptive strategies

Given the above, the rationale for anti osteoclastic agents is clear. In addition to bisphosphonates, recombinant OPG and OPG mimetics, and anti RANK-L constructs, are under development (Body, Greipp et al. 2003; Body, Facon et al. 2006; Heath, Vanderkerken et al. 2007). An OPG mimetic has shown some anti resorptive effects in the 5T2MM murine model of myeloma, and no adverse effects in a preliminary phase 1 trial (Body, Greipp et al. 2003). More attention, however, has focused on the anti RANK-L monoclonal antibody, denusomab (Amgen). In 2006, a head to head comparison between pamidronate and denusomab demonstrated prompt and substantial reduction in urinary and serum N-telopeptide, a marker of bone resorption in both pamidronate and denusomab treated patients. However, the reduction in urinary and serum N-telopeptide was sustained for 84 days in the denusomab treated patients whilst the reductions were not sustained in the pamidronate treated patients.. The authors concluded that denusomab compared favourably given the lack of adverse effects and a sustained reduction in bone turnover for at least 84 days following a single s.c. dose (Body, Facon et al. 2006). However, apparent benefits, indicated by markers of bone resorption, are of limited significance until they are backed up by clear evidence of a reduction in hard clinical endpoints eg actual skeletal events. Hard clinical advantage is yet to be demonstrated for denusomab or other anti-resorptive strategies.

Bone formation is also inhibited in myeloma hence the rationale for bone anabolic strategies

Given the limited benefit of anti-resorptive strategies, attention has switched to bone anabolic strategies. In addition to increased recruitment and activity of osteoclasts, it has become apparent that osteoblast activity is also inhibited. Histomorphometric evidence suggests that early in the disease process there is an increase in the recruitment of both osteoclasts and osteoblasts. There are well documented reports of myeloma bone disease with an osteosclerotic phenotype (Roberts, Rinaudo et al. 1974; McCluggage, Jones et al. 1995; Mulleman, Gaxatte et al. 2004). However, typically as disease progresses osteoblast function is inhibited, osteoclast activity is stimulated resulting in net osteolysis (Bataille, Chappard et al. 1991). However, the molecular mechanism responsible for the inhibition of osteoblast function is not yet accounted for. Following biochemical analysis of markers of bone metabolism, Abildgaard et al. suggested that an unknown factor in myeloma was responsible for inhibition of later stage osteoblast differentiation and function leading to the uncoupling of osteoblast and osteoclast activity and net bone lysis (Abildgaard, Glerup et al. 2000).

Targeting Dkk-1 and other inhibitors of Wnt signalling

Evidence is now accumulating to suggest that the Wnt/β-catenin signalling pathway inhibitor, dickkopf (Dkk-1) contributes to osteolytic disease in myeloma via the inhibition of osteoblast differ entiation and therefore, may be if not the only osteoblast inhibitor active in myeloma bone disease, one of the critical factors driving osteolytic bone disease. The first important discovery to suggest this was made following gene array analysis of genes upregulated in myeloma (Tian 2003) .

These authors screened patients with multiple myeloma and radiological evidence of lytic lesions and compared them with patients with myeloma and no radiological evidence of lytic lesions. Other control groups included patients with other lymphoid malignancies eg Waldenström macroglobulinaemia, a condition related to myeloma but which does not lead to lytic disease, patients with monoclonal gammopathy of undetermined significance (MGUS), a precursor condition of myeloma and healthy controls. They identified overexpression of four genes in patients with myeloma and lytic lesions. One of these genes codes for Dkk-1. Furthermore, expression of the DKK-1 gene increased from a basal level in healthy controls, to patients with MGUS, to patients with myeloma and at least one lytic lesion and was highest in patients with myeloma with more than one lytic lesion. Bone marrow plasma levels of Dkk-1 were measured by enzyme linked immunosorbent assay (ELISA) in all patients and a similar trend was seen. The authors proceeded to examine the effects of recombinant Dkk-1, in the presence of bone morphogenic protein-2 on C2C12 osteoblast precursors and inhibition of osteoblast differentiation with the addition of anti-Dkk-1. They confirmed that the addition of Dkk-1 inhibited osteoblast differentiation and that this inhibition was reversed using an anti-Dkk-1 antibody. The correlation between serum Dkk-1 levels and osteolytic bone disease has been subsequently confirmed by other groups (Politou, Heath et al. 2006).

Figure 3 – Treatment with anti Dkk-1prevents myeloma induced reductions in trabecular and cortical bone volume in the 5T2MM murine model of myeloma (Heath et al 2009).

Since then a number of in vivo studies have demonstrated the effectiveness of targeting Dkk-1 in various murine models of myeloma (Yaccoby, Ling et al. 2006; Edwards, Edwards et al. 2007). In particular, our group has shown a reduction in osteolytic lesions, increases in bone volume and bone formation rate using an anti-Dkk-1 antibody (BHQ880, Novartis) in the 5T2MM murine model of myeloma (Figure 3) (Heath, Chantry et al. 2009). A humanised version of this anti-body is currently the subject of a large, multi-centered phase 2 trial in patients with myeloma and results are eagerly anticipated.

There is also increasing evidence to suggest that other inhibitors of Wnt/β-catenin signalling may play a role in the reduction of osteoblast numbers and inhibition of osteoblast function in myeloma bone disease and as such constitute potential therapeutic targets. For example, the role of sFRP-2 has been investigated (Oshima, Abe et al. 2005). These authors took conditioned media from the human derived myeloma cell lines RPMI8226 and U266 and primary myeloma cells and observed the effect on bone morphogenic protein induced alkaline phosphatase activity in osteoblasts precursors and in vitro mineralization. They found reduced alkaline phosphatase activity and mineralization. These cell lines and most of the primary myeloma cells used expressed sFRP-2 but not sFRP-1, sFRP-3 or Dkk-1.

Dkk-1 is also not the only Wnt component found to be upregulated in gene micro-array analyses of myeloma (De Vos, Couderc et al. 2001; Zhan, Hardin et al. 2002; Davies, Dring et al. 2003). De Vos et al. (2001) detected up-regulation of the gene encoding the Wnt inhibitor sFRP3/FRZB. Davies et al. (2003) demonstrated that FRZB was the most up-regulated gene during the progression from monoclonal gammopathy of uncertain significance to myeloma. Zhan et al (2002) demonstrated down regulation of Wnt 10B and up regulation of Wnt 5A and sFRP-3 (aka FRZB). These findings therefore suggest that there a number of different components and inhibitors of the Wnt/β-catenin signalling pathway implicated in the pathogenesis of myeloma bone disease.

Other osteoblast inhibitors active in myeloma bone disease

Interleukin-7 (Il-7)

There is also evidence to suggest other molecules and pathways play a role in osteoblast inhibition in myeloma bone disease. Investigators have reviewed the expression of seven of the known inhibitors of osteoblast differentiation, some of which act upon the Wnt signalling pathway (Dkk-1, sFRP-1, sFRP-2, sFRP-3, sFRP-4) and some which act on other pathways (gremlin, noggin, Il-7) in myeloma cell lines (Giuliani, Colla et al. 2005). They began by confirming the reduction of markers of osteoblast differentiation and nodule formation in the colony forming assay (colony forming unit – osteoblast, CFU-OB) when osteoblast precursors were co-cultured with myeloma cell lines. The expression and activity of the osteoblast differentiation transcription factor Runx2/Cbfa1 mRNA measured by RT-PCR was also reduced. They also noted that the inhibitory effect was significantly more marked with increased cell to cell contact reiterating the role of the very late antigen 4 (VLA-4) integrin system. They then co-cultured human bone marrow derived osteoblast precursors (Human BM PreOBs) with Il-7 and Dkk-1 and found a significant inhibitory effect with Il-7 but contrary to previous findings, not with Dkk-1. They therefore suggest that Il-7 is a key inhibitor of osteoblast differentiation and inhibitior of Runx2/Cbfa and furthermore may also influence osteoclast numbers via the reduction of OPG production in differentiated osteoblasts. This inhibiton was blocked by the addition of Il-7 antibodies but not by the addition of Dkk-1 antibody. Il-7 is secreted by T-cells and is known to stimulate osteoclast activation via up regulation of RANKL (Giuliani, Colla et al. 2002) but its’ effects on osteoblasts were previously unknown. The authors suggest that Il-7 probably does not act via Wnt signalling demonstrating the likely contribution of other pathways on osteoblast differentiation.

Hepatocyte growth factor

Hepatocyte growth factor (HGF) has also been implicated (Grano, Galimi et al. 1996). These authors identified HGF production by osteoclasts and expression of the HGF receptor by both osteoclasts and osteoblasts. These data suggested HGF mediated autocrine regulation of osteoclasts and paracrine regulation of osteoblasts and that HGF therefore acted as a coupling factor between osteoclast and osteoblast activity. Subsequently, a number of studies noted expression of HGF by myeloma cell lines and purified primary myeloma cells (Borset, Hjorth-Hansen et al. 1996; Seidel, Borset et al. 1998; Standal, Abildgaard et al. 2007). Standal et al. have recently demonstrated that HGF inhibits BMP induced expression of alkaline phosphatase in human mesenchymal stem cells and the murine osteoblast pre-cursor cell line C2C12.

Interleukin-3 (Il-3)

Similarly, recent studies have suggested that Il-3 plays a dual role in the pathogenesis of myeloma bone disease (Lee, Chung et al. 2004). Investigators demonstrated that Il-3 induced osteoclastogenesis and increased myeloma cell growth. More recently, the same group confirmed that Il-3 was significantly increased in the bone marrow plasma of patients with myeloma compared to healthy controls and that Il-3 was also able to inhibit basal and BMP induced osteoblast differentiation (Ehrlich, Chung et al. 2005).

TGF-β signalling also regulates osteoblast differentiation

Given the importance of osteoblast inhibition in the pathogenesis of myeloma bone disease, attention has also focussed on other pathways that may regulate osteoblastogenesis. Transforming growth factor beta (TGF-β) signalling is one such pathway (Geiser, Zeng et al. 1998). The TGF-β superfamily of cytokines, comprising TGF-βs, the bone morphogenic proteins (BMPs) and the activin/inhibin system, is a highly conserved system performing a wide variety of biological functions ranging from regulation of embryogenesis to regulation of reproduction and widespread tissue homeostasis (Guo and Wang 2009).