Giant cell tumour of bone: new treatments in development

Giant cell tumour of bone (GCTB) is a rare osteolytic tumour that is responsible for approximately 6 % of all primary bone tumours. Reported annual incidence ranges between 1 and 6 per 10 million persons [1, 2] to approximately 1 per million in the US, Western Australia, Japan and Sweden [3]. It typically affects adults aged between 20 and 50 years [4‐6], with a slightly higher incidence among females (1.7 per 10 million in females versus 1.5 per 10 million in males) [1, 4‐6]. GCTB is typically located in the epiphysis of bones, causing localised tenderness and swelling, reduced joint mobility, and pain that is often severe and intractable [6]. It usually develops in long bones but can also occur in unusual locations. The Enneking staging classification, based on radiological, histological and clinical features, is the most commonly used (Table 1) [7]. There is also a radiological grading system established by Campanacci et al. [8] that classifies GCTB into three radiographic types (I, intramedullary lesion confined to bone; II, thinned, expanded cortex, III, cortical breakout), and is roughly comparable with the staging system of Enneking et al. [7]

Symptoms are variable; some patients may be asymptomatic until they develop a pathologic fracture while others complain of pain at the adjacent joint and limited range of motion. There may also be swelling and even a visible mass, if the tumour has grown for a long time. Other commonly reported symptoms include muscular or nerve pain [6].

The tumour is locally aggressive and destructive, and it grows rapidly, destroying bone and spreading into surrounding soft tissues [9]. If it is surgically resected, there is a substantial probability of recurrence, which seems to be greater in some locations associated with more difficult treatment, such as the distal radius and the proximal femur [10]. In the absence of treatment, the continued and unchecked tumour growth leads to complete destruction of the bone, physical deformity and the possibility of loss of limb.

The most common site of metastasis is the lung, occurring at a frequency of 1–6 % [11, 12]. Pulmonary metastases are usually histologically benign and their course is indolent. The standard of care is surgical resection, and prognosis is generally good. If resection is not possible, they can be left untreated [6, 11, 13].

Rarely, in less than 1 % of cases, GCTB may undergo malignant transformation that is known to result in a poor prognosis for the patient [6, 14]. The malignancy may arise as a result of dedifferentiation of the primary tumour or secondary to radiation therapy (approximately 50 % of cases) [11]. The most commonly observed transformation is to a high-grade sarcoma, usually an osteosarcoma, however, in rare cases this transformation may result in the formation of a fibrosarcoma or a classically denominated malignant fibrous histiocytoma. The mean time after initial GCTB diagnosis to malignant transformation is around 19 years in patients with spontaneous transformation and around 9 years in post-radiation cases [11].

The histopathology of GCTB reveals the presence of marked haemorrhage and three major cell types: multinucleated giant cells, stromal cells and mononuclear cells of the monocyte/macrophage lineage [15, 16].

The spindle-like stromal cells are the main neoplastic components, and appear to be activated by fibroblasts that secrete type I and III collagen and possess parathormone receptors [17, 18]. They promote giant cell formation by expressing and secreting a variety of chemotactic factors (cytokines such as interleukin [IL]-6, IL-8, IL-11, IL-17, IL-34, basic fibroblast growth factor [b-FGF], tumour necrosis factor [TNF]-α, vascular endothelial growth factor [VEGF], macrophage colony-stimulating factor [M-CSF], RANKL, cathepsin K; chemokines such as IL-8, TGF-β1 and stromal derived factor-1 [SDF-1]; and enzymes such as matrix metalloproteinase [MMP]-9 and MMP-13) [19‐22]. All these factors serve to engage and differentiate circulating monocytes into macrophages [23]. Of these factors, SDF-1 appears to act as a chemoattractant involved in the recruitment of monocytes [23]. Furthermore, some studies have correlated the expression of VEGF and MMP-9 with the extent of bone destruction and probability of recurrence [24].

Giant cells are directly responsible for the increased bone resorption observed within the lesion [25, 26]. They are considered to be reactive macrophages that have acquired osteoclastic activity as a result of their stimulation by stromal cells, which modifies their gene expression pattern within the osseous environment [27]. Giant cells also drive increased expression of a key mediator in osteoclastogenesis: the RANK receptor [28]. Activation of this receptor by RANKL, which is secreted by stromal cells, promotes osteoclast formation, activation, function, and survival [29‐32]. Thus, leading to the increased level of bone resorption observed within the GCTB lesion. In addition, the activated osteoclasts, in turn, release tumour growth factors into the bone microenvironment, initiating a tumour/bone vicious cycle (Fig. 1a) [33‐35].

The underlying cause of the increased RANKL expression by stromal cells is unknown, however, this phenomenon is reduced after elimination of the giant cells [32]. Conversely, giant cells are clearly dependent on RANKL signalling by stromal cells [32, 36]. Thus, it is possible that GCTB promotes a pathological variation of the normal physiological interdependence of osteoblast and osteoclast populations in bone [37, 38].

Cytogenetic abnormalities have been observed in up to 72 % of patients with GCTB, yet to date, no uniform aberrations have been identified [15, 39, 40]. Telomeric associations (reductions in telomere length with an average loss of 500 base pairs) are the most frequent chromosomal aberrations and the telomeres most commonly affected are 11p, 13p, 14p, 15p, 19q, 20q and 21p [39, 40].

There is also a hypothesis that the origin of GCTB could be linked to a form of bone injury, as in some cases GCTB appears in locations associated with prior trauma [41, 42]. In this scenario, GCTB could be considered a local reactive condition secondary to a haemorrhage due to bone injury and/or defective collagen in the matrix or in the vessel wall. It is possible that the haemorrhage serves to provide fresh monocytes and plasma proteins that initiate activation of stromal cells, which in turn stimulate conversion of giant cells into active osteoclasts [41]. Once the primary lesion occurs, the stromal cells would be capable of re-forming the tumour in secondary tumour sites or after surgical removal, thanks to their proliferative and tumour-initiating properties [17, 43, 44]. However, it seems that other transformational factors are required since injection of isolated stromal cells into immunocompromised mice does not produce giant cells [45, 46]. Some studies suggest that metastases could result from tumour emboli travelling to distant sites [47, 48].

Surgical removal of the tumour with wide excision or intralesional curettage and placement of cement (polymethyl methacrylate) has been historically the preferred treatment for GCTB [10, 49‐52]. The challenge with surgery is to remove as much of the tumour as possible while leaving the joint intact. Wide excision is associated with poor functional outcome and greater surgical complications [53‐55]. Therefore, intralesional curettage has been the mainstay of treatment for the majority of patients with stage I or II tumours. Wide excision is usually reserved for more aggressive stage III tumours with extraosseous extension or otherwise unresectable tumours [6, 56, 57]. However, sometimes the tumour is unresectable or surgery is not recommended due to age, patient comorbidities or risk of severe morbidity, such as joint removal or loss of limbs.

Aggressive GCTBs may require wide excision and reconstruction with a modular endoprosthesis; the most commonly used synthetic grafts are made from polymethyl methacrylate (PMMA). These grafts are known to generate an exothermic reaction that increases thermal necrosis of tumour cells and an inflammatory reaction, consequently resulting in an improved patient recovery and tumour removal [58, 59].

The main complications associated with surgery include pathologic fracture and postoperative infection. Postoperative infection occurs in 2–25 % of patients, and its incidence is probably greater with more extensive surgery involving en bloc resection and placement of an endoprosthesis [60‐63]; whereas pathologic fracture is associated with an increased rate of recurrence and a poorer functional outcome [64].

There is a recognised tendency for GCTBs to recur locally in many cases following surgery, even in the soft tissues adjacent to the primary bone location [5, 10, 65, 66]. In one of the largest published cohorts, a multicentre retrospective study in 294 Scandinavian patients, Kivioja et al. [52] reported recurrence rates ranging between 20 % for patients with PMMA cementation following intralesional curettage, and 56 % for patients without cementation. In contrast, wide excision is reported to be associated with a lower risk of local recurrence (0–12 %) than intralesional curettage (12–65 %) [5, 67‐69]. The use of improved surgical techniques, such as extensive mechanical burr drilling of the tumour wall after curettage or adjuvant cryoablation with liquid nitrogen, has further decreased recurrence rates in some centres, but these techniques have not yet been widely adopted. In the study by Malawer et al. [70] only 2.3 % of patients recurred after primary treatment with cryosurgery, although this percentage increased to 7.9 % when second-line treatments were also considered.

Currently, there is no standard or approved first-line medical treatment for GCTB. Surgical treatment may be combined with chemical adjuvant therapies. Some of the treatments commonly applied to the affected area are: alcohols [59, 71, 72], phenol [71, 73], hydrogen peroxide [71, 74, 75], and zinc chloride [76]. Hydrogen peroxide has been found to increase the penetration of phenol into the surrounding tissues [75]. Use of chemical adjuvants has been shown to reduce the percentage of recurrences in some studies [77], although others failed to demonstrate any impact [78]. Furthermore, these adjuvants must be used with caution, to avoid chemical burns.

Radiation therapy has been used to treat GCTB since 1932 [79] and its efficacy has been demonstrated by several studies in patients for whom surgery was not feasible [80, 81]. Specialised techniques such as 3-D conformal radiotherapy (RT) and intensity-modulated radiotherapy (IMRT) have been associated with good local control rates in patients with GCTB in locations that are not accessible by surgical resection [82, 83]. However, some reports have suggested an increased risk of malignant transformation into post-radial sarcoma [84].

The better safety profile of the new drugs available to inhibit osteoclastogenesis has decreased the use of RT in GCTB [85].

Embolisation is made by hyperselective catheterisation and embolisation of the arteries that feed the pathological lesion with the most appropriate embolic agent. Typically, Gelfoam, polyvinyl alcohol (PVA) particles, and coils are used for embolisation; other agents include tissue adhesive, ethanol, and microfibrillar collagen. Occlusion of the vessels decreases the volume of the tumour, but multiple procedures are frequently necessary [86]. Photoablation with an argon laser is another therapy that can lead to successful tumour necrosis [87].

Given the high vascularity and morbidity associated with surgical resection and/or radiation therapy, embolisation has been reported to be useful within 24–48 h prior to these therapies [88, 89], to prevent recanalisation. The combined use of preoperative embolisation and adjuvants, including radiation therapy and intraoperative phenol and nitrogen, can decrease local recurrence to less than 10 % [90].

Serial embolisation is also used as primary treatment in some patients with GCTB of the extremities, especially for tumours with large cortical defects or joint involvement and for those with large GCTBs of the sacrum. This procedure has a low morbidity rate and has been shown to be effective in preserving function and relieving pain in selected patients [91‐93].

A case series also suggested that preoperative treatment with denosumab induces dramatic sclerosis and reconstitution of cortical bone, achieving tumour necrosis in 90 % of patients. The authors reported that, after denosumab treatment, subsequent surgical resection was easier in cases of aggressive tumours and that denosumab should also be considered as a stand-alone treatment in patients who are poor surgical candidates or in cases where the tumour is in a location difficult to treat surgically [101]. There are also some case reports of successful use of denosumab in children [102], although it has not been formally assessed in this population and is not recommended for use.

Due to their anti-resorptive properties, some exploratory studies tested the efficacy of bisphosphonates in GCTB. It was shown that nitrogen-containing bisphosphonates induce apoptosis in both giant cells and stromal cells in vitro [106]. In a case–control study, pamidronate and zoledronate reduced local tumour recurrence (4.2 vs 30 % in the control group, p = 0.056) and controlled disease progression when used orally or intravenously as adjuvant therapy to intralesional curettage [107]. In 25 patients with recurrent and metastatic GCTB treated with bisphosphonates, stabilisation of disease was achieved in most cases refractory to conventional treatment [108]. In addition, there are case reports of successful local administration of zoledronic acid as adjuvant therapy during surgery [109]. However, they are not approved for use in this indication and more evidence is needed.

The knowledge of GCTB pathophysiology is rapidly evolving. The identification of the chemotactic factors secreted by stromal cells and involved in monocyte transformation into giant cells provides an opportunity to discover innovative treatments. The monoclonal antibody denosumab is the first drug agent with proven efficacy in GCTB by targeting one of these factors (RANKL). The main pending questions with denosumab include the evaluation of its possible benefits as neoadjuvant therapy [112], the optimal duration and schedule of treatment at long term to avoid recurrences, and its long-term safety. Some angiogenesis inhibitors have also been tested, such as calcitonin and interferon. IFN-α inhibits the expression of b-FGF and IL-8, two angiogenic factors. Other candidate therapies could be monoclonal antibodies directed against the involved cytokines or enzymes, such as anti-IL6, cathepsin inhibitors, anti-M-CSF or MMP-specific inhibitors [113]. The newer antibody–drug conjugates (ADCs), a novel class of highly potent drugs composed of an antibody (a whole antibody or an antibody fragment) linked to a cytotoxic drug could revolutionise treatment of GCTB [114]. Although few ADCs are currently available [115], there are more than 20 compounds currently in clinical development, specific for a wide range of biological targets expressed by tumour cells [116]. It is hoped that, in the near future, some of them could be suitable for GCTB, in view of promising results in other cancers.

It also seems that targeting the neoplastic stromal cells could fight directly against the origin of tumour. Therapies blocking proliferation of stromal cells, such as drugs inhibiting cell cycle progression or telomerase activity could be effective. First, it would be necessary to identify specific markers for the stromal cells.

Recent findings suggest that the haemorrhagic component plays a fundamental role in the development of giant cells. In some instances, GCTB could be a reactive condition secondary to massive intraosseous haemorrhage, which attracts monocytes and forces their quick proliferation and conversion into multinucleated cells. There is also the hypothesis that poor matrix support to the vessels may underlie the haemorrhage that precedes tumour formation. Currently, the use of embolisation techniques and occlusion of the vessels helps reduce recurrence. Other treatments aimed to occlude the vessels and reinforce local osseous matrix support, such as laser and hormone therapies, could be also effective.

A more deep investigation on genetic predisposition may help to identify individuals at higher recurrence risk, in whom more aggressive therapies should be undertaken. For example, amplification of 20q11.1 seems to be a prognostic marker for adverse outcome [117] and warrants further investigation.