Review

Endocrine aspects of bone metastases Lorenz C Hofbauer, Tilman D Rachner, Robert E Coleman, Franz Jakob

Skeletal lesions are a frequent complication of breast and prostate cancer and a hallmark of multiple myeloma. Endocrine and paracrine factors modulate various aspects of bone metastases, including tumour proliferation, skeletal susceptibility to tumour homing, the microenvironment needed to support tumour persistence, and the initiation of a vicious cycle between tumour and bone-resident cells that further promotes tumour growth. Endocrine changes, such as oestrogen or vitamin D deficiency, contribute to a fertile bone microenvironment that might promote bone metastases. Bone health could be impaired further by existing cancer treatments, especially sex hormone deprivation. In this Review, we discuss the effect of hormones and associated local factors on cross-talk between bone metabolism and tumour biology. We review the biology of osteolytic and osteosclerotic lesions, with a focus on endocrine aspects, and outline potential therapeutic targets. We also summarise endocrine aspects of the pathogenesis and clinical presentation of bone metastases and provide an update on existing and future treatments.

Introduction Breast and prostate cancer are the most common malignancies in women and men respectively, and recent data suggest that their prevalence is increasing.1 Enhanced screening programmes such as prostatespecific antigen measurement and mammography, environmental factors, and increased life expectancy might contribute to this finding. Improved diagnosis and better treatment, including the use of more potent and specific regimens and adjuvant endocrine therapies, have prolonged survival.2 At present, 60–80% of patients with breast or prostate cancer develop bone metastases and, in light of the therapeutic advances, this number is likely to increase further because of longer survival times.3 In multiple myeloma, which is less common than breast or prostate cancer, osteolytic bone lesions occur in 90% of patients.4 Bone metastases result in so-called skeletal-related events—including pathological fractures, severe pain, and neurological compression syndromes—that can need surgery or radiotherapy (figure 1), and can also lead to hypercalcaemia and reduced quality of life.5 Thus, the prevention and treatment of skeletal metastases remains a clinical challenge across many disciplines. Bone metastases in breast cancer are mainly osteolytic, but some have a mixed or even osteosclerotic appearance (figure 1).3 In prostate cancer, bone metastases are predominantly osteosclerotic, but can have osteolytic features (figure 1). In recent years, endocrine pathways that are operative in bone and tumour biology have been deciphered as key regulators of tumour–bone cross-talk during the development of bone metastases.5,6 Although the roles of particular hormones, such as oestrogens and androgens, are well established and are therapeutically targeted in breast and prostate cancer, the effects of vitamin D and hormone-associated factors are more subtle and are incompletely understood. The bone health of patients with bone metastases can be impaired further by sex hormone deprivation therapy (frequently used in breast and prostate cancer), which changes bone remodelling (figure 2 and panel 1) and promotes rapid bone loss.13,14 Antiresorptive

agents including bisphosphonates and denosumab, an inhibitor of receptor activator of nuclear factor-κB ligand (RANKL), have emerged as new therapies for both bone metastases and treatment-related osteoporosis.15,16 Tumourinduced osteolytic bone loss can occur by enhanced osteoclast action. Additionally, the idea that inhibition of bone formation can enhance malignant bone disease through production of activin, dickkopf (DKK)1, and sclerostin has come into focus. The development of boneanabolic strategies targeting these pathways—for example, activin antagonists and DKK1 antibodies—opens up new perspectives for the treatment of malignant bone disease.8,9 In this Review, we summarise how particular hormones affect the bone metastatic process, with a focus on breast A

B

C

E

Lancet Diabetes Endocrinol 2014 Published Online January 24, 2014 http://dx.doi.org/10.1016/ S2213-8587(13)70203-1 Division of Endocrinology and Metabolic Bone Diseases, Department of Medicine III (Prof L C Hofbauer MD, T D Rachner MD) and Centre for Regenerative Therapies Dresden (Prof L C Hofbauer), TU Dresden, Dresden, Germany; Sheffield Cancer Research Centre, Weston Park Hospital, Sheffield, UK (Prof R E Coleman MD); Orthopaedic Center for Musculoskeletal Research, Wuerzburg, Germany (Prof F Jakob MD) Correspondence to: Prof Lorenz C Hofbauer, Division of Endocrinology, Diabetes and Bone Diseases, Department of Medicine III, Dresden TU Medical Centre, Fetscherstraße 74, D-01307 Dresden, Germany [email protected]

H

F

G

I

D

Figure 1: Clinical presentation of bone metastases Pathological fractures of the spine as seen on (A) CT scan and (B) MRI (red arrowheads); the hip on (E) x-ray and (F) CT scan (red arrowheads); or (C) the long bones (x-ray, red arrowhead) are typical hallmarks of bone metastases, as are neural compression syndromes, resulting from several vertebral fractures (red arrowheads in B). The phenotype of bone metastases can be osteolytic (red arrowheads in C and D; breast cancer), osteosclerotic (pale green arrow in F; breast cancer), or mixed with both osteolytic (red arrowheads in CT scans A and G; prostate cancer) and osteosclerostic lesions (pale green arrows in CT scans A and G; prostate cancer). Despite their osteosclerotic appearance, these bone lesions also fracture easily (red arrowheads in E and F). Other imaging modalities to visualise bone metastases (pale green arrows in H and I) include PET or CT (H) and bone scintigraphy (I).

www.thelancet.com/diabetes-endocrinology Published online January 24, 2014 http://dx.doi.org/10.1016/S2213-8587(13)70203-1

1

Review

A

MSC

Vitamin D BMP

Pre-osteoblast

Wnt RANKL RUNX2 M-CSF Osterix

Osteoblast

Apoptotic osteoblast

Sclerostin

Glucocorticoids Apoptotic osteoclast Osteoclast

Pre-osteoclast HSPC

c-fms

Lining cell

RANK

Osteocyte Bone

B

PTH

Sclerostin

Oestrogens Androgens

Wnt OPG RANKL

MSC/osteoblast RANKL RANK

Sphingosine-1-phosphate Ephrins Semaphorins F-actin

RANKL

Lysosomes

Osteoclast

Cathepsin K HCI Collagen

αvβ3

Sealing zone

Osteocyte Bone

Figure 2: Endocrine and paracrine regulation of bone remodelling (A) Osteoblasts differentiate from MSCs under the control of the transcription factors RUNX2 and osterix, vitamin D, BMPs, and Wnt signalling. In theory, osteoblasts can develop into osteocytes and lining cells, or undergo apoptosis under the control of glucocorticoids. Mature osteoblasts produce an extracellular matrix rich in type 1 collagen that is mineralised and serves as a scaffold. Osteocytes are embedded in the mineral where they act as mechanosensors and regulate mineral homoeostasis. MSCs, osteoblasts, and osteocytes express RANKL, a key osteoclast regulator. Osteoclasts differentiate from HSPCs in the presence of M-CSF (which binds to c-fms) and RANKL. This process is controlled by the systemic and local hormone milieu. RANK activation by RANKL promotes osteoclast proliferation, fusion, and activation, which thus increases the pool of active osteoclasts. (B) Osteoblast and osteocyte-derived factors (Wnt molecules, OPG, and RANKL) and hormones (PTH) targeting the osteoblast control bone resorption by osteoclasts. OPG is a decoy receptor that inhibits RANKL and bone resorption. Bone resorption requires the unique ability of osteoclasts, which are multinucleated cells that migrate through F-actin rearrangement and adhere to bone by αvβ3-integrins. This process creates a sealing zone, an acidic microenvironment that allows degradation of bone matrix with proteases such as cathepsin K. Osteoclasts also secrete local factors that control osteoblast functions. Both osteoblast and osteoclast functions are regulated by androgens and oestrogens through high-affinity receptors. Osteocytes orchestrate bone remodelling by producing the bone formation inhibitor sclerostin and the osteoclast-promoting cytokine RANKL, which enhances bone resorption. Osteoclast-derived factors, including sphingosine-1-phosphate, ephrins, and semaphorins in turn regulate osteoblast function. MSC=mesenchymal stromal cell. BMP=bone morphogenetic protein. RANKL=receptor activator of nuclear factor-κB ligand. M-CSF=macrophage colony-stimulating factor. HSPC=haemopoietic stem and progenitor cell. PTH=parathyroid hormone. OPG=osteoprotegerin.

cancer, prostate cancer, and multiple myeloma. We discuss the role of sex hormones and hormone ablation on skeletal metastases and bone health, and review the latest therapeutic advances in the emerging field of osteooncology, and their interactions with the endocrine system.

Sex hormones and the RANK–RANKL pathway in the breast cancer–bone interface Sex hormones have a major role in the pathogenesis of breast cancer. The effect of oestrogens in breast cancer depends on the patient’s menopausal status, which defines the concentrations of circulating oestrogens, and on the hormone receptor status of the tumour, which affects its susceptibility to oestrogen exposure. Since normal breast tissue expresses both oestrogen and progesterone receptors, 2

tumours that are positive for these two hormone receptors tend to be more differentiated with a better overall prognosis than hormone receptor-negative tumours.17 Oestrogens enhance the proliferation of oestrogen receptor-positive tumours, which provides the rationale for anti-hormone therapies that have substantially improved survival. Oestrogens are also essential regulators of skeletal remodelling (panel 1).18 In women, the presence of oestrogens promotes acquisition of peak bone mass and reduces bone turnover. After menopause, oestrogen deficiency is the main cause of bone loss in women. Some of the concurrent oestrogen effects on bone metabolism and the metastatic process are mediated by the RANK– RANKL–osteoprotegerin system (figure 2). During early menopause, women have raised RANKL concentrations in mesenchymal stromal cells, B lymphocytes, and

www.thelancet.com/diabetes-endocrinology Published online January 24, 2014 http://dx.doi.org/10.1016/S2213-8587(13)70203-1

Review

T lymphocytes compared with premenopausal women, which is reversed by hormone replacement therapy.19 After menopause, the conversion of adrenal androgens into oestrogens by the aromatase enzyme in fat tissue is responsible for residual systemic oestrogen concentrations. However, local aromatase expression in the skeleton and adipose tissue can create a unique endocrine microenvironment.20 In hormone receptor-positive breast cancer, this conversion is targeted by aromatase inhibitors, including letrozole, anastrozole, and exemestane, leading to almost complete suppression of oestradiol. Aromatase inhibition has become a cornerstone in adjuvant therapy of postmenopausal women with hormone receptorpositive breast cancer. However, its use is also associated with rapid bone loss and increased fracture risk, which emphasises the importance of normal postmenopausal concentrations of oestrogens for bone health.21 Furthermore, increased bone turnover after aromatase inhibition can mobilise haemopoietic stem cells from the bone marrow niche, thus contributing to reactivation of dormant cancer cells or conditioning of premetastatic niches for disseminated tumour cells.22 This concept warrants clinical validation with antiresorptive agents such as bisphosphonates and denosumab. Progesterone and progestins affect both bone and breast cancer biology. The role of progesterones was emphasised by the results of the Women’s Health Initiative and the Million Women Study, which indicated that women receiving hormone replacement therapy had a significantly increased risk of breast cancer.23,24 Until then, hormone replacement therapy was used widely on the basis of observational studies, which indicated that their use was associated with reduced cardiovascular disease, relief of menopausal symptoms (such as hot flushes and sexual dysfunction), and prevention of postmenopausal bone loss.25 However, the Women’s Health Initiative and the Million Women Study showed a roughly two-times higher risk of breast cancer after combination therapy of progesterone and oestrogens than with oestrogen alone (which caused a roughly 1·3-times higher risk).23,24 Indeed, subgroup analyses of the Women’s Health Initiative Estrogen-Alone trial in patients receiving only conjugated equine oestrogens showed a significant reduction in invasive breast cancer compared with patients given placebo.26 A 10-year followup of this study also showed reduced breast cancer incidence (0·27% in the treatment group vs 0·35% in the placebo group; HR 0·77; 95% CI 0·62–0·95) after oestrogen use for 5·9 years.27 These effects were especially apparent in women starting hormone replacement therapy more than 5 years after the onset of menopause.26 In turn, these results suggested a notable role of progesterone in the pathogenesis of breast cancer. Preclinical studies had previously shown that the progestin medroxyprogesterone acetate promotes the growth of mouse mammary tumours,28 and recent studies have linked this effect to the RANK–RANKL pathway.29,30

Panel 1: Cellular basis of bone remodelling Bone remodelling—the coordinated sequence of bone resorption by osteoclasts— followed by synthesis of a bone matrix by osteoblasts (figure 2) underlies constant skeletal turnover.7 Bone remodelling occurs under systemic and local control, with integration of endocrine, paracrine, neural, and mechanical stimuli. Androgens, oestrogens, and vitamin D are key systemic regulators of bone metabolism. Recently, the local role of cytokines, chemokines, and growth factors (such as the activin and inhibin system) on bone cell homoeostasis has been increasingly acknowledged.7–9 Osteoclasts differentiate from haemopoietic stem and progenitor cells in the presence of macrophage colony-stimulating factor and receptor activator of NF-kB ligand (RANKL) (figure 2A). In particular, RANKL, which is upregulated in oestrogen deficiency and vitamin D insufficiency, stimulates RANK on osteoclasts to promote their proliferation and activation, thus increasing the pool of active osteoclasts. Osteoprotegerin, a decoy receptor that inhibits RANKL and bone resorption, is upregulated in osteoblasts by oestrogens.10 Osteoclasts adhere to bone by αvβ3-integrin and rearrange their cytoskeleton (figure 2B), thus creating a sealing zone, which is acidified and enriched with collagen-degrading enzymes such as cathepsin K.7 Osteoblasts differentiate from mesenchymal stromal cells through the actions of the transcription factors RUNX2 and osterix, vitamin D, bone morphogenetic proteins (BMPs), and Wnt signalling—a complex pathway with several ligands, receptors, and inhibitors, such as sclerostin and dickkopf (DKK)-1. Mature osteoblasts produce a new extracellular matrix rich in type 1 collagen that is mineralised and serves as a natural scaffold. Osteoblasts undergo apoptosis, develop into lining cells, or give rise to osteocytes (figure 2A), which are embedded in the mineralised matrix where they act as mechanosensors and regulate mineral homoeostasis.11 Osteocytes orchestrate bone remodelling by producing the bone formation inhibitor sclerostin and the osteoclast-promoting cytokine RANKL, which enhances bone resorption.11,12 Osteoblast and osteoclast functions are affected by osteotropic cancer types, through direct cellular interactions or interference from key cytokines of bone biology, which are either directly produced by the cancer cells or increased by paracrine mechanisms.

Medroxyprogesterone acetate upregulates RANKL expression in mammary epithelial cells, activates RANK signalling, induces mammary stem cell expansion, and protects against tumour cell apoptosis. Deletion of RANK in mammary epithelial cells reduces the occurrence of medroxyprogesterone acetate-driven breast cancer and suppresses the capacity of breast cancer stem cells.29 In a proof of principle study, inhibition of RANKL also reduced tumorigenesis in rodents.30 Although these results were obtained in animals, they highlight the potential role of progesterone in mammary carcinogenesis and link it to the RANKL pathway. Since RANKL can be targeted therapeutically by the monoclonal antibody denosumab, it will be important to ascertain whether denosumab affects breast cancer development in human beings. Treatment of menopausal symptoms in women with breast cancer remains a challenge, especially because their symptoms tend to be more severe than those of postmenopausal women without breast cancer. Studies of hormone replacement therapy in patients after breast cancer have produced ambiguous results. A safety study of hormone replacement therapy after breast cancer was stopped because of an increased

www.thelancet.com/diabetes-endocrinology Published online January 24, 2014 http://dx.doi.org/10.1016/S2213-8587(13)70203-1

3

Review

relapse rate.31 However, a 10-year follow-up trial showed no differences in new breast cancer events and distant metastases.32 Thus, in light of the renaissance of hormone replacement therapy, more research is needed to clarify the safety profile of hormone replacement in women who have a history of breast cancer. A large body of evidence exists about the role of vitamin D receptor signalling and breast cancer biology. However, how vitamin D affects homing of breast cancer cells into bone is not fully understood; investigations of this issue have been confounded by the experimental design of preclinical studies involving vitamin D deprivation. In principle, secondary hyperparathyroidism that results from vitamin D insufficiency might enhance bone turnover through the RANK–RANKL pathway, which, along with oestrogen deficiency, might activate osteoclasts (figure 2), and could result in conditioning of a premetastatic niche for tumour cells.22,33 Whether repletion of vitamin D deficiency could prevent bone metastases in breast cancer needs to be assessed.

Androgens and vitamin D in prostate cancer– bone cross-talk Serum concentrations of androgens or other sex hormones are not associated with the risk of developing prostate cancer.34 However, in established prostate cancer, about 80% of men with advanced disease respond to androgen deprivation therapy, and this treatment has become the standard of care for advanced prostate cancer.35 Androgen deprivation therapy is also considered as monotherapy for localised high-risk prostate cancer.36 Androgen deprivation is based on suppression of androgen production and androgen receptor signalling, which can require a combination of gonadotropinreleasing hormone agonists (leuprorelin or goserelin) or  antagonists (degarelix) and androgen receptor antagonists (flutamide or bicalutamide). Since 5α-reductase converts testosterone to the high-affinity androgen receptor ligand 5α-dihydrotestosterone, specific 5α-reductase enzyme inhibitors, such as finasteride and dutasteride, can be used. Prostate-specific antigen, which is expressed by most prostate cancers, is regulated by androgens and is a useful tumour marker to monitor therapy. However, castration-resistant prostate cancer is a clinical problem that often develops as a result of reactivated androgen receptor signalling caused by increased intracrine androgen synthesis or the conversion of adrenal androgens into testosterone and 5α-dihydrotestosterone by tumour cells.37 In advanced disease, adrenal androgen synthesis can be blocked by abiraterone, an inhibitor of 17α-hydroxylase and 17,20-lyase, which depletes the precursor for local androgen production.38 Another option is the use of the new potent androgen receptor antagonist enzalutamide.39 Men receiving androgen deprivation therapy experience rapid bone loss, which suggests that testosterone is the main sex hormone regulating male bone mass. However, 4

the conversion of androgens to oestrogens by aromatisation is also crucial for the health of the male skeleton.40 Androgen deprivation therapy has been shown to promote epithelial-to-mesenchymal transition in prostate cancer cells, which is a prerequisite for skeletal metastasis.41 The use of androgen replacement therapy, for example for pre-existing hypogonadism, in men with prostate cancer is under debate. The view that androgen replacement unequivocally stimulates tumour growth in men after radical prostatectomy, external-beam radiotherapy, or brachytherapy has been questioned.42 Prostate cancer and adverse outcomes have also been linked to vitamin D deficiency and secondary hyperparathyroidism. Although higher prediagnosis vitamin D concentrations were associated with an improved prognosis of prostate cancer, raised parathyroid hormone concentrations were associated with a 43% increased risk of developing the disease.43 Preclinical evidence also suggests that the expression of vitamin D-metabolising enzymes is relevant for prostate cancer development and propagation. Secondary hyperparathyroidism induced by severe vitamin D deficiency promoted bone metastasis in prostate and breast cancer in mice,44,45 although this experimental setting affects several variables (vitamin D and parathyroid hormone) and enhances bone turnover.44

Oestrogens, vitamin D, and multiple myeloma Emerging in-vitro evidence suggests that oestrogens affect myeloma growth and activity. However, the effects of 17β-oestradiol on myeloma seem to be diverse, and antiproliferative effects were shown in some, but not all, myeloma cell lines. Oestrogens inhibit interleukin 6 production by immune cells, and thereby abolish interleukin 6-dependent myeloma proliferation.46 Furthermore, oestrogens can reverse the inhibitory effects of myeloma on osteogenic differentiation of osteoblasts through regulation of the RANKL–osteoprotegerin pathway.47 Little in-vitro evidence indicates that phytoestrogens can affect myeloma biology. The phytoestrogen genistein inhibited proliferation and induced apoptosis in a range of haematological malignancies, including transformed myeloma cell lines, and seemed to target the NF-κB signalling pathway.48–50 Although these findings are interesting, especially in the context of disease prevention, and warrant further investigation, they have not been validated in in-vivo models or tested formally in clinical trials. Few data exist about the direct effects of vitamin D on myeloma cells in vitro. Vitamin D has been shown to reduce cell viability and to increase apoptosis in murine myeloma cells.51 Moreover, several investigators have assessed the antitumour potential of calcitriol in preclinical models of solid and haematological malignancies, including myeloma,52 but few have specifically addressed the role of vitamin D in myeloma bone disease in human beings. Vitamin D deficiency is

www.thelancet.com/diabetes-endocrinology Published online January 24, 2014 http://dx.doi.org/10.1016/S2213-8587(13)70203-1

Review

highly prevalent in patients with multiple myeloma (typically elderly people), and one study showed that 40% of affected patients had 25-hydroxyvitamin D deficiency (