The metastatic microenvironment of breast cancer: Clinicalimplications

R.E. Coleman a,*, W. Gregory b, H. Marshall b, C. Wilson a, I. Holen a

aAcademic Unit of Clinical Oncology, CR-UK/YCR Sheffield Cancer Research Centre, Weston Park Hospital, Witham Road,Sheffield S10 2SJ, UKbClinical Trials Research Unit, University of Leeds, Leeds, UK

Keywords:BisphosphonatesAdjuvant therapyBone marrow microenvironmentMetastasis

a b s t r a c t

Metastasis to bone, and indeed potentially to other sites, results from the numerous interactions be-tween cancer cells, haematopoietic stem cells and normal bone cells within the bone marrow micro-environment. These interactions are in turn influenced by multiple endocrine, paracrine and physicalfactors. Bone-targeted treatments may modify the course of the disease via both direct and indirectinhibitory effects on this “vicious cycle” of growth factor and cytokine signalling between tumour andbone cells.

Improvements in both disease free (DFS) and overall survival in women with early breast cancer havebeen demonstrated in several large randomised adjuvant trials of oral clodronate and intravenouszoledronic acid. The evidence for a beneficial impact on disease outcome is particularly strong in patientswith low levels of reproductive hormones, including pre-menopausal women receiving ovarian sup-pression therapy and those who have passed through menopause at the time of diagnosis. A recent meta-analysis of postmenopausal women treated with adjuvant bisphosphonates showed an 18% improve-ment in DFS (hazard ratio [HR] ! 0.82; 95%CI 0.74e0.92, 2P ! <0.001), with reductions in relapse ratesnot only in bone but also at extra-skeletal and loco-regional sites. These exciting findings are beginningto change clinical practice.

! 2013 Elsevier Ltd. All rights reserved.

Introduction

Metastasis is complex and comprised of a multistage process ofspread of malignant tumour cells from the primary tumour tosecondary organs. Initially tumour cells must acquire an invasivephenotype in order to invade the surrounding stroma and disruptthe vascular endothelium. Once within the blood stream, thetumour cells must survive in the circulation, and evade bothphysical damage and attack by the immune system. Some of thecirculating tumour cells will find a favourable vascular site toadhere to and subsequently invade through the capillary wall into anewmicroenvironment (Fig.1). Here the disseminated tumour cells(DTC) need to adapt to in order to survive and either start to form amicrometastasis or undergo genetic reprogramming to enter into aquiescent or dormant state that can last for years.

Several metastasis suppressor genes, defined by their ability toprevent the development of metastasis by inducing apoptosis or

dormancy once the cells have lodged at the secondary site, appearable to respond to the microenvironment and regulate thisdormancy state [1]. These genes produce proteins that can partic-ipate in a diverse range of signalling pathways, sometimes inhib-iting not just one but multiple steps in the metastatic cascade [2].The tumour microenvironment is clearly an important regulator ofcancer progression [3] at multiple levels (Fig. 2) and a critical factorfor the survival and growth of DTC at preferential metastatic sites[4].

Tumour dormancy describes the stage in tumour progressionafter treatment of the primary cancer during which malignant cellsremain occult until a range of poorly understood immunological,angiogenic and tumour factors promote subsequent growth andclinical recurrence. For example, in patients with breast cancer,relapse can occur years or even decades after the initial diagnosis.Dormant tumour cells appear to be more resistant to adjuvanttherapies, and are probably largely responsible for the continuingrate of recurrence and death of patients despite the enlarging rangeof currently available adjuvant systemic treatments [5]. Impor-tantly, mathematical models of breast cancer dynamics indicatethat the period of tumour quiescence or dormancy is typically

* Corresponding author. Tel.: "44 114 2265213; fax: "44 114 2265678.E-mail address: r.e.coleman@sheffield.ac.uk (R.E. Coleman).

Contents lists available at SciVerse ScienceDirect

The Breast

journal homepage: www.elsevier .com/brst

0960-9776/$ e see front matter ! 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.breast.2013.07.010

The Breast 22 (2013) S50eS56

followed by rapid growth after emergence from dormancy, ratherthan a slow, steady growth of micro-metastatic disease [6].

The presence of detectable circulating tumour cells (CTC) in theperipheral blood of patients, as well as DTC in the bonemarrow hasbeen shown by many groups to be an independent prognosticfactor [7]. However, not all patients with detectable CTC or DTCappear to relapse and in many patients that do relapse, neither CTCnor DTC can be detected despite multiple sampling. In addition toevaluating the crude number of these potentially metastatic cells,there is great interest in exploiting the technologies now availableto performmolecular characterization of individual cells in order toidentify novel targets that might be exploited to keep residualtumour cells in a dormant state indefinitely after initial local and

systemic treatments. Currently, decision-making for the use ofadjuvant treatments is essentially based on tumour characteristics.However, our increasing understanding of the complex cross-talkamong dormant cancer cells, the microenvironment, and sur-rounding hosts cells is teaching us that treatment decisions andapproaches should be based on more than the genetic profile of theprimary tumour alone.

The development of skeletal metastases is perhaps the best-understood example of the so-called “seed and soil hypothesis”first described by Sir Stephen Paget in the 19th century [8]. Bonemetastasis involves complex interactions between cancer cells,osteoblasts and osteoclasts, and both haematopoetic and endo-thelial stem cells within the bone microenvironment. The presence

Fig. 1. Schematic representation of the traditional view of metastasis. Cells degrade the extracellular matrix, intravasate into the vascular space, survive the hostile environment ofthe circulation, adhere to endothelial cells, extravasate into the target organ and develop a metastatic focus. From Mocellin S et al. [43] with permission.

Fig. 2. The tumour microenvironment can be considered at a local, regional or organ confined and distant level. At all three levels there is an ever-increasing complexity withmultiple factors interacting simultaneously in time and space.

R.E. Coleman et al. / The Breast 22 (2013) S50eS56 S51

of tumour in bone ultimately results in activation of osteoclasts,leading to an increased rate of bone resorption. Additionally, bonemarrow derived stem cells are of fundamental importance in thedevelopment of metastases at other sites, preparing the environ-ment for tumour cells to establish a metastasis [9]. Drugs that areable to target bone, notably the bisphosphonates but also deno-sumab, an inhibitor of receptor activator of nuclear factor kappa-bligand (RANKL), a key regulator of bone cell function, provide anadditional strategy to prevent metastasis within bone and, poten-tially, at extra-skeletal sites [10].

Bisphosphonates are potent inhibitors of osteoclastic boneresorption, with proven efficacy in reducing tumour-associatedskeletal complications in advanced cancer [11]. More recently,clinical studies have investigated the adjuvant use of these drugs inbreast cancer, with evaluation of their impact on bone density,metastases, and survival. Additionally, numerous preclinical ex-periments have shown that the development of bone metastasescan be inhibited by either the bisphosphonates or RANKL inhibi-tion, through both bone-mediated and possible direct antitumourmechanisms [10]. Furthermore, synergy between potent amino-bisphosphonates such as zoledronic acid with chemotherapy hasbeen demonstrated in mouse models, although the clinical rele-vance of these observations remains uncertain [12]. The emergingclinical trial results suggest an increasing role for adjuvantbisphosphonates in the treatment of early-stage breast canceralthough, as is discussed below, benefits appear to be confined tothe postmenopausal setting [10].

The bone microenvironment and metastasis

Disseminated tumour cells have been reported in the bonemarrow of 30%e40% of early-stage breast cancer patients at thetime of diagnosis [7]. Most DTC die, but circulating breast cancercells do have a particular affinity for bone. These circulating cellscan adhere to the vessels and sinusoids of the bone marrow, theninvade into themarrowand inter-trabecular spaces with the help ofadhesion molecules, and compete with haematopoietic stem cellsfor the osteoblastic niche adjacent to the bone surface [9]. Tumourcells have also been shown to exhibit chemotactic responses toareas of bone undergoing resorption [13].

Healthy bone is in a constant state of remodelling, a process thatis essential to preserve the structural integrity of bone. Bone-

derived osteoblasts and osteoclasts work together through the in-fluence of cytokines and other humoral factors to couple formationand resorption. Osteoblasts, derived from fibroblast precursors,produce collagen matrix and contribute to bone formation, whilethe multinucleated osteoclasts are cells derived from themacrophage-monocyte lineage, and are themajor mediator of bonedegradation or resorption.

Tumour cells have the capacity to produce a wide range of cy-tokines and growth factors that can promote the production ofRANKL and macrophage colony-stimulating factor (M-CSF) fromosteoblasts, leading to activation of osteoclasts and disturbance ofthe normal balance of new bone formation and bone resorption.These multiple interactions between metastatic tumour cells andthe bone microenvironment may contribute to the development ofmetastases both within and outside bone.

There are several phases to the development of bonemetastasis,with destruction of bone occurring late (Fig. 3). Initially tumourcells colonise the bone surface competing with haematopoieticcells for the osteoblastic niche where they can enter a dormantquiescent state [14]. Subsequently, cells may escape from quies-cence, progress and either activate osteoclasts to establish a bonemetastasis or disseminate elsewhere. Elegant animal studies havedemonstrated that metastasis is multidirectional with dissemina-tion not only from the primary to the metastatic site but also fromone metastatic site to another (eg bone to liver) and even back tothe primary site and organ of origin [15].

Mundy’s group [16] has described a “vicious cycle” that occurswhen active cancer cells are present adjacent to the bone matrixestablishing a bone metastasis. Products produced by the tumourinduce breakdown of bone, causing release of factors into the localenvironment that may cause stimulation and further growth ofmalignant cells, that in turn leads to yet further bone resorption.At the microscopic level, osteoclasts are visible between cancercells and the bone surface that is being destroyed [17]. Osteoclastsare activated by cytokines and growth factors produced directly orindirectly by the tumour cell. As bone matrix is broken down bythe active osteoclasts, a rich supply of mitogenic factors is releasedthat can lead to increased growth and proliferation of the breastcancer metastases [18]. The overall effect is the creation of aself-sustaining vicious cycle with multidirectional interactionsbetween cancer cells, osteoclasts, osteoblasts, and the bonemicroenvironment.

Fig. 3. The development of bone metastases can be considered in several stages: colonisation, quiescence, progression either locally or dissemination to another site and overtmetastasis.

R.E. Coleman et al. / The Breast 22 (2013) S50eS56S52

Bone targeted treatments to prevent metastasis

There are two classes of bone-targeted agents in routine clinicaluse, the bisphosphonates and denosumab, a highly specific inhib-itor of RANK ligand. Although there are promising data withdenosumab in castrate resistant prostate cancer indicating that thisagent can delay bone metastases [19], the trials of denosumab inearly breast cancer have not yet reported and so currently all ourinformation on this therapeutic approach is restricted to the use ofbisphosphonates.

Bisphosphonates are analogs of endogenous pyrophosphate, inwhich a carbon atom replaces the central oxygen atom. They have apowerful affinity for bone, with 40e70% of an intravenous admin-istration binding rapidly to the bone surface, and concentratingpreferentially at sites of increased bone formation or resorption,with the remainder excreted in the urine. The half-life ofbisphosphonates in circulation is less than an hour [20]. However,the half-life in bone is very long with biological effects of potentagents such as zoledronic acid still evident years after administra-tion of a single dose [21]. Once deposited on the bone surface,bisphosphonates are ingested by osteoclasts engaged in boneresorption. These agents interfere with bone resorption by pro-ducing a direct toxic apoptotic effect on osteoclasts, and by inhib-iting their differentiation and maturation [22].

Bisphosphonates fall into two classes, based on the structure ofthe active R2 side chain and the presence or absence of a nitrogen-containing moiety. First-generation bisphosphonates, such asclodronate and etidronate, do not contain nitrogen. These agentssubstitute into the production of adenosine triphosphate, whichthen becomes a toxic adenosine triphosphate analogue that poi-sons the osteoclast. Nitrogen-containing bisphosphonates, such aspamidronate, alendronate, risedronate, ibandronate, and zole-dronic acid, interfere with cell signalling and block the prenylationof small signalling proteins that are essential for osteoclast functionand survival [22]. However, within the family of bisphosphonates,there are more similarities in pharmacologic effects than differ-ences, although side effect profiles, rates of oral absorption, andpotency do differ. Although differences in the molecular mecha-nism of action exist for the two classes of agents, all therapeuticbisphosphonates have a final inhibitory effect on osteoclastfunction.

Bisphosphonates as adjuvant therapy in early-stage breastcancer

Preclinical data have provided the biological basis for a role ofbisphosphonates in inhibiting the development of bone metasta-ses. In vitro studies have consistently demonstrated thatbisphosphonates can inhibit critical steps in development of me-tastases in the bone, including adhesion and invasion [23]. Animaltumour model systems have shown also that bisphosphonates caninhibit the development of bone metastases in vivo, reduce theburden of tumour within bone, and improve survival in miceinjected with human breast cancer cells [12,24,25]. Although mostanimal models suggest that the primary antitumour effect ofbisphosphonates is manifested in the bone, some data indicate alsoan effect of bisphosphonates on extra-skeletal metastases.Bisphosphonates also possess anti-angiogenic, anti-invasive, andimmunomodulatory properties [26,27]. Additionally, nitrogen-containing bisphosphonates can directly induce tumour cellapoptosis and inhibit tumour cell proliferation, and act synergisti-cally with cytotoxic chemotherapy agents commonly used in breastcancer treatment [12,24,28]. The high doses of bisphosphonatesthat have been used in many laboratory studies suggesting directanticancer activities are incompatible with the clinical doses and

schedules approved for the treatment of cancer patients. However,beneficial effects of combining chemotherapy with zoledronic acidon tumour response within the breast to neoadjuvant therapy havebeen suggested [29] while short term changes in growth indexwithin the primary breast tumour and angiogenesis markers inserum have been recently reported in an exploratory biomarkerstudy comparing chemotherapy alone to chemotherapy plus zole-dronic acid [30]. Whether these presumably direct antitumour ef-fects of bisphosphonates play a clinically significant role in thetreatment or prevention of cancer in humans remains unproven.

In advanced breast cancer, clear improvements in survival havenot been seen. This may be because the studies were underpow-ered to detect survival advantages or, more likely, reflect the futilityof modifying the bone microenvironment to influence the largeburden of disease present with overt metastases. However, in theadjuvant setting, where the disease burden is microscopic, andmore likely to be in a dormant sitewithin the osteoblastic niche andpotentially more receptive to cellular changes in the bone marrow,the anticancer effects of bone targeted treatments seen in thepreclinical models are more likely to be mirrored in patients.

Phase II exploratory studies in women with early-stage, high-risk breast cancer have reported that monthly zoledronic acid, incombination with standard anticancer therapy, increases theclearance of DTC and reduces the number and persistence of DTC inbone marrow compared with standard therapy alone [31e33]. Thezoledronic acid-mediated reduction in DTC persistence might beone of the mechanisms underlying the observed clinical benefits inthe adjuvant studies described below. However, further studies areneeded to determine whether the disease modifying benefits seenin some patient subsets correlate with decreases in DTC levels orwhether the combined treatment in some patients is simplyencouraging the DTC to leave the bone microenvironment anddevelop metastasis elsewhere.

In patients with early-stage breast cancer, inconsistent effectson disease outcomes had been reported in the early clinical trials ofadjuvant bisphosphonate treatment [10]. However, results from theABCSG-12 trial showed that zoledronic acid treatment every 6months for 3 years significantly improved disease-free survival(DFS) in pre-menopausal women with endocrine-sensitive early-

0 .5 1 1.5 2 2.5 3 3.5 4

DIFFERENCE IN HAZARD RATIO EFFECT BETWEEN MENOPAUSAL GROUPS

COMPONENT GROUP

Loco-regional

recurrence

New 2nd primaries

Skeletal distant

recurrence

Non-skeletal distant

recurrence

All IDFS except

skeletal distant

recurrence

All IDFS events

Fig. 4. Interaction forest plot displaying hazard ratio differences for zoledronic acidversus control between postmenopausal women and those who were pre, peri orunknown menopausal status by IDFS endpoint component. For instance, a HR for allIDFS excluding skeletal distant recurrence in postmenopausal women of 0.70 and inthe pre, peri and unknown group of 1.32 combine to give a HR difference of 1.32/0.70 ! 1.87 (HRs are log distributed, so a difference equates to dividing the HRs)(reference [36] on line supplement).

R.E. Coleman et al. / The Breast 22 (2013) S50eS56 S53

stage breast cancer treated with ovarian suppression therapy(goserelin) and tamoxifen or anastrozole [34]. Updated results fromthis trial at a median of 84 months follow-up [35] showed that theimprovement in DFS with adjuvant zoledronic acid treatment wasmaintained (Hazard ratio (HR) ! 0.71, 95% confidence intervals(CI)! 0.55 to 0.92, P! .011) and that overall survival was improved.

Improvements in DFS and overall survival were also reported inthe cohort of patients with established menopause (>5 years sincelast menstrual period) treated in the AZURE trial [36]. In this study,3360 patients with stage II or III breast cancer (unselected bymenopausal status or oestrogen receptor expression) wererandomly assigned to receive standard adjuvant systemic therapywith or without zoledronic acid every 3e4weeks for six doses, thenevery 3e6 months thereafter, for a total of 5 years. After a medianfollow-up of 59months, no significant differencewas seen in DFS oroverall survival in the study population as a whole. However, inpre-specified subgroup analyses, a significant improvement in DFS

was seen with zoledronic acid in the 1041 with establishedmenopause. 5-year invasive DFS was 71% in the control arm and78.2% in those treated with zoledronic acid (adjusted HR 0.75; 95%CI 0.59e0.96, P ! 0.02). In addition, zoledronic acid seemed toimprove overall survival in this group of women, with 5-yearoverall survival rates of 79% in the control group and 85% in pa-tients treated with zoledronic acid (adjusted HR 0.74; 95% CI 0.55e0.98, P ! 0.04). Treatment effects were not influenced by ER statusor tumour grade.

The impact of menopause as a treatment effect modifier wasstrongly linked to the site of first recurrence. For skeletal recurrencethe effects of zoledronic acid did not differ significantly by meno-pausal groups (heterogeneity test c12 ! 0.14, P ! 0.70), with a smallnon-significant benefit in favour of zoledronic acid. However, forthe other components of recurrence there was a statistically sig-nificant large difference in treatment effect according to meno-pausal status, with benefit in postmenopausal women and harm inall other women (heterogeneity test c12 ! 14.00, P ! <.001). This isshown in Fig. 4.

Taken together, the data from both ABCSG-12 and AZURE sug-gest that the beneficial effects of adjuvant zoledronic acid treat-ment are dependent on an environment with low reproductivehormone levels. The mechanisms underlying this observation areunder investigation. Similar beneficial effects in postmenopausalpatients have been seen in several other trials of bisphosphonates,and a consistent pattern of benefit seems to be emerging [36e38]as suggested by a recent meta-analysis of published studies (seeFig. 5) [39].

The results from NSABP-B34 [38] and the previous clodronatestudy by Powles et al. [40] also suggest that the beneficial effects ofbisphosphonates in solid tumours may be less dependent on thetype of agent chosen and more on the hormonal status of the pa-tient. There is an ongoing clinical trial (NCT00127205) comparingzoledronic acid, ibandronate and clodronate. This trial has recentlycompleted accrual and will address the relative efficacy of 3 yearstreatment with either oral clodronate (1600 mg daily) versusoral ibandronate (50 mg daily) versus zoledronic acid (4 mg

0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2

Study Odds Reduction (+/- S.D.)

AZURE: >5 YEARS POSTMENOPAUSE

ABCSG-12*

GAIN: POSTMENOPAUSAL

NSABP B-34: AGE 50

ZO-FAST

Z-FAST

E-ZO-FAST

TOTAL: -18% +/- 5

Z = -3.37, P = .00075

2

6

(heterogeneity) = 8.46 P = .21

ODDS RATIO

2

6

* Induced menopause

Fig. 5. Meta-analysis of large published studies of adjuvant bisphosphonates in earlybreast cancer in which outcomes by menopausal status or age can be determined.Adapted from a poster presentation by Gregory et al. at ASCO 2012 [44].

Fig. 6. Schematic illustration of the cellular interactions in the different pre-metastatic niches in the pre- (A) and postmenopausal (B) bone narrow microenvironment. (A) The highlocal concentration of inhibin, oestrogen and progesterone act to decrease the tumour suppressors Smad2/3, through direct tumour cell receptor inhibition, and inhibition ofosteoclast bone resorption mediated release of activin/TGFb. (B) The loss of inhibin, oestrogen and progesterone leads to higher local concentrations of activin and TGFb, andsubsequently increase in tumour suppressors Smad2/3. The size of lettering denotes local concentration). From Wilson et al. [41].

R.E. Coleman et al. / The Breast 22 (2013) S50eS56S54

intravenously monthly for 6 months, then every 3 months) in bothpre and postmenopausal women.

The interactions between reproductive hormones and the TGFbfamily of growth factors may provide at least part of the explana-tion for the differences seen according to ovarian function in theresponse to bisphosphonates. In pre-menopausal women, thecycling endocrine effect of oestrogen, progesterone and inhibin hasthe potential to inhibit activin and TGFß signalling. Inhibin binds tothe activin type II receptor, blocks activin function and therebyreduces the activity of members of the SMAD family of tumoursuppressor proteins that are known to have important effects onboth bone and cancer cell functions. At menopause, serum oes-tradiol, inhibin and follistatin decline over several years and folliclestimulating hormone (FSH) levels rise. In the absence of inhibin, theTGFß ligand, activin becomes the dominant signalling molecule inbone. Both activin and TGFb will support homing of disseminatedtumour cells to bone by increasing the chemokine CXCL4, andstimulating osteoclastic bone resorption leading to greater avail-ability of bone-derived growth factors. In a postmenopausalwoman, the loss of inhibin, oestrogen and progesterone will reduceinhibition of the TGFß superfamily, and thus the relative biologicalactivity of activin and TGFß will increase [42,43] (Fig. 6).

Quite how the variable effects of reproductive hormones on thebone microenvironment explain the differential effects on extra-skeletal recurrence according to menopausal status remains un-clear. However, the recently demonstrated ability of breast cancercells to reseed from bone to other distant sites and back to thebreast [15], perhaps suggests that the bone microenvironment isthe key co-ordinator in themetastatic process, determining the fateof cancer cells not only within bone but also at other distant sites.Treatments that modify this environment have the potential to doharm as well as good.

Conclusions

Bisphosphonates are a promising group of compounds in theadjuvant breast cancer setting. Preclinical studies provide goodproof of principle for the role of bisphosphonates in preventing thegrowth and development of bone metastases. There is alsoincreasing evidence of clinically relevant anticancer response tobisphosphonates in disease outside the skeleton. Recent adjuvantstudies in breast cancer patients have shown efficacy in a lowoestrogen environment (postmenopausal women, and pre-menopausal women treated with GnRH agonists), with improve-ments in both local and distant DFS. An individual patient meta-analysis conducted by the Early Breast Cancer Clinical TrialsCollaborative Group is underway. If this confirms the findings re-ported within individual trials of adjuvant bisphosphonates inpostmenopausal women, a new treatment approach shouldbecome available.

Conflict of interest statement

Robert Coleman has received consultancy and speaker fees fromAmgen and Bayer and given expert testimony on behalf of Novartis.

Walter Gregory, Helen Marshall, Caroline Wilson and IngunnHolen have no relevant disclosures to make.

References

[1] Taylor J, Hickson J, Lotan T, Yamada DS, Rinker-Schaeffer C. Using metastasissuppressor proteins to dissect interactions among cancer cells and theirmicroenvironment. Cancer Metastasis Rev 2008;27:67e73.

[2] Horak CE, Lee JH, Marshall JC, Shreeve SM, Steeg PS. The role of metastasissuppressor genes in metastatic dormancy. APMIS 2008;116:586e601.

[3] Hu M, Polyak K. Microenvironmental regulation of cancer development. CurrOpin Genet Dev 2008;18:27e34.

[4] Massague J. New concepts in tissue-specific metastases. Clin Adv HematolOncol 2003;1:576e7.

[5] Páez D, Labonte MJ, Bohanes P, Zhang W, Benhanim L, Ning Y, et al. Cancerdormancy: a model of early dissemination and late cancer recurrence. ClinCancer Res 2012;18:645e53.

[6] Demicheli R, Terenziani M, Bonadonna G. Estimate of tumor growth time forbreast cancer local recurrences: rapid growth after wake-up? Breast CancerRes Treat 1998;51:133e7.

[7] Braun S, Vogl FD, Naume B, Janni W, Osborne MP, Coombes RC, et al. A pooledanalysis of bone marrow micrometastasis in breast cancer. N Engl J Med2005;353:793e802.

[8] Paget S. The distribution of secondary growths in cancer of the breast. Lancet1889;8:98e101.

[9] Martínez AS, Huelsken J. The niche under siege: novel targets for metastasistherapy. J Intern Med 2012 Dec 18. http://dx.doi.org/10.1111/joim.12024.[Epub ahead of print].

[10] Coleman R, Gnant M, Morgan G, Clezardin P. Effects of bone-targeted agentson cancer progression and mortality. J Natl Cancer Inst 2012;104:1059e67.

[11] Coleman RE, McCloskey EV. Bisphosphonates in oncology. Bone 2011;49:71e6.

[12] Ottewell PD, Monkkonen H, Jones M, Lefley DV, Coleman RE, Holen I. Anti-tumor effects of doxorubicin followed by zoledronic acid in a mouse model ofbreast cancer. J Natl Cancer Inst 2008;100:1167e78.

[13] Saidak Z, Boudot C, Abdoune R, Petit L, Brazier M, Mentaverri R, et al.Extracellular calcium promotes the migration of breast cancer cells throughthe activation of the calcium sensing receptor. Exp Cell Res 2009;315:2072e80.

[14] Comen EA. Tracking the seed and tending the soil: evolving concepts inmetastatic breast cancer. Discov Med 2012;14:97e104.

[15] Kim MY, Oskarsson T, Acharyya S, et al. Tumour self-seeding by circulatingcancer cells. Cell 2009;139:1315e26.

[16] Käkönen SM, Mundy GR. Mechanisms of osteolytic bone metastases in breastcarcinoma. Cancer 2003;97:834e9.

[17] Shimamura T, Amizuka N, Li M, Freitas PH, White JH, Henderson JE, et al.Histological observations on the microenvironment of osteolytic bonemetastasis by breast carcinoma cell line. Biomed Res 2005;26:159e79.

[18] Weilbaecher KN, Guise TA, McCauley LK. Cancer to bone: a fatal attraction. NatRev Cancer 2011;11:411e25.

[19] Smith MR, Saad F, Coleman R, Shore N, Fizazi K, Tombal B, et al. Denosumaband bone metastasis-free survival in men with castration-resistant prostatecancer: results of a global phase 3, randomised, placebo-controlled trial.Lancet 2012;379:39e46.

[20] Chen T, Berenson J, Vescio R, Swift R, Gilchick A, Goodin S, et al. Pharmaco-kinetics and pharmacodynamics of zoledronic acid in cancer patients withbone metastases. J Clin Pharmacol 2002;42:1228e36.

[21] Brown JE, Ellis SP, Lester JE, Gutcher S, Khanna T, Purohit OP, et al. Prolongedefficacy of a single dose of the bisphosphonate zoledronic acid. Clin Cancer Res2007;13:5406e10.

[22] Rogers MJ, Crockett JC, Coxon FP, Mönkkönen J. Biochemical and molecularmechanisms of action of bisphosphonates. Bone 2011;49:34e41.

[23] Bosch-Barrera J, Merajver SD, Menéndez JA, Van Poznak C. Direct antitumouractivity of zoledronic acid: preclinical and clinical data. Clin Transl Oncol2011;13:148e55.

[24] Ottewell PD, Lefley DV, Cross SS, Evans CA, Coleman RE, Holen I. Sustainedinhibition of tumor growth and prolonged survival following sequentialadministration of doxorubicin and zoledronic acid in a breast cancer model.Int J Cancer 2010;126:522e32.

[25] Holen I, Coleman RE. Anti-tumour activity of bisphosphonates in preclinicalmodels of breast cancer. Breast Cancer Res 2010;12:214e26.

[26] Stresing V, Fournier PG, Bellahcène A, Benzaïd I, Mönkkönen H, Colombel M,et al. Nitrogen-containing bisphosphonates can inhibit angiogenesis in vivowithout the involvement of farnesyl pyrophosphate synthase. Bone 2011;48:259e66.

[27] Rogers TL, Holen I. Tumour macrophages as potential targets of bisphospho-nates. J Transl Med 2011;9:177.

[28] Daubiné F, Le Gall C, Gasser J, Green J, Clézardin P. Antitumor effects of clinicaldosing regimens of bisphosphonates in experimental breast cancer bonemetastasis. J Natl Cancer Inst 2007;99:322e30.

[29] Coleman RE, Winter MC, Cameron D, Bell R, Dodwell D, Keane MM, et al. Theeffects of adding zoledronic acid to neoadjuvant chemotherapy on tumourresponse: exploratory evidence for direct anti-tumour activity in breastcancer. Br J Cancer 2010;102:1099e105.

[30] Winter MC, Wilson C, Syddall SP, Cross SS, Evans A, Ingram CE, et al.Neoadjuvant chemotherapy with or without zoledronic acid in early breastcancer e a randomised pilot study evaluating biomarkers. Clin Cancer Res2013;19:2755e65.

[31] Aft R, Naughton M, Trinkaus K, Watson M, Ylagan L, Chavez-MacGregor M,et al. Effect of zoledronic acid on disseminated tumour cells in women withlocally advanced breast cancer: an open label, randomised, phase 2 trial.Lancet Oncol 2010;11:421e8.

[32] Rack B, Jückstock J, Genss EM, Schoberth A, Schindlbeck C, Strobl B. Effect ofzoledronic acid on persisting isolated tumour cells in patients with earlybreast cancer. Anticancer Res 2010;30:1807e13.

R.E. Coleman et al. / The Breast 22 (2013) S50eS56 S55

[33] Solomayer EF, Gebauer G, Hirnle P, Janni W, Lück HJ, Becker S, et al. Influenceof zoledronic acid on disseminated tumour cells in primary breast cancerpatients. Ann Oncol 2012;23:2271e7.

[34] Gnant M, Mlineritsch B, Schippinger W, Luschin-Ebengreuth G, Pöstlberger S,Menzel C, et al. Endocrine therapy plus zoledronic acid in premenopausalbreast cancer. N Engl J Med 2009;360:679e91.

[35] Gnant M, Mlineritsch B, Luschin-Ebengreuth G, Stoeger H, Dubsky P, Jakesz R,et al. Long-term follow-up in ABCSG-12: significantly improved overall sur-vival with adjuvant zoledronic acid in premenopausal patients withendocrine-receptor-positive early breast cancer. Cancer Res2011;71(24Suppl.) [absS1-2,95s].

[36] Coleman RE, Marshall H, Cameron D, Dodwell D, Burkinshaw R, Keane M, et al.Breast-cancer adjuvant therapy with zoledronic acid. N Engl J Med 2011;365:1396e405.

[37] Coleman R, de Boer R, Eidtmann H, Llombart A, Davidson N, Neven P,et al. Zoledronic acid (zoledronate) for postmenopausal women with earlybreast cancer receiving adjuvant letrozole (ZO-FAST study): final 60-month results. Ann Oncol 2013;24:398e405.

[38] Paterson AH, Anderson SJ, Lembersky BC, Fehrenbacher L, Falkson CI, King KM,et al. Oral clodronate for adjuvant treatment of operable breast cancer (Na-tional Surgical Adjuvant Breast and Bowel Project protocol B-34): a multi-centre, placebo-controlled, randomised trial. Lancet Oncol 2012;13:734e42.

[39] Mobus V, Diel IJ, Elling D, Harbeck N, Jackisch C, Thomssen C, et al. GAIN(German Adjuvant Intergroup Node Positive) Study: a phase-iii multicentertrial to compare dose dense, dose intense ETC (iddETC) vs. EC-TX andibandronate vs. observation in patients with node-positive primary breastcancer e 1st interim efficacy analysis. Cancer Res 2011;71(24 Suppl.) [absS2-4,100s].

[40] Powles T, Paterson A, McCloskey E, Schein P, Scheffler B, Tidy A. Reduction inbone relapse and improved survival with oral clodronate for adjuvant treat-ment of operable breast cancer. Breast Cancer Res 2006;8. R13.

[41] Wilson C, Holen I, Coleman RE. Seed, soil and secreted hormones: potentialinteractions of breast cancer cells with their endocrine/paracrine microenvi-ronment and implications for treatment with bisphosphonates. Cancer TreatRev 2012;38:877e89.

[42] Hadji P, Coleman R, Gnant M, Green J. The impact of menopause on bone,zoledronic acid, and implications for breast cancer growth and metastasis.Ann Oncol 2012;23:2782e90.

[43] Mocellin S, Keilholz U, Rossi CR, Nitti D. Circulating tumor cells: the ‘leukemicphase’ of solid cancers. Trends Mol Med 2006;12:130e9.

[44] Gregory W, Marshall H, Bell R, Cameron DA, Coleman RE. Adjuvant zoledronicacid (ZOL) in postmenopausal women with breast cancer and those renderedpostmenopausal: results of a meta-analysis. J Clin Oncol 2012;30(suppl.). [Abs513].

R.E. Coleman et al. / The Breast 22 (2013) S50eS56S56