Morphologie (2015) 99, 63—72

Disponible en ligne sur

ScienceDirectwww.sciencedirect.com

GENERAL REVIEW

The use of animal models in multiplemyelomaL’utilisation des modèles animaux dans le myélome multiple

H. Libouban

GEROM groupe études remodelage osseux et biomatériaux—LHEA, IRIS-IBS institut de biologie en santé,LUNAM université, CHU d’Angers, 49933 Angers cedex, France

Available online 17 April 2015

KEYWORDSMurine model ofmyeloma;Osteolysis;BALB/c;SCID;SCID-hu;5T2MM;5T33MM;5TGM1;C57BL/KalwRij;Xenograft;Pristane

Summary In myeloma, the understanding of the tissular, cellular and molecular mechanismsof the interactions between tumor plasma cells and bone cells have progressed from in vitroand in vivo studies. However none of the known animal models of myeloma reproduce exactlythe human form of the disease. There are currently three types of animal models: (1) injectionof pristane oil in BALB/c mice leads to intraperitoneal plasmacytomas but without bone marrowcolonization and osteolysis; (2) injection of malignant plasma cell lines in immunodeficient miceSCID or NOD/SCID; the use of the SCID-hu or SCID-rab model allows the use of fresh plasma cellsobtained from MM patients; (3) injection of allogeneic malignant plasma cells (5T2MM, 5T33)in the C57BL/KalwRij mouse induces bone marrow proliferation and osteolytic lesions. Thesecells did not grow in vitro and can be propagated by injection of plasma cells isolated frombone marrow of a mouse at end stage of the disease into young recipient mice. The 5TGM1 is asubclone of 5T33MM cells and can grow in vitro. Among the different models, the 5TMM modelsand SCID-hu/SCID-rab models were extensively used to test pathophysiological hypotheses andto assess anti-osteoclastic, anti-osteoblastic or anti-tumor therapies in myeloma. In the presentreview, we report the different types of animal models of MM and describe their interests andlimitations.© 2015 Elsevier Masson SAS. All rights reserved.

MOTS CLÉSRésumé Dans le myélome, la compréhension des mécanismes tissulaires, cellulaires etmoléculaires des interactions entre les plasmocytes tumoraux et les cellules osseuses a pro-

Modèle murin demyélome ;Ostéolyse ;BALB/c ;SCID ;

gressé à partir des études in vitro et in vivo. Cependant, aucun des modèles animaux de myélomene reproduit exactement le myélome humain. Il existe actuellement trois types de modèlesanimaux : (1) l’injection d’huile pristane chez des souris BALB/c conduit à des plasmacytomesintrapéritonéaux mais sans qu’il y’ait colonisation de la moelle osseuse et développement d’uneostéolyse ; (2) l’injection de lignées de plasmocytes malins chez des souris immunodéficientes

E-mail address: helene.libouban@univ-angers.fr

http://dx.doi.org/10.1016/j.morpho.2015.01.0031286-0115/© 2015 Elsevier Masson SAS. All rights reserved.

64 H. Libouban

SCID-hu ;5T2MM ;5T33MM ;5TGM1 ;C57BL/KalwRij ;Xénogreffe ;Pristane

SCID ou NOD/SCID ; l’utilisation du modèle de souris SCID-hu ou SCID-rab permet d’injecter desplasmocytes tumoraux frais obtenus à partir de patients atteints de myélome ; (3) l’injectionde cellules plasmocytaires allogéniques (5T2MM, 5T33) chez la souris C57BL/KalwRij induit uneprolifération tumorale dans la moelle osseuse et des lésions ostéolytiques. Ces cellules ne secultivent pas in vitro et peuvent être propagées par injection de plasmocytes isolés, à par-tir de moelle osseuse de souris au stade terminal de la maladie, dans des souris receveusesjeunes. La lignée 5TGM1 est un sous-clone des cellules 5T33MM et peut se développer in vitro.Parmi les différents modèles de myélome, les modèles 5TMM et les modèles SCID-hu/SCID-rabont été largement utilisés pour tester des hypothèses physiopathologiques et évaluer l’effetde thérapies anti-ostéoclastiques, de thérapies anti-ostéoblastiques ou des thérapies anti-tumorales du myélome. Cette revue de la littérature décrit les différents types de modèlesexistants avec leurs intérêts et leurs limitations.© 2015 Elsevier Masson SAS. Tous droits réservés.

I

MlatTaiaoianmcOcowalt

ebf�iafaaIibstoibrc

ma

aIb([g

uUasM

•

•

mrM

Iw

ToqhPp[psti

ntroduction

yeloma (MM) is a hematological malignancy whose preva-ence is 1 in 10,000 people; the median survival ispproximately three years. MM is due to the prolifera-ion of malignant plasma cells (PC) in the bone marrow.he overt disease is clinically characterized by bone painnd asthenia. Biologically, a monoclonal protein (M-protein)s found in blood and/or urine in 98% of cases. Skeletalbnormalities characterized by osteolytic foci and diffusesteolysis is observed on radiographs in 90% of patients dur-ng the time course of the disease. Osteolysis is due ton increased osteoclastogenesis induced by the tumor andot to the malignant PCs, which do not have the cytologicachinery for resorbing the calcified bone matrix. Hyper-

alcemia, which reflects bone lysis, is often associated.steolysis causes serious clinical problems (fractures, spinalord compression, bone pain. . .). Pathophysiology and etiol-gy of osteolysis in myeloma are still imperfectly known. Itas shown that the destruction of the bone tissue was notssociated with tumor PCs themselves, but by the stimu-ation of osteoclast activity which occurs in the vicinity ofumor nodules.

Osteolytic cavities are generated in response to a vari-ty of local factors produced by the malignant PCs andy the bone marrow microenvironment. These factorsavoring osteolysis include tumor necrosis factor alpha (TNF-), interleukin-1� (IL-1�), interleukin-6 (IL-6), macrophage

nflammatory protein-1-alpha and beta (MIP-1� and MIP-1�)nd ligand for receptor activator of nuclear transcriptionactor-�B (RANKL) [1]. However, none of these factorsppears to be solely responsible for bone destruction. Inddition, the implication of several of these factors (IL-6,L-1, TNF-�) in osteolysis has been shown in vitro but theirmplication in vivo is uncertain [2]. In MM, interaction withone marrow microenvironment is essential for MM progres-ion. The interaction of stromal cells with tumor PCs throughhe adhesion molecules VCAM-1/�4-ß1 increases osteoclast-genesis and resorption activity [3]. The VCAM-1/�4-ß1nteraction is also involved in RANK-L/OPG and IL-6 released

y stromal cells [4,5]. IL-6 association with its solubleeceptor promotes the proliferation and survival of tumorells. Additional growth factors are also released by the

tdo

icroenvironment that can promote the tumor growth suchs IGF-1 [6].

In overt MM, a decrease in bone formation is alsossociated, leading to an uncoupling in bone remodeling.nhibition of osteoblastogenesis is due to inhibitors releasedy PCs and depressing the Wnt-signaling pathway: DKK1Dickkopf-1) and Sfrp2 (secreted frizzled-related protein 2)7,8] or acting on the osteoblastic precursors: hepatocyterowth factor (HGF), IL-3 and IL-7 [9—11].

In MM, there is a vicious circle in which tumor PCs stim-late bone cells which in turn stimulate the tumor growth.nderstanding of the tissular, cellular and molecular mech-nisms of interactions between tumor PCs and bone cellsuffers from the low number of available animal models.odels are interesting:

to evaluate pathophysiology of the disease and its effecton bone remodeling;

to test the effect of therapeutic with an anti-osteoblastic,anti-osteoblastic or anti-tumor activity.

At present, only mouse models are available and reflectore or less perfectly the human disease. In the present

eview, we report the different types of animal models ofM and describe their interests and limitations.

nduction of plasmacytomas in BALB/c miceith pristine oil

he injection of the mineral oil pristane into the peritoneumf BALB/c mice induces plasmacytomas with a high fre-uency [12]. This model has been discovered in 1969 andas constituted at that time the first one to study MM.ristane oil induces a chronic inflammatory tissue wherelasmacytomas develop with a 60% frequency after 16 weeks12]. Plasmacytomas can then be transferred to other miceretreated with pristane. Plasmacytomas predominantlyecrete a monoclonal IgA in 60% of cases. This is a limita-ion of the model because IgG MM is the most frequent formn humans [13]. The main disadvantages of the model are

hat PCs are restrictively localized in the peritoneum ando not extend to the bone marrow; so bone lesions are notbserved. This model does not faithfully reproduce human

taR

tvobnote

lmmteccncmfoirodds

ltt(hc[boimmpib

T

T1c[tMf

The use of animal models in multiple myeloma

MM and thus cannot be used to study the pathophysiologyof MM-induced bone lesions. However, it is well adapted tostudy the gene expression profile in PCs during tumor pro-gression [14,15]. Another interest of the pristane model isthe possibility to study of the oncogenic rearrangements ofc-myc [16,17]. Beside to cytogenetic studies, the role of IL-6 has also been investigated in this model. IL-6 is now wellknown to be involved in survival and proliferation of PCs inMM [18]. It has been shown that BALB/c mice deficient forthe IL-6 gene failed to develop plasmacytoma after pristaneinjection [19,20].

Experimental therapies on the plasmacytoma model havebeen tested but they were focused only on the effect oftumor growth [12]. One interesting study was the use ofzoledronic acid (ZOL, the most potent bisphosphonate ofthe 3rd generation) currently used to prevent bone lesions inhuman MM. It is well known that ZOL is effective in reducingand delaying skeletal events in MM patients [21]. The directeffect of ZOL on PCs was less evident in vivo. A reduction ofplasmacytoma development and an increase in survival havebeen observed in BALB/c mice, which have received ZOL onthe 1st day after pristane injection [22].

Immuno-deficient mice: the SCID model

It consists in the injection of human PCs in severe combinedimmunodeficiency (SCID) mice, so it represents a xenograftmodel. However, the choice of the plasma human tumorline appears critical because of they can be of myeloma-tous or lymphoblastoid origin and their Epstein-Barr virusstatus may differ (EBV- or EBV+). The malignant human cellscolonize the bone marrow microenvironment and develop inthe marrow cavities. Osteolytic lesions occur and are simi-lar to those described in humans; in addition, a M-protein issecreted and can be dosed in the serum [23—27].

Some of these myeloma lines, as the KPMM2, are wellcharacterized [28]. KPMM2 is EBV- originated from a MMpatient. These cells are strongly dependent to IL-6 as inhuman MM. In addition, their proliferation is induced by IL-6 via an autocrine mechanism. Intravenous injection of thecells in SCID mice induces a major development within thehematopoietic bone marrow and lymph nodes [27]. Oste-olytic lesions are observed 30—40 days after inoculation andmice exhibit hind limb paralysis at the end course of thedisease. In the KPMM2 model, injection of anti-human IL-6receptor significantly reduces the development of the dis-ease. Other studies have been done with human PCs from aleukemia line transformed by the Epstein-Barr virus (ARH77)[23]. SCID mice develop paralysis of the lower limbs 28 to35 days after inoculation of ARH77 cells. Hypercalcemia isobserved within 5 days after the onset of paralysis. Histolog-ical examination of various organs shows a PC infiltration ofthe liver, spleen, spine and long bones. A significant increaseis observed in MIP-1� level in the bone marrow environment.In contrast, serum or medullar level of local factors such asIL-6, IL-1, TNF TGF-�, TNF-�, lymphotoxin and PTHrp werenot significantly increased, although ARH77 are able to pro-

duced IL-6 in vitro. Cell lines JJN3 or KPMM2 have not beenoften used as an animal model for MM. In contrast, ARH77,despite the disadvantage of the EBV+, status remained themost widely used cell line in the SCID model. Several types of

bvop

65

reatment were tested in the ARH77 model like ibandronate,n antisense construct against MIP-1� and the recombinantANK antagonist molecule RANK-Fc [29—31].

The route of inoculation has also an influence on theumor development in the SCID model [32]. The intra-enous route is preferred and is associated with typicalsteolytic lesions. Intraperitoneal injection is accompaniedy locations in the pancreas, spleen, liver, intestine, kid-ey but bone marrow is rarely invaded. The intra-cardiacr intra-osseous combines injection direct intramedullaryumor growth associated with osteolytic lesions but alsoxtramedullary proliferations.

Most of the SCID models which were described in theiterature used MM cell lines. The transplant of fresh pri-ary cells induced growth in SCID mice and production ofonoclonal immunoglobulin but without colonization into

he bone marrow [33,34]. Because of the difficulty to gen-rate bone marrow colonization of fresh primary human MMells using conventional SCID mice, xenografts of fresh MMells were tested in a highly immunodeficient mouse; theon-obese diabetic SCID mice (NOD SCID). These mice areharacterized by the absence of NK cells. Therefore, theajority of injection studies with xenogeneic PCs is per-

ormed in NOD SCID mice since year 2000 [26]. Injectionf PCs from blood of MM patients using the intracardiac orntra-osseous route, leads to colonization into bone mar-ow and osteolytic lesions [26]. The time development tobserve symptoms was fluctuant between one to five monthsepending of the PCs origin. Bones lesions are observed inifferent skeletal sites: long bones, vertebrae and also thekull.

The SCID model can be refined by grafting fragments ofiving human fetal long bone in the recipient animals, wherehey constitute a ‘‘marrow micro-environment’’ favoringhe development of transplanted human PCs. This modelreferred as SCID-hu) is associated with a homing of theuman PCs within the fetal bone graft. Primary human MMells can also be directly injected into the grafted bone35]. Osteolytic activity develops only in the xenogeneicony graft. Several studies on this model were focusedn MM therapy such as anti-angiogenic drug thalidomide,nhibitor of DKK1 and the use of RANK-FC [36—38]. All treat-ents decrease osteolysis and MM progression. However, thisodel is cumbersome to implement and may raise ethicalroblems. The authors have also developed a similar modeln which SCID mice are grafted with a piece of rabbit fetalone (referred as SCID-rab) grafted similarly [39].

he 5TMM model in the C57BL/KalwRij mouse

he C57BL/KalwRij mouse was described by Radl et al. in978. With age, mice of this strain develop several mono-lonal B-cell proliferative disorders with a frequency of 80%40]. When mice are older than two years, 0.5% develops aypical MM or a Waldenström’s disease. Several spontaneousM have been isolated and are called the 5TMM series. Dif-

erent 5TMM lines were obtained after transplantation of

one marrow cells into syngeneic young mice by the intra-enous route (Fig. 1). Eight 5TMM were isolated and mostf them were characterized in detail: isotype of the M-rotein, bone marrow involvement, presence of osteolytic

66

Figure 1 Protocol for generation and amplification of the5TMM models. PCs were isolated from bone marrow of 5TMMmice at end stage of the disease. Bone marrow is flushed fromlong bone and PCs are isolated after a centrifugation gradi-ent. Isolated PCs are intravenously injected into the tail veinof young healthy mice. Development of the MM is monitoredby electrophoresis of the serum. In MM positive mice, the M-protein is observed on the electrophoresis gel whereas theelectrophoresis profile of control mice did not show M-protein.Protocole pour la production et l’amplification des modèles5TMM. Les plasmocytes sont isolés à partir de la moelle osseusede souris 5TMM au stade terminal de la maladie. Un flush de lamoelle osseuse est effectué à partir des os longs et les plasmo-cytes sont isolés en réalisant un gradient de centrifugation. Lesplasmocytes isolés sont injectés par voie intraveineuse dans laveine caudale de souris saines et jeunes. Le développement duMM est contrôlé par électrophorèse du sérum. Chez les sourispositives pour le MM, la protéine monoclonale est observée surld

lama(at

gtsaOw(iswemwte5ais5uli5tdsn

mtte�cile

5rbtttcomsep(tcliib

e gel d’électrophorèse tandis que le profil d’électrophorèsee souris contrôles ne montre pas de M-protéine.

esions and growth pattern [41]. Seven cell lines producen IgG immunoglobulin with a light chain. The circulatingonoclonal M-protein reflects the extent of the tumor load

nd can be easily monitored by electrophoresis of the serumFig. 1). Two lines have been particularly studied: the 5T2MMnd 5T33MM. The 5T2MM represents a model situation ofhe most common forms of the human MM with a moderate

st

t

H. Libouban

rowth and osteolytic lesions. In addition and in contrasto the human disease, osteolysis represents the unique con-equence of the inoculation of 5T2MM cells because of thebsence of renal lesions due to light chain deposition [42].n the other hand, the 5T33MM is a very aggressive lineith a rapid tumor growth and involvement of other organs

non-hematopoietic), a condition, which is rarely observedn humans. Kinetic developments of these two models aretrongly different: 5T2MM cells can be detected at nineeeks after inoculation whereas 5T33 can be observed asarly as two weeks after inoculation [43]. In the 5T2MModel, osteolytic lesions can be observed from the 11theek and the end-stage occurs at 16-week post-injection of

he PCs. In the 5T33MM model, the terminal stage of the dis-ase is obtained 4—5 weeks after inoculation. A subclone ofT33MM cells, which can grow in vitro, has been establishednd named the 5TGM1 cells [41,44]. When it is inoculatedntravenously to young syngeneic mice, it reproduces theame pathological characteristics than those observed in theT33MM. More recently, both 5TGM1 and 5T33MM cell weresed to generate fluorescent cells allowing visualization andocalization of MM cells in whole body by fluorescence imag-ng [45,46]. Since the discovery of the 5TGM1subclone, theT33MM has been used less frequently. Moreover, despitehe capacity of 5TGM1 to grow in vitro, the 5TGM1 modelid not replace the 5T2MM model, which is remaining exten-ively used to study pathophysiology of MM and to evaluateew therapeutics.

Several studies have shown that 5T2MM and 5T33MModels have similar pathophysiological characteristics to

hose of human MM: the cytokine network involved andhe adhesion molecules are the same. 5T2MM and 5T33MMxpress the following adhesion molecules: CD44, �4/�1 and5/�1 [43,47,48]. Moreover in in vitro experiments, 5TGM1ells establish direct contacts with stromal cells via thenteraction of �4/�1 and VCAM-1. This finding is particu-arly interesting as interaction of PCs with stromal cells isssential for the tumor growth [2].

Osteolytic lesions have been well characterized in theT2MM mice [43,49,50]. X-ray assessment of bone lesionsevealed that lesions are preferentially localized in longones (proximal tibia, distal femur and humerus). Inhe mouse, these zones have physiologically a high boneurnover, this leads to a preferential localization of theumor and secondarily promotes the tumor growth as severalytokines are mitogenic factors for PCs. The development ofsteolytic lesions is well-studied using digital X-ray or mam-ography films [49,50]. Bone lesions appear as numerous

mall and round cavities that develop from the endostealnvelope. Lesions are preferentially localized in the meta-hyses and also systematically observed at the tibial crestFig. 2). Lesions can be observed in the acetabulum and inhe iliac crest when using an aggressive subclone of 5T2MMells characterized by a faster development rate [51]. Boneesions in the 5T2MM are therefore similar to those observedn human with one exception: the absence of bone lesionsn the skull. However, the tibial crest of the mouse has aony architecture, which mimics the human bones of the

kull where the diploe (a kind of cancellous bone with largerabeculae) is limited by two thick cortices.

Histological examination of invaded bone shows thathe marrow is replaced by tumor nodules at the terminal

The use of animal models in multiple myeloma 67

Figure 2 X-ray image of the pelvis and hind legs of aC57BL/KalwRij mouse injected with 5T2MM cells at end stage ofthe disease. Note the existence of numerous osteolytic lesionson both distal femur and proximal tibia.Radiographie du bassin et des membres postérieurs d’unesouris C57BL/KalwRij injectées avec des cellules 5T2MM et austade terminal de la maladie. Notez l’existence de nombreuses

Figure 4 Scanning electron microscopy micrographs of theendosteal surface from vertebral body of a 5T2MM mouse. Thewhole surface is eroded. Note the typical appearance of erodedsurfaces formed of valleys and pits.Image de microscopie électronique à balayage de la surfaceendostée d’un corps vertébral chez une souris 5T2MM. Toute las(

lsoi(bone resorption, it confirms radiological findings, allows

lésions ostéolytiques à la fois sur le fémur distal et le tibiaproximal.

stage of the disease. Around the tumor nodules, there is aconsiderable increase in the number of osteoclasts. Theyare responsible for the resorption of trabecular as well ascortical bone. At the end stage of the disease, trabecular

bone has completely disappeared and cortical perforationsare evidenced (Fig. 3). Scanning electron microscopy wasused to visualize and quantify the endosteal resorption on

qhn

Figure 3 Goldner trichrome stained section of the femur from 5TAt early stage, cortical bone and trabecular bone are preserved wheand several cortical perforations were observed (arrow). Note the eperiosteum.Coupe histologique de fémur colorée au trichrome de Goldner etterminal (B) de la maladie. Au stade précoce, l’os cortical et l’os

trabéculaire a presque entièrement disparu et plusieurs perforatioprolifération tumorale extra-médullaire, sous le périoste.

urface est érodée. Noter l’aspect typique des surfaces érodéesaspect de « coups de dents » ostéoclastiques).

ong bones. Eroded surface occupied almost the whole boneurface (Fig. 4). It has been shown that eroded surfacesccupy 76.2 ± 17.1% in 5T2MM mice cells vs. 19.2 ± 7.1%n control mice [49]. X-ray microcomputed tomographymicroCT) with 3D reconstruction can accurately visualize

uantitative analysis of trabecular and cortical bone, andistological analysis is possible because the technique ison-destructive for the samples (Fig. 5).

2MM mice at early stage (A) and end stage (B) of the disease.reas at the end stage, trabecular bone has almost disappearedxistence of extramedullary extension of the tumor under the

effectuée chez des souris 5T2MM à un stade précoce (A) ettrabéculaire sont préservés tandis qu’au stade terminal, l’osns corticales sont observées (flèche). Noter l’existence d’une

68 H. Libouban

Figure 5 X-ray microtomographic images of the tibia of a control mouse (A) and of a 5T2MM mouse at end stage of the disease.Note the existence of numerous cortical perforations extending from the endosteal to the periosteal surface of the bone and thecomplete disappearance of the trabecular network.Microtomographie-X du tibia d’une souris contrôle (A) et d’une souris 5T2MM au stade terminal de la maladie. Noter l’existence den dostt

vMttost([wtipsasRsoicM1iiIss

mfT

M

mPwcsmsw[ao

mrbbmo(vpmaaaia5owb

ombreuses perforations corticales s’étendant de la surface enrabéculaire.

The three models 5T2MM, 5T33MM and 5TGM1 have pro-ided very useful data on key actors in the pathogenesis inM and on the testing of new drugs. In MM, osteolysis is

he consequence of both an increase in osteoclastic resorp-ion and a decrease in osteoblastic formation. Alterationf the bone turnover is due to the release of local factorsecreted by PCs directly or indirectly via normal cells ofhe bone microenvironment. Identification of these factorstermed osteoclast activating factors [OAF] by Mundy et al.52]) came first from in vitro studies but their roles in vivoere confirmed by using the 5TMM models. The key role of

he RANK/RANKL/OPG system in osteolysis has been shownn 5T2MM mice treated with a recombinant osteoprotegerinrotein. Administration of OPG prevents osteolysis and pre-erves trabecular bone [53]. More recently it has been shownn increase of soluble RANKL in the serum of 5T2MM mice aseen in MM patients [54,55]. The increased level of solubleANKL is associated with dramatic bone destruction stronglyuggesting a key role for soluble RANKL in the pathogenesisf osteolysis [54]. In addition soluble RANKL is not detectedn 5T33MM mice, thus confirming that 5T2MM model is morelosely related to the human disease. Increased resorption inM is also due to macrophage inflammatory protein-1� (MIP-�) produced and released by PCs. In the 5TGM1 model, thencrease in MIP-1� stimulates osteoclastic resorption; thisncrease is, in part, dependent from RANKL pathway [56].nhibition of MIP-1� in vivo has an effect both on osteoly-is and on the tumor burden leading to an increase in theurvival of mice.

Another specific aspect of MM concerns the molecularechanisms leading to a reduced bone formation. Several

actors have been identified: DKK1, sFRP-2, IL-3, Runx2 andGF-� [2].

Most factors implicated in the pathogenesis ofM are produced and secreted in the bone marrow

iCcs

ée vers la surface périostée et la disparition totale du réseau

icroenvironment, which is essential for the growth ofCs. In addition to local factors, physical interaction of PCsith various cells of the bone marrow (stromal cells, osteo-lasts, endothelial cells) is required. Several studies haveuggested that PCs modify bone marrow at the cellular andolecular level to promote growth and survival. A recent

tudy has shown an increase in stromal cell number andith a mirror decrease in osteoblasts number in MM patients

57]. This has been confirmed in the 5TGM1 model wheren increase in stromal cells associated with a decrease ofsteoblast number correlates with the tumor burden [57].

Another way to show the key role played by theicroenvironment in MM progression is to modify the bone

emodeling prior to the inoculation of PCs. Indeed, it haseen shown that a high bone remodeling level (as inducedy ovariectomy in mice) accelerates PC growth in the 5T2MMurine model and considerably increases the number of

steolytic foci (Fig. 6) [58]. Moreover, a subline of 5T2MMthe 5THL cells) has been selected by the altered microen-ironment resulting in a more aggressive cell line withreservation of the original clone phenotype [51]. 5THLice are characterized by a M-protein detectable 6 weeks

fter cells injection and osteolytic lesions start to appeart eight weeks; end stage of the disease being obtainedt 10 weeks. Another method for increasing bone remodel-ng is to feed animals with a low calcium diet. This induces

secondary hyperparathyroidism and animals grafted withTHL PCs have a considerable increase in the extend of oste-lytic foci together with a more severe disease associatedith frequent paraplegia due to invasion of the spinal canaly PCs (Fig. 7) [59]. The microenvironment is also enriched

n microparticles (fragments of PC membrane covered withD138) released by the malignant cells but the role of theseellular debris remains unknown [60]. Up to recently, notudy on the 5TMM model was conducted to evaluate the

The use of animal models in multiple myeloma 69

Figure 6 X-ray microtomographic image of the tibia of a5T2MM mouse 10 weeks after inoculation of PCs. This mousewas characterized by an accelerated MM development, inducedby a high bone remodeling level in the bone microenvironment.Note that the osteolysis is more pronounced compared to what isobserved at a later time in a normal 5T2MM mouse (see Fig. 5B).Microtomographie-X du tibia d’une souris 5T2MM, 10 semainesaprès inoculation de plasmocytes tumoraux. Cette souris aprésenté un développement accéléré du MM, induit par unniveau élevé de remodelage osseux dans le microenviron-nement osseux. Noter que l’ostéolyse est plus prononcée parrapport à ce qui est observé après une période plus longue dans

Figure 7 Hematoxylin-eosin stained section of lumbar verte-brae of a 5THL mouse. Note the several cortical perforations(arrows) that allow proliferation of PCs under the periosteum;PCs are thus proliferating in the meninges (m). This prolifera-tion induces compression of the spinal cord (SC) which is themechanism responsible for paraplegia.Coupe histologique colorée à l’hématoxyline-éosine et obtenueà partir d’une vertèbre lombaire de souris 5THL. Noter les per-forations corticales (flèches) et l’accumulation de plasmocytestumoraux sous le périoste; les plasmocytes prolifèrent à traversles méninges (m) et cette prolifération induit une compressiondp

rissr(p5osmMdsafi

le modèle standard 5T2MM (voir Fig. 5B).

existence of gene alterations. Recent data have identifieddeletion in the SAMS1 tumor gene suppressor in the 5TGM1model, which is correlated with human findings [61]. Thesedata open the way of new possible investigations in thepathogenesis of MM using the 5TMM model.

As the 5TMM model constitutes a good model to evaluatepathophysiological aspects of MM, it is also well adapted toevaluate various therapeutic strategies in MM. The effect oftreatment combining interferon-� and melphalan was testedin 5T33MM mice [62]. This combination significantly reducestumor growth and leads to an increased animal survival.The use of antibodies against the IgG2a� M-protein reducestumor growth in most animals [63]. Bisphosphonates arecurrently the standard way to prevent and treat osteolyticdisease in MM. These compounds have anti-osteoclasticproperties by inducing apoptosis of these cells. Several typesof bisphosphonates have been tested in the 5TMM models.Initially, oral pamidronate (APD) was used in the 5T2MM[64]. Treated animals showed preservation of bone mass

and partial restoration of osteolytic lesions. However, APDhas no effect on the tumor mass itself [58,64]. The amino-bisphosphonate, ibandronate has also been tested in bothmodels 5T2MM and 5T33MM [65]. It induces a significant

twsn

e la moelle épinière (SC), mécanisme responsable de para-légie.

eduction in the occurrence of osteolytic lesions; however,t has no effect on tumor growth or animal survival. Ithould be noted that in these models, the doses of bispho-phonates used are generally 100 times higher than thoseequired in other animal models of bone hyper-resorptionsuch as the ovariectomized mice). New generations of bis-hosphonates such as ZOL or apomine were tested in theT2MM model and showed interesting results as they actn several aspect of MM disease: osteolysis, tumor growth,urvival and angiogenesis [66,67]. In these studies, treat-ent was started at early stage of the disease when the-protein is detectable in the serum. ZOL and apomine haveifferent modes of action: ZOL decreases osteoclastogene-is whereas apomine does not, suggesting a direct effect ofpomine on PCs in vivo. This direct effect is not observedor all the other bisphosphonates in vivo but has been shownn vitro when PCs are cultured in presence of ZOL added

o the medium [68]. A comparison between ZOL and APDas recently done in the 5THL model. Previous studies have

hown that APD preserved the trabecular bone mass but hado effect on the amount of cortical perforations. A microCT

7

ae[

pswiIarifmtsIgm[

ft

•

•

•

i

C

Ateearhnh

D

Tc

R

[

[

[

[

[

[

[

[

[

[

0

nalysis evidenced that ZOL was superior to APD and consid-rably reduced the number and size of cortical perforations69].

Because of advances in the comprehension of MM patho-hysiology, novel therapeutic approaches have emerged andome were tested in the 5TMM model. The proteasome,hich constitutes the physiological way to degrade proteins

n eucaryotic cells, was found to play a key role in MM.t is involved in the regulation of the RANK/RANKL systemnd proteasome inhibitors (such as bortezomid) are cur-ently used in the treatment of human MM [70]. Proteasomenhibitors have also been shown to promote osteoblast dif-erentiation and bone formation [71]. Treatment of 5T2MMice with bortezomid reduces the incidence of bone lesions,

umor growth and angiogenesis [72]. The effect of protea-ome inhibitors in vivo is mediated by PTH receptors [73].n parallel to these treatments, radioimmunotherapy tar-eting syndecan (CD138 antigen) were conducted in 5T3MMice and it has been suggested to use in combined therapies

74—76].Because MM always induced a marked reduction in bone

ormation, several trial have been done with drugs knowno stimulate osteoblasts by acting on the Wnt pathway:

fluoride was used in clinical practice but induced fluorosiswithout healing of MM lesions [77,78] however, the drugwas never evaluated in the mouse;

parathyroid hormone was successfully tested in the SCID-hu and SCID-rab model and promoted bone formation [79],however, clinical data are contradictory [80];

stimulation of the Wnt pathway by lithium chloride pro-tects 5TGM1 mice against osteolysis and decreases tumorgrowth [81].

The use of anti-DKK1 antibodies in the 5T2MM micenduces similar effects [82].

onclusion

s usual with animal models, the development of such sys-ems is very interesting since they offer the possibility tovaluate new hypothesis and therapeutics strategies. How-ver, they do not represent exactly the human diseasend differences are noted when trying to predict humanesponse to drugs. Animal models are interesting to testypothesis and understand the pathophysiological mecha-isms of the disease and exploring new genetic or pathwaysypotheses.

isclosure of interest

he author declares that he has no conflicts of interest con-erning this article.

eferences

[1] Roodman GD. Pathogenesis of myeloma bone disease.Leukemia 2009;23:435—41.

[2] Edwards CM, Zhuang J, Mundy GR. The pathogenesis of thebone disease of multiple myeloma. Bone 2008;42:1007—13.

[

H. Libouban

[3] Michigami T, Shimizu N, Williams PJ, Niewolna M, DallasSL, Mundy GR, et al. Cell-cell contact between mar-row stromal cells and myeloma cells via VCAM-1 andalpha(4)beta(1)-integrin enhances production of osteoclast-stimulating activity. Blood 2000;96:1953—60.

[4] Giuliani N, Bataille R, Mancini C, Lazzaretti M, BarilleS. Myeloma cells induce imbalance in the osteoprote-gerin/osteoprotegerin ligand system in the human bonemarrow environment. Blood 2001;98:3527—33.

[5] Uchiyama H, Barut BA, Mohrbacher AF, Chauhan D, AndersonKC. Adhesion of human myeloma-derived cell lines to bonemarrow stromal cells stimulates interleukin-6 secretion. Blood1993;82:3712—20.

[6] Feliers D, Woodruff K, Abboud S. Potential role of insulin-like growth factor binding protein-4 in the uncouplingof bone turnover in multiple myeloma. Br J Haematol1999;104:715—22.

[7] Oshima T, Abe M, Asano J, Hara T, Kitazoe K, Sekimoto E, et al.Myeloma cells suppress bone formation by secreting a solubleWnt inhibitor, sFRP-2. Blood 2005;106:3160—5.

[8] Tian E, Zhan F, Walker R, Rasmussen E, Ma Y, Barlogie B, et al.The role of the Wnt-signaling antagonist DKK1 in the develop-ment of osteolytic lesions in multiple myeloma. N Engl J Med2003;349:2483—94.

[9] Ehrlich LA, Chung HY, Ghobrial I, Choi SJ, Morandi F, Colla S,et al. IL-3 is a potential inhibitor of osteoblast differentiationin multiple myeloma. Blood 2005;106:1407—14.

10] Giuliani N, Colla S, Morandi F, Lazzaretti M, Sala R, Bonomini S,et al. Myeloma cells block RUNX2/CBFA1 activity in human bonemarrow osteoblast progenitors and inhibit osteoblast formationand differentiation. Blood 2005;106:2472—83.

11] Standal T, Abildgaard N, Fagerli UM, Stordal B, Hjertner O,Borset M, et al. HGF inhibits BMP-induced osteoblastogen-esis: possible implications for the bone disease of multiplemyeloma. Blood 2007;109:3024—30.

12] Gado K, Silva S, Paloczi K, Domjan G, Falus A. Mouse plasma-cytoma: an experimental model of human multiple myeloma.Haematologica 2001;86:227—36.

13] Potter M. Experimental plasmacytomagenesis in mice. HematolOncol Clin North Am 1997;11:323—47.

14] LeGrand J, Park ES, Wang H, Gupta S, Owens Jr JD, Nel-son PJ, et al. Global gene expression profiling in mouseplasma cell tumor precursor and bystander cells reveals poten-tial intervention targets for plasma cell neoplasia. Blood2012;119:1018—28.

15] Park ES, Shaughnessy Jr JD, Gupta S, Wang H, Lee JS, WooHG, et al. Gene expression profiling reveals different path-ways related to Abl and other genes that cooperate with c-Mycin a model of plasma cell neoplasia. BMC Genomics 2007;8:302.

16] Potter M, Wiener F. Plasmacytomagenesis in mice: model ofneoplastic development dependent upon chromosomal translo-cations. Carcinogenesis 1992;13:1681—97.

17] Silva S, Kovalchuk AL, Kim JS, Klein G, Janz S. BCL2 acceleratesinflammation-induced BALB/c plasmacytomas and promotesnovel tumors with coexisting T(12;15) and T(6;15) transloca-tions. Cancer Res 2003;63:8656—63.

18] Barille S, Bataille R, Amiot M. The role of interleukin-6 and interleukin-6/interleukin-6 receptor-alpha complex inthe pathogenesis of multiple myeloma. Eur Cytokine Netw2000;11:546—51.

19] Hilbert DM, Kopf M, Mock BA, Kohler G, Rudikoff S. Interleukin6 is essential for in vivo development of B lineage neoplasms.J Exp Med 1995;182:243—8.

20] Lattanzio G, Libert C, Aquilina M, Cappelletti M, Ciliberto G,

Musiani P, et al. Defective development of pristane-oil-inducedplasmacytomas in interleukin-6-deficient BALB/c mice. Am JPathol 1997;151:689—96.

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

The use of animal models in multiple myeloma

[21] Henk HJ, Teitelbaum A, Perez JR, Kaura S. Persistency withzoledronic acid is associated with clinical benefit in patientswith multiple myeloma. Am J Hematol 2012;87:490—5.

[22] Avcu F, Ural AU, Yilmaz MI, Ozcan A, Ide T, Kurt B, et al.The bisphosphonate zoledronic acid inhibits the developmentof plasmacytoma induced in BALB/c mice by intraperitonealinjection of pristane. Eur J Haematol 2005;74:496—500.

[23] Alsina M, Boyce B, Devlin RD, Anderson JL, Craig F, Mundy GR,et al. Development of an in vivo model of human multiplemyeloma bone disease. Blood 1996;87:1495—501.

[24] Hjorth-Hansen H, Seifert MF, Borset M, Aarset H, Ostlie A,Sundan A, et al. Marked osteoblastopenia and reduced boneformation in a model of multiple myeloma bone disease insevere combined immunodeficiency mice. J Bone Miner Res1999;14:256—63.

[25] Mitsiades CS, Mitsiades NS, Bronson RT, Chauhan D, Munshi N,Treon SP, et al. Fluorescence imaging of multiple myeloma cellsin a clinically relevant SCID/NOD in vivo model: biologic andclinical implications. Cancer Res 2003;63:6689—96.

[26] Pilarski LM, Hipperson G, Seeberger K, Pruski E, Coupland RW,Belch AR. Myeloma progenitors in the blood of patients withaggressive or minimal disease: engraftment and self-renewalof primary human myeloma in the bone marrow of NOD SCIDmice. Blood 2000;95:1056—65.

[27] Tsunenari T, Koishihara Y, Nakamura A, Moriya M, Ohkawa H,Goto H, et al. New xenograft model of multiple myeloma andefficacy of a humanized antibody against human interleukin-6receptor. Blood 1997;90:2437—44.

[28] Goto H, Shimazaki C, Tatsumi T, Yamagata N, Fujita N, TsuchiyaM, et al. Establishment of a novel myeloma cell line KPMM2carrying t(3;14)(q21;q32), which proliferates specifically inresponse to interleukin-6 through an autocrine mechanism.Leukemia 1995;9:711—8.

[29] Choi SJ, Oba Y, Gazitt Y, Alsina M, Cruz J, Anderson J, et al.Antisense inhibition of macrophage inflammatory protein 1-alpha blocks bone destruction in a model of myeloma bonedisease. J Clin Invest 2001;108:1833—41.

[30] Cruz JC, Alsina M, Craig F, Yoneda T, Anderson JL, Dallas M,et al. Ibandronate decreases bone disease development andosteoclast stimulatory activity in an in vivo model of humanmyeloma. Exp Hematol 2001;29:441—7.

[31] Sordillo EM, Pearse RN. RANK-Fc: a therapeutic antagonist forRANK-L in myeloma. Cancer 2003;97:802—12.

[32] Asosingh K, Radl J, Van Riet I, Van Camp B, Vanderkerken K. The5TMM series: a useful in vivo mouse model of human multiplemyeloma. Hematol J 2000;1:351—6.

[33] Ahsmann EJ, van Tol MJ, Oudeman-Gruber J, Lokhorst H, Uyt-dehaag FG, Schuurman HJ, et al. The SCID mouse as a modelfor multiple myeloma. Br J Haematol 1995;89:319—27.

[34] Feo-Zuppardi FJ, Taylor CW, Iwato K, Lopez MH, Grogan TM,Odeleye A, et al. Long-term engraftment of fresh humanmyeloma cells in SCID mice. Blood 1992;80:2843—50.

[35] Yaccoby S, Barlogie B, Epstein J. Primary myeloma cellsgrowing in SCID-hu mice: a model for studying the biol-ogy and treatment of myeloma and its manifestations. Blood1998;92:2908—13.

[36] Fulciniti M, Tassone P, Hideshima T, Vallet S, NanjappaP, Ettenberg SA, et al. Anti-DKK1 mAb (BHQ880) as apotential therapeutic agent for multiple myeloma. Blood2009;114:371—9.

[37] Pearse RN, Sordillo EM, Yaccoby S, Wong BR, Liau DF, Colman N,et al. Multiple myeloma disrupts the TRANCE/osteoprotegerincytokine axis to trigger bone destruction and promote tumorprogression. PNAS 2001;98:11581—6.

[38] Yaccoby S, Johnson CL, Mahaffey SC, Wezeman MJ, Barlogie B,Epstein J. Antimyeloma efficacy of thalidomide in the SCID-humodel. Blood 2002;100:4162—8.

71

39] Yata K, Yaccoby S. The SCID-rab model: a novel in vivo systemfor primary human myeloma demonstrating growth of CD138-expressing malignant cells. Leukemia 2004;18:1891—7.

40] Radl J, Hollander CF, van den Berg P, de Glopper E. Idiopathicparaproteinaemia. I. Studies in an animal model — the ageingC57BL/KaLwRij mouse. Clin Exp Immunol 1978;33:395—402.

41] Radl J. Multiple myeloma and related disorders. Lessons froman animal model. Pathol Biol (Paris) 1999;47:109—14.

42] Libouban H, Onno C, Pascaretti-Grizon F, Gallois Y, Moreau MF,Baslé MF, et al. Absence of renal lesions in C57BL/KaLwRijmice with advanced myeloma due to 5T2MM cells. Leuk Res2006;30:1371—5.

43] Vanderkerken K, De Raeve H, Goes E, Van Meirvenne S, Radl J,Van Riet I, et al. Organ involvement and phenotypic adhesionprofile of 5T2 and 5T33 myeloma cells in the C57BL/KaLwRijmouse. Br J Cancer 1997;76:451—60.

44] Garrett IR, Dallas S, Radl J, Mundy GR. A murine model ofhuman myeloma bone disease. Bone 1997;20:515—20.

45] Alici E, Konstantinidis KV, Aints A, Dilber MS, Abedi-ValugerdiM. Visualization of 5T33 myeloma cells in the C57BL/KaLwRijmouse: establishment of a new syngeneic murine model ofmultiple myeloma. Exp Hematol 2004;32:1064—72.

46] Oyajobi BO, Munoz S, Kakonen R, Williams PJ, Gupta A, Wide-man CL, et al. Detection of myeloma in skeleton of miceby whole-body optical fluorescence imaging. Mol Cancer Ther2007;6:1701—8.

47] Asosingh K, Gunthert U, Bakkus MH, De Raeve H, Goes E, VanRiet I, et al. In vivo induction of insulin-like growth factor-Ireceptor and CD44v6 confers homing and adhesion to murinemultiple myeloma cells. Cancer Res 2000;60:3096—104.

48] Menu E, Asosingh K, Van Riet I, Croucher P, Van Camp B,Vanderkerken K. Myeloma cells (5TMM) and their interac-tions with the marrow microenvironment. Blood Cells Mol Dis2004;33:111—9.

49] Libouban H, Moreau MF, Lesourd M, Baslé MF, Chappard D.Osteolytic bone lesions in the 5T2 multiple myeloma model:radiographic, scanning electron microscopy and microtomo-graphic studies. J Histotechnol 2001;24:81—6.

50] Vanderkerken K, Goes E, De Raeve H, Radl J, Van CampB. Follow-up of bone lesions in an experimental multiplemyeloma mouse model: description of an in vivo techniqueusing radiography dedicated for mammography. Br J Cancer1996;73:1463—5.

51] Libouban H, Moreau MF, Basle MF, Bataille R, Chappard D. Selec-tion of a highly aggressive myeloma cell line by an alteredbone microenvironment in the C57BL/KaLwRij mouse. BiochemBiophys Res Commun 2004;316:859—66.

52] Mundy GR, Raisz LG, Cooper RA, Schechter GP, Salmon SE. Evi-dence of the secretion of an osteoclast stimulating factor inmyeloma. N Engl J Med 1974;291:1041—6.

53] Croucher PI, Shipman CM, Lippitt J, Perry M, Asosingh K, HijzenA, et al. Osteoprotegerin inhibits the development of osteolyticbone disease in multiple myeloma. Blood 2001;98:3534—40.

54] Buckle CH, De Leenheer E, Lawson MA, Yong K, Rabin N,Perry M, et al. Soluble rank ligand produced by myeloma cellscauses generalised bone loss in multiple myeloma. PLoS One2012;7:e41127.

55] Terpos E, Szydlo R, Apperley JF, Hatjiharissi E, Politou M,Meletis J, et al. Soluble receptor activator of nuclear factorkappaB ligand-osteoprotegerin ratio predicts survival in mul-tiple myeloma: proposal for a novel prognostic index. Blood2003;102:1064—9.

56] Oyajobi BO, Franchin G, Williams PJ, Pulkrabek D, GuptaA, Munoz S, et al. Dual effects of macrophage inflam-

matory protein-1alpha on osteolysis and tumor burden inthe murine 5TGM1 model of myeloma bone disease. Blood2003;102:311—9.

7

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

2

57] Noll JE, Williams SA, Tong CM, Wang H, Quach JM, Purton LE,et al. Myeloma plasma cells alter the bone marrow microen-vironment by stimulating the proliferation of mesenchymalstromal cells. Haematologica 2014;99:163—71.

58] Libouban H, Moreau MF, Baslé MF, Bataille R, Chappard D.Increased bone remodeling due to ovariectomy dramaticallyincreases tumoral growth in the 5T2 multiple myeloma mousemodel. Bone 2003;33:283—92.

59] Libouban H, Le Drevo MA, Moreau MF, Baslé MF, ChappardD. Dietary calcium deficiency increases osteolysis and mod-ifies gene expression profile in an animal model of multiplemyeloma. Calcif Tissue Int 2008;82:S35.

60] Benameur T, Chappard D, Fioleau E, Andriantsitohaina R, Mar-tinez MC, Clere N, et al. Plasma cells release membranemicroparticles in a mouse model of multiple myeloma. Micron2013;54—55:75—81.

61] Noll JE, Hewett DR, Williams SA, Vandyke K, Kok C, To LB,et al. SAMSN1 is a tumor suppressor gene in multiple myeloma.Neoplasia 2014;16:572—85.

62] Manning LS, Chamberlain NL, Leahy MF, Cordingley FT. Assess-ment of the therapeutic potential of cytokines, cytotoxic drugsand effector cell populations for the treatment of multiplemyeloma using the 5T33 murine myeloma model. Immunol CellBiol 1995;73:326—32.

63] Croese JW, Vissinga CS, Boersma WJ, Radl J. Immune regulationof mouse 5T2 multiple myeloma. I. Immune response to 5T2 MMidiotype. Neoplasma 1991;38:457—66.

64] Radl J, Croese JW, Zurcher C, van den Enden-Vieveen MH,Brondijk RJ, Kazil M, et al. Influence of treatment with APD-bisphosphonate on the bone lesions in the mouse 5T2 multiplemyeloma. Cancer 1985;55:1030—40.

65] Dallas SL, Garrett IR, Oyajobi BO, Dallas MR, Boyce BF, BaussF, et al. Ibandronate reduces osteolytic lesions but not tumorburden in a murine model of myeloma bone disease. Blood1999;93:1697—706.

66] Croucher PI, De Hendrik R, Perry MJ, Hijzen A, Shipman CM, Lip-pitt J, et al. Zoledronic acid treatment of 5T2MM-bearing miceinhibits the development of myeloma bone disease: evidencefor decreased osteolysis, tumor burden and angiogenesis, andincreased survival. J Bone Miner Res 2003;18:482—92.

67] Edwards CM, Mueller G, Roelofs AJ, Chantry A, Perry M, Rus-sell RG, et al. Apomine, an inhibitor of HMG-CoA-reductase,promotes apoptosis of myeloma cells in vitro and is asso-ciated with a modulation of myeloma in vivo. Int J Cancer2007;120:1657—63.

68] Derenne S, Amiot M, Barillé S, Collette M, Robillard N, BerthaudP, et al. Zoledronate is a potent inhibitor of myeloma cell

growth and secretion of IL-6 and MMP-1 by the tumoral envi-ronment. J Bone Miner Res 1999;14:2048—56.

69] N’Diaye M, Libouban H, Aguado E, Bouvard B, AudranM, Chappard D. Unwrapping microcomputed tomographic

H. Libouban

images for measuring cortical osteolytic lesions in the 5T2murine model of myeloma treated by bisphosphonate. Micron2015;68:98—106.

70] Webb SL, Edwards CM. Novel therapeutic targets in myelomabone disease. Br J Pharmacol 2014;171:3765—76.

71] Garrett IR, Chen D, Gutierrez G, Zhao M, Escobedo A, RossiniG, et al. Selective inhibitors of the osteoblast proteasomestimulate bone formation in vivo and in vitro. J Clin Invest2003;111:1771—82.

72] Deleu S, Lemaire M, Arts J, Menu E, Van Valckenborgh E, VandeBroek I, et al. Bortezomib alone or in combination with the his-tone deacetylase inhibitor JNJ-26481585: effect on myelomabone disease in the 5T2MM murine model of myeloma. CancerRes 2009;69:5307—11.

73] Zangari M, Berno T, Yang Y, Zeng M, Xu H, Pappas L, et al.Parathyroid hormone receptor mediates the anti-myelomaeffect of proteasome inhibitors. Bone 2014;61:39—43.

74] Cherel M, Gouard S, Gaschet J, Sai-Maurel C, Bruchertseifer F,Morgenstern A, et al. 213Bi radioimmunotherapy with an anti-mCD138 monoclonal antibody in a murine model of multiplemyeloma. J Nucl Med 2013;54:1597—604.

75] Gouard S, Pallardy A, Gaschet J, Faivre-Chauvet A, Bruchert-seifer F, Morgenstern A, et al. Comparative analysis of multiplemyeloma treatment by CD138 antigen targeting with bismuth-213 and melphalan chemotherapy. Nucl Med Biol 2014;41Suppl.:e30—5.

76] Sanchez E, Chen H, Berenson JR. In vivo models of multiplemyeloma (MM). Biochem Pharmacol 2014;89:313—20.

77] Harley JB, Schilling A, Glidewell O. Ineffectiveness of fluoridetherapy in multiple myeloma. N Engl J Med 1972;286:1283—8.

78] Kyle RA, Jowsey J, Kelly PJ, Taves DR. Multiple-myelomabone disease. The comparative effect of sodium fluorideand calcium carbonate or placebo. N Engl J Med 1975;293:1334—8.

79] Pennisi A, Ling W, Li X, Khan S, Wang Y, Barlogie B, et al.Consequences of daily administered parathyroid hormone onmyeloma growth, bone disease, and molecular profiling ofwhole myelomatous bone. PLoS One 2010:5.

80] Kang MG, Won EJ, Choi HW, Kim HR, Choi HJ, Park HR, et al.Serum parathyroid hormone is a new potential risk factor inmultiple myeloma. Biomed Res Int 2014;2014:804182—282.

81] Edwards CM, Edwards JR, Lwin ST, Esparza J, Oyajobi BO,McCluskey B, et al. Increasing Wnt signaling in the bone mar-row microenvironment inhibits the development of myelomabone disease and reduces tumor burden in bone in vivo. Blood2008;111:2833—42.

82] Heath DJ, Chantry AD, Buckle CH, Coulton L, Shaughnessy Jr

JD, Evans HR, et al. Inhibiting Dickkopf-1 (Dkk1) removes sup-pression of bone formation and prevents the development ofosteolytic bone disease in multiple myeloma. J Bone Miner Res2009;24:425—36.