Metaplastic woven bone in bone metastases: A Fourier-transform infrared analysis and imaging of bone quality (FTIR)

Disponibleenlignesur

ScienceDirect

www.sciencedirect.com

ORIGINAL ARTICLE

Metaplastic woven bone in bone

metastases: A Fourier-transform infrared analysis and imaging of bone quality (FTIR)

Le tissu osseux métaplasique dans les métastases osseuses : analyse et imagerie de la qualité osseuse en microscopie infra-rouge à transformée de Fourier (FTIR)

D. Chappard

, G. Mabilleau

, C. Masson

, A. Tahla

, E. Legrand

aGroupeétudesremodelageosseuxetbiomatériaux,Gerom,SFR42—08,universitéd’Angers,IRIS-IBS, institutdebiologieensanté,CHUd’Angers,49933Angerscedex,France

bServicederhumatologie,CHUd’Angers,49033Angerscedex,France

cDépartementdechirurgieosseuse,CHUd’Angers,49033Angerscedex,France

KEYWORDS Bonemetastasis;

FTIR;

Wovenbone;

Osteosclerosis;

Infra-redmicroscopy

Summary Most osteolytic tumors are in fact mixed and contain an osteoblastic compo- nent associated with the predominant osteolytic areas. This metaplastic woven bone is always evidenced by histological analysis even in the absence of radiological expression.

Metaplastic bone formation reflects the activation of new osteoblasts coming from the stimulation ofthe dormant lining cells. Twelve patients with secondary metastases ofthe iliac crest evidenced by hot spots on a ⁹⁹Tc-MBP san were diagnosed by histomorphome- try on bone biopsies. Fourier Transformed InfraRed analysis and Imaging (FTIRI) was used on 4␮m thick sections of undecalcified bone. The mineralization degree, carbonate sub- stitution, crystallinity and the cross-links ratio of collagen (1660/1690cm⁻¹ bands) were determined.Thematrixcharacteristicswereanalyzedandimagedinthepre-existingresidual bone and in the metaplastic woven bone in the vicinity of the tumor cells. FTIRI pro- vided imagesofthephosphate, amideandcombination ofpeakratio afterhavingselected the peaks of interest. In addition, the matrix properties can be measured and compared between the old and newly-formed bones. Woven bone appeared poorly calcified with a low phosphate/amide ratio (P=0.03) crystallinity (P<0.0001) and carbonate substitution (P=0.003). Collagen was less mature as evidenced by lower cross-links (P=0.01). Woven bone associated with bone metastasis appears poorly mineralized and rapidly elaborated

∗Correspondingauthor.

E-mailaddress:daniel.chappard@univ-angers.fr(D.Chappard).

https://doi.org/10.1016/j.morpho.2018.02.002

1286-0115/©2018ElsevierMassonSAS.Allrightsreserved.

byosteoblasts.Thecollagenousphaseofthebonematrixhasalowlevelofreticulation.FTIRI isapowerfultooltomeasureandvisualizethevariouscomponentsofthebonematrixinhuman diseases.

©2018ElsevierMassonSAS.Allrightsreserved.

Résumé Laplupartdestumeursostéolytiquessontenfaitmixtesetcontiennentuncomposant ostéoblastiqueassociéauxzonesostéolytiquesquiprédominent.Cetosmétaplasiquefibreux, nonlamellaire,esttoujoursmisenévidencehistologiquementmêmeenl’absenced’expression radiologique. La formation d’osmétaplasique reflète l’activation de nouveaux ostéoblastes provenantdelastimulationdescellulesbordantes.Douzepatientsprésentantdesmétastases osseusessecondairesmisesenévidencepardesfoyerschaudsenscintigraphieau⁹⁹Tc-MBPont étédiagnostiquésenhistomorphométriesurdesbiopsiesosseusesdelacrêteiliaque.L’analyse et l’imagerie infrarougesà transforméede Fourier(FTIRI) ont étéutilisées sur des coupes de 4␮md’épaisseur d’osnon décalcifié.Laminéralisation, letaux de substitution pardes carbonates,lacristallinitéetlerapportdescross-linksducollagène(bandesà1660/1690cm⁻¹) ontétédéterminés.Lescaractéristiquesdelamatriceosseuseontétéanalyséesetimagéesdans l’osrésiduelpréexistantetdansl’osmétaplasiqueauvoisinagedescellulestumorales.LaFTIRIa fournidesimagesdeladistributiondesphosphates,desamidesetdesrapportsdecombinaisons depicsaprèsavoirsélectionnélespicsd’intérêt.Deplus,lespropriétésdelamatriceontpu êtremesuréesetcomparées entrel’os résidueletl’os métaplasique.Cedernier estapparu faiblementcalcifiéavecunrapportphosphate/amide(p=0,03),unecristallinité(p<0,0001) etunesubstitutioncarbonate(p=0,003)réduits.Lecollagèneétaitmoinsmaturecommeen témoignentuneplusfaiblequantitédecross-links(p=0,01).L’osmétaplasiqueaucoursdes métastasesosseusesapparaîtfaiblementminéraliséetrapidementélaboréparlesostéoblastes.

Laphasecollagéniquedelamatriceosseuseprésenteunfaibletauxderéticulation.LaFTIRI estunoutilpuissantpourmesureretvisualiserlesdifférentscomposantsdelamatriceosseuse danslesostéopathieshumaines.

©2018ElsevierMassonSAS.Tousdroitsr´eserv´es.

Introduction

In patients with advanced cancer, bone is frequently involved, particularly is the case of breast and prostate adenocarcinomas that have been recognized as highly osteophilic tumors several centuries ago [1,2]. In these twotypesof malignancies, bonemetastases areobserved in 60—70% of patients. Malignant cells have a particu- lar propensity for homing tothe bone morrow since they expressvarious chemokines or synthetize variousproteins that favor their anchorage in the bone marrow microen- vironment. Bone marrow, in turn, can provide a ‘‘fertile soil’’ for the growth of metastatic cells: a large number of growth factors and cytokines are naturally present to control the growth and differentiation of hematopoietic cells [3].Also, bone remodeling(which is normallymain- tained atequilibrium by acomplexnetwork of cytokines) is altered by malignant cells. After arrival in the bone marrowmicroenvironment,tumorcellsinteractwithbone remodeling and disrupt the coordinated action of osteo- clasts(boneresorbingcells)andosteoblasts(boneforming cells) [4]. Typically metastases are usually considered as osteosclerotic or osteolytic, based on the X-ray appear- anceofthebonelesions[5].Osteocleroticmetastasesare commonin prostate cancer while lung, breast, kidney or gut tumors typicallyproduce osteolytic metastases.How- ever,most of theosteolytictumorsarein factmixedand

containanosteoblasticcomponentassociatedwiththepre- dominantosteolytic areas.This osteoblastic component is stimulated by the release of growth factors (secreted by the malignantcellssuch asendothelin-1)or buriedin the bonematrixduringtheosteoclasticresorptionphase[6,7].

Amongthem,transforminggrowthfactor␤(TGF-␤)isstored athigh concentrations.Severalstudieshavereportedthat it can activate tumor cellprogression via the smad path- way,this producinga ‘‘viciouscircle’’betweenmalignant andbonecells[8].Duringthephysiologicalremodelingpro- cess,TGF-␤controlsthecouplingbetweenboneformation andresorption.TGF-␤inhibitsosteoclastdifferentiationand promotestherecruitmentofnewosteoblastsintheeroded areas.Inosteolyticmetastases,osteoclastactivityisconsid- erablyincreased(byanumberofcytokinesreleasedbythe tumorcells),thusreleasinghighamountofTGF-␤thatcan provoketheformationofmetaplasticboneinthevicinityof theosteolyticfoci[8,9].

Recently, new histological methods have appeared to characterize thequalityof thebonematrix in humandis- eases. Fourier transformed infrared analysis and imaging (FTIRI) is an interesting technique for imaging the min- eralandtheorganicphasesofthematrixintissuesections [10,11].Theaimofthepresentretrospectivestudywasto analyzethebonematrixinpatientswithmixedmetastases containinganoticeableamountofwovenbone.Undecalci- fiedbonesectionswereusedandthematrixcharacteristics

FTIRanalysisofbonematrixinmetastases 3 Table 1 Patients characteristics with type of metastasis on X-ray images and osteoid parameters determined by histomorphometry.

Patient Gender X-ray Tumororigin OV/BV(%) OS/BS(%) O.Th(␮m)

1 Female Mixed Breast 8.8^a 45.7^a 11.9

2 Male Osteosclerosis Lung 12.6^a 75.6^a 21.9^a

3 Male Mixed Prostate 2.5 13.3 13.6

4 Male Osteosclerosis Prostate 22.7^a 100^a 27.9^a

5 Male Osteosclerosis Prostate 2.9 13.9 10.9

6 Male Osteosclerosis Prostate 12.9^a 32.6^a 14.1

7 Male Osteosclerosis Prostate 31.6^a 59.1^a 24.3^a

8 Female Mixed Unknown 2.9 11.2 14.3

9 Male Osteosclerosis Unknown 29.0^a 100^a 27.3^a

10 Female Mixed Breast 6.2^a 12.1 14.8^a

11 Female Mixed Breast 7.8^a 25.5^a 9.5

12 Female Osteosclerosis Breast 17.9^a 69.1^a 15.7^a

a Increasedvs.normalvalues,i.e.<6%forosteoidvolumeOV/BV;<20%forosteoidsurfacesOS/BSand15␮mforosteoidthickness O.Th.

were determined and imaged in the pre-existing residual bone andin the newly-formedbone in thevicinity of the tumorcells.

Patients

Twelve patients (five females, seven males) from the rheumatologyunitwerereferred forthe histopathological diagnosis of a bone metastasis at the ilium. The char- acteristics of patients are given in Table 1. All patients had a ⁹⁹Tc—MBP bone scan that evidenced hot lesions.

Seven patients presented osteosclerotic lesions and five hadmixedbonylesionsassociatingscleroticandlytic foci onX-rays.Transiliac bone biopsywasdone in theseareas under local anesthesia as previously reported [12]. Six patientshadasingletetracyclinebonelabelingdonewith demethylchlortetracyline (150mg/d during two days) to evaluatetheextentofosteoidsurfaceswithactivemineral- ization.TheEthicalSubcommitteeofouruniversityhospital approved the use of patient material asthis work is ret- rospectivefromabone samplecollection(complementary histologicalstudy).

Material and methods

Atransiliacbonebiopsywasobtainedinallpatientsunder local anesthesia witha 7mm internal diametertrephine, 2cm below the iliac crest and 2cm behind the antero- superioriliacspine(Commeca,49070Beaucouzé,France).

Specimenswerefixedin anethanol-basedfluid[13].Bone cores were fixed for 24hours, dehydrated in acetone andembeddedundecalcifiedinpoly (methylmethacrylate) (pMMA) Sections (7␮m in thickness for histomorphometry and 4␮m for FTIRI) were cut dry ona heavy duty micro- tomeequippedwithtungstencarbideknives(LeicaPolycutS with50knives).TheywerestainedwithGoldner’strichrome for the identification of osteoid and hematoxylin-phloxin- saffron(HPS)forcytologicalexamination.Osteoclastswere

detected by an histoenzymologic method based on the tartrateresistant acidphosphatase(TRAcP)determination [14]. A complete histomorphometric analysis was done accordingtotherecommendationsoftheAmericanSociety forBoneandMineralResearch[15].Onlyosteoidparameters (osteoid volume, surface and thickness) were considered here.

Fourier-transforminfraredmicrospectroscopy (FTIRandFTIRI)

Analysiswas done according to previous specifications on 4 ␮m thick sections [16,17]. Sections were sandwiched between two barium fluoride (BaF2) optical windows (IR grade).Samples were then analyzed with a BrukerHype- rion 3000 microscope (Bruker optic, Ettlingen, Germany) equippedwith a liquid nitrogen-cooled Mercury Cadmium Telluride (MCT) detector and purged with dried air, and interfacedwitha vertex70spectrometer (Bruker). A 15X Cassegrainobjectiveandcondenserwithanumericalaper- tureof 0.4wereused.Infrared spectrawerecollected at a spectral resolution of 4cm⁻¹ in transmission mode on 32 co-added scans. FTIR data processing was performed withtheBruker’sproprietaryOpusSoftware(release6.5).A backgroundspectrumwascollected,containingdata from the environment, the instrument and the optical disks.

Sequential raw spectra for each regions analyzed were averaged and the contribution of the embedding pMMA wasspectrallysubtractedafter normalizationof the peak located at 1730cm⁻¹. Twenty spectra (15␮m apart in a 45×60␮m square matrix) were obtained in threediffer- entregions of independent trabeculaeper bone samples.

Analysis was done separately in the pre-existing and in thin trabeculae composed of metaplastic woven bone. In allinstances,at leastfivetrabeculae wereanalyzed from each sectionin the twodifferentlocations. Spectrawere cut between 800—1800cm⁻¹, baseline corrected with an elasticcorrectionmethod,vectornormalizedintheamide 1 region (1585—1710cm⁻¹) and averaged. Spectra were also subjected to curve fitting in the 1 3 phosphate

Figure1 Histologicalfindingsinbiopsies frompatientswithbonemetastasisassociatedwithwoven bone.A.Prostateadeno- carcinoma:notetheconsiderableextentofthinneo-trabeculaemadeofwovenboneandcoveredwithosteoidtissue(inred)and anchoredatthesurfaceofpre-existingtrabeculae(pb);Goldner’strichrome.B.Samepatient,undecalcifiedbonesectionshowing theconsiderableextentoffixationoftetracyclineinactivelymineralizingsurfaces.C.Patientwithabreastadenocarcinoma.Note theconsiderablestimulationofosteoblastogenesisandtheappearanceofnewbone(orange)laidonpre-existingbone(pb);Mc:

malignantcellsandstromareaction.D.Histoenzymologicalidentificationofnumerousosteoclasts(inbrown)bytheTRAcPreaction inapatientwithasclerosismetastasisfromaprostateadenocarcinoma.

band (900—1200cm⁻¹) and the amide I- amide II region (1485—1720cm⁻¹) according to second derivativeand the Levenberg-MaquardtalgorithmusingtheGrams/AI8.0soft- wareasreportedpreviously[18].

TheevaluatedIRspectralparameterswere:

• mineraltomatrixratio,calculatedastheintegratedarea of the ␯1 ␯3 phosphate band over the amide 1 region (1585—1720cm⁻¹)andusedasanindexofmineralcontent [10];

• mineral crystallinity, which reflects the apatite size and perfection and calculated asthe area ratio of the 1020cm⁻¹overthe1030cm⁻¹subbands[11].Alowerratio indicatesahighermineralcrystallinity;

• collagen maturity, relatedto therelative proportionof thetwomajorcollagen cross-linkspyridinium(Pyr,non- reducible)anddeH-DHLNL(reducible)andcalculatedas thearearatioofthe1660and1690cm⁻¹subbands[19];

• carbonates tophosphatesratio, which reflects thecar- bonate content in bone and calculated as ratio of integratedareasofthe2CO32− region(850—890cm⁻¹) and␯1␯3phosphateband[20].

ForFTIRI,sectionswereimagedusinga64×64element focalplane array(FPA) detector(coupled withthemicro- scope)toprovide4096spectrasimultaneouslyfrominterest areas.Theimageswerecollectedataresolutionof8cm⁻¹

with32co-addedscansforbackgroundandsamples,inthe transmission mode in the spectralregion 900—2000cm⁻¹. Images of the trabeculae wereobtained by a stitching of 3×3elementalimagesprovidedbytheFPAwiththeOpus software. Three spectral images were acquired on sep- arated area from each sample, each area containing an pre-existingtrabecularandfociofwovenbone.Eachimage was ratioed against the background imaging spectra col- lected under identical conditions. Color-coded images of mineraltomatrixratio,mineralcristallinity,collagenmatu- rity and carbonate content (usingthe intensity of the ␯3 carbonatebandat1415cm⁻¹overthe␯1␯3phosphateband distributions were produced in each groups for qualita- tive comparison by using a Matlab routine (release 7.10;

Math Works, Natick, MA) [21]. The spectroscopic calcu- latedparameterswereexpressedasacolor-codedimages, the color of a pixel corresponding to a certain range of values.

Statisticalanalysis

Statistical analysis was performed with the Systat^® sta- tistical software release 11.0 (Systat inc. San José; CA).

Data are expressed as mean±SD. Differences between groups were analysed by the non-parametric Mann and Whitney U test. Differences were considered significant whenP<0.05.

FTIRanalysisofbonematrixinmetastases 5

Results

Histologicalanalysis

Inthisseriesofpatients,theoriginoftheprimarytumorwas abreastadenocarcinomainfourcases,fiveprostateadeno- carcinoma, one lung adenocarcinomaand two cases with adenocarcinomasfromanunknownorigin.Inallcases,the histologicalexamination revealed numerous foci of meta- plasticbone (wovenbone) anchoredat thesurfaceof the previous trabeculaefromthe patient(‘‘old bone’’)inside the cancellous space (Fig. 1A). In the patient with the lungadenocarcinoma,metaplasticbone wasalsoobserved underthe periosteum.The woven texturecouldeasily be observedunderpolarizationmicroscopy.Theosteoidparam- eterswereincreasedinninepatientsindicatinganincreased osteoblastic activity (Table 1) and the amount of meta- plasticbonedifferedbetweenthepatientsandthosewith prostateadenocarcinomahadthehighestamountofmeta- plastic bone that tended to completely fill the marrow cavity. These thintrabeculae appeared witha greencore whenstainedbyGoldner’strichromeandwerecoveredby thick osteoid seams. O.Th was increased in six patients indicatinganalteredmineralization.Inpatientswithtetra-

Figure 2 Typical FTIR spectrum of the bone matrix in a patient with metastasis comprising woven bone anchoredat thesurfaceofpre-existingbone.Pre-existingtrabeculae:black spectrum;wovenbonefromthemetaplasticnewtrabeculae:

greycurve.

cyclinelabeling,themetaplasticbone washeavilylabeled bythefluorescentmarker(Fig.1B).OnHPSstainedsections, themetastatic cells supported by the collagenous stroma

Figure3 ParametersderivedfromFTIRspectrameasuredinthepre-existingboneandthenewly-formedtrabeculaeinpatients withbonemetastases.A.Mineraltoorganicmatricratio.B.Carbonatesubstitution.C.Crystallinity.D.Collagencross-links.Abscissa axes:arbitraryunits.

reaction couldbe evidenced. Numerous pseudo-epithelial seamsofosteoblasts,apposingnewboneat thesurfaceof thepre-existingtrabeculae,werealsoevidenced(Fig.1C).

TRAcPstainedsectionsevidencednumerous fociofactive osteoclastsinerodedareasduginthevicinityofthemalig- nantcells(Fig.1D).Osteoclastswerefoundatthesurfaceof pre-existingtrabeculaebuttheyalsoappearedtoremodel themetaplasticbone.

FTIRanalysis

The typicalspectrum of thebone matrix was differentin thepre-existingboneandinthemetaplasticbonefromthe newly-formedtrabeculae(Fig.2).The␯1␯2phosphatepeak wasmarkedlyreduced inthe metaplastictrabeculae.The indexesfromtheFTIRspectraappearonFig.3.Mineraliza- tion(phosphate/amideIratio)wassignificantlyreducedin newbonewhencomparedtothepre-existingbonefrompre- existingtrabeculae. Similarly, the carbonate substitution, crystallinityandtheamountofcross-linkswerereducedin thewovenbonefromthemetaplasticfoci.

FTIRIanalysis

FTIRIanalysisconfirmedtheseresultsandprovided2Dspa- tialimagingof themolecularmaps derived fromthe FTIR

spectra.Examplesoftypicalimagesaregivenforametas- tasisfromabreastcancer(Fig.4)andfromaprostatecancer (Fig.5).Theamountofphosphatewasmarkedlyreducedin thethinmetaplastictrabeculae(Figs.4B—5B)andthesur- faceareaofthewovenbonewaslessthannotedonthevideo images(Figs.4A—5A).Onthecontrary,theamideImapiden- tifiedalargerareabecausecollagenwasalsopresentinthe osteoid seams thatcover all thesetrabeculaeand anchor them onthe pre-existingones (Figs. 4C—5C).The map of thecross-linksconfirmedthattheywerebyfarlessnumer- ousin thesenewly-formedmetaplastictrabeculaethan in thepre-existingtrabecularbone(Figs.4D—5D).

Discussion

Inthisseriesofpatientswithmetastaticadenocarcinomas attheiliacbone,fociofwovenbonewerealwaysidentified atthesurfaceofthepre-existingtrabeculaeinthecancel- lousspace;theyhaveatypicalcandelabraaspectwiththin trabeculae.Inthepatientwithalungcancer,thepresence ofwovenbonealsoextendinginthesub-periostealareasis alsocharacteristicofamarrowmetastasis[22].Allpatients hadabonemetastasisthatwasclinicallyclassifiedeitheras osteoscleroticormixedonX-rays.Inaddition,onlypatients withhotspotsonthe⁹⁹Tc-MBPbonescanwereselectedprior

Figure4 FTIRIimagesobtainedinafemalepatientwithmetastasisfromabreastadenocarcinoma.A.Videoimageofa4␮m thickundecalcifiedsection.B.Spatialdistributionofphosphates.C.Spatialdistributionoftheamidegroupsofcollagen.D.Spatial distributionofthecollagencross-links,notethereducecontentinthewovenbone.Thelook-uptableisgivenatthebottom.

FTIRanalysisofbonematrixinmetastases 7

Figure5 FTIRIimagesobtainedinamalepatientwithsclerosismetastasisfromaprostateadenocarcinoma.A.Videoimageof a4␮mthickundecalcifiedsection.B.Spatialdistributionofphosphates,thepre-existingbonehasahigherphosphatecontent.C.

Spatialdistributionoftheamidegroupsofcollagenofboththemineralizedwovenboneandosteoidtissue.D.Spatialdistribution ofthecollagencross-links,notethereducecontentinthewovenbone.Thelook-uptableisgivenatthebottom.

to analysisbecause it indicates a high osteoblastic activ- itycausinganelevateduptakeoftheradionuclide-labeled bisphosphonate[23].Sometypesofcancerssuchasbreast andprostate adenocarcinomas arespecially metastasizing tobone,becausetheypossessuniquecapacitiesthatfacili- tatetheiranchoringinthemarrow[2,5].

Osteoblasticactivityisincreasedinmetaplastictrabec- ulaewithplumpedcellsandenlargedosteocytes[24].This metaplasticbonemaybeevidencedasburiedwhiteareas on X-rays but is always evidenced by histology even in the absence of radiological expression [25]. Metaplastic bone formation reflects the activation of newosteoblasts or the stimulation of the dormant lining cells at the sur- faceofbonetrabeculae.Theyelaboraterapidlyplexiform trabeculae made of woven bone, which are anchored on the pre-existing bone surfaces as shown by histology or microcomputedtomography[7,26—28].Thebonematrixis madeofrandomlydisposedbundlesofcollagenfiberswhich aresecondarilymineralized.Metaplasticboneisformedby either the release of growth factors by the tumor (e.g., endothelin-1 and IL-6 in prostate carcinoma) or due to the excess in osteoclastic activity which release locally highamountsofTGF-␤which isburiedinthebonematrix [29,30].

In this study, we found that the bone matrix of the newly-formedbone trabeculae wasincompletely mineral- izedwith a marked decreasein mineral crystallinity (i.e.

poorly made hydroxyapatite crystals). A reduction in the carbonate/phosphateratioalsoindicatesanimmaturebone matrixsincethe␤-carbonatesubstitution(i.e.,replacement of a phosphate group by a carbonate in the hydroxyap- atitecrystal)increasesupon aging[31,32]. Similarfinding wereobserved in rats which developed sclerosing metas- tasesafterinjectionofLNCaPC4-2Bprostatecancercellsin thetibia.Ramanmicrospectroscopyidentifiesadecreasein themineralizationdegree andhydroxyapatitecrystallinity [33].

Wovenboneissynthetizedbyosteoblastswithan anar- chictextureinwhichcollagenfibershaverandomdirections.

Thisbone alsomineralizesrapidlyandisreportedtohave anhigherdegreeofmineralizationthanlamellarbone[34].

However,itsbiomechanicalpropertiesarepoorerthanthose of lamellarbone becausethe irregular collagen-fibril ori- entationandhighermineralizationdegreemakeitcapable toabsorbless energy[35,36].Itshouldhoweverbenoted thatthesestudieswereconductedinthegrowingskeleton ofhealthyanimals.Inthisstudy,wovenbonewasobserved in apathologic condition where the neoplastic cells have

considerablyincreasedtheactivityofosteoblasts.Thecon- sequence is that the newly-formed bone is incompletely mineralizedwithareducedcollagen maturityin thebone matrix.Similarfindingshave beenobservedarounddental implants:theplacementofadentalimplantisfollowedby rapiddepositionofwovenboneallowingaprimaryanchor- age but the degree of mineralization and biomechanical propertiesarereducedwhencomparedwithlamellarbone [37,38].In fibrousbonedysplasia, the lamellartextureof boneisreplacedbywovenbone,whichislessmineralized and has a reduced mechanical strength [39,40]. In bone metastases,thepresenceofahighamountofincompletely calcified bone has sometimes important consequences on calciummetabolism: hypocalcemiais ametabolicdisturb- ance which is common in advanced prostate or breast cancerduetothe‘‘hungrybone’’syndrome[41,42].Insuch patients, the increase in the amount of woven bone and osteoblasticactivityareresponsibleforanincreasetrapping ofcalciumin thesclerosisareas.Treatment withbisphos- phonatesis the standard of care for patients withbreast cancerwhohavepredominantlyosteolyticlesionsasthese compoundshaveamarkedanti-osteoclasticactivy.However, bisphosphonatesarenowknowntodepressbothosteoclastic andosteoblasticactivityin asecondtimeduetothecou- plingbetweenthesecellpopulationsduringtheremodeling process.Bisphosphonatesarenowusedtoreducethebone formingactivitythusavoidingthe‘‘hungrybonesyndrome’’

[43,44].

Conclusion

Metaplasticbonepresentinbonemetastasesfromdifferent typesofadenocarcinomaarebuiltbystimulatedosteoblasts fromthetumor.Theyrapidlyelaboratetrabeculaemadeof wovenboneanchoredatthesurfaceofthepre-existingtra- beculaebutwovenbonehasapoorquality.Themineralto matrixratiowasmarkedlyreduced(evidencingalowmin- eralizationdegree),thecarbonatecontentandcrystallinity werereduced(indicating apoormaturationofhydroxyap- atite),andthe collagen maturitywasreducedwitha low amountofcross-links.Theamountofthiswovenbonewith apoorqualitymayhaveclinicalconsequences.

Disclosure of interest

Theauthorsdeclarethattheyhavenocompetinginterest.

Acknowledgements

Authors thank Mrs. Laurence Lechat for secretarial assis- tance.TheyalsothankMrsChristineAudrainandAleksandra MieczkowskaforherskillfultechnicalassistanceswithFTIRI.

This work wassupported by financial participationof the FrenchMinistryforResearch.

References

[1]Coleman RE. Skeletal complications of malignancy. Cancer 1997;80:1588—94.

[2]Guise TA, Mundy GR. Cancer and bone. Endocr Rev 1998;19:18—54.

[3]FidlerIJ.Thepathogenesis ofcancermetastasis:the‘‘seed andsoil’’hypothesisrevisited.NatRevCancer2003;3:453—8.

[4]RizzoliR,BodyJJ,BrandiML,Cannata-AndiaJ,ChappardD, ElMaghraouiA,etal.Cancer-associatedbonedisease.Osteo- porosInt2013;24:2929—53.

[5]ClézardinP,TetiA.Bonemetastasis:pathogenesisandthera- peuticimplications.ClinExpMetastasis2007;24:599—608.

[6]Fournier PGJ,Juárez P, Guise TA. Tumor-bone interactions:

thereisnoplacelikebone.In:HeymannD,editor.Bonecan- cer:progressionandtherapeuticapproaches.2ndEd.London:

Acad.Press;ElsevierInc.;2014.p.13—28.

[7]ChappardD,BouvardB,BasléMF,LegrandE,AudranM.Bone metastasis:histologicalchangesandpathophysiologicalmech- anismsinosteolyticorosteoscleroticlocalizations.Areview.

Morphologie2011;95:65—75.

[8]BaselgaJ,RothenbergML,TaberneroJ,SeoaneJ,DalyT,Clev- erlyA, etal. TGF-betasignalling-related markers incancer patientswithbonemetastasis.Biomarkers2008;13:217—36.

[9]Guise TA, Chirgwin JM. Transforming growth factor-beta in osteolyticbreastcancerbonemetastases.ClinOrthopRelat Res2003;415S:32—8.

[10]Boskey AL, Pleshko N, S.B. D, Mendelsohn R. Applications of Fourier transform infrared (FT-IR) microscopy to the study of mineralization in bone and cartilage. Cells Mater 1992;2:209—20.

[11]Paschalis EP, DiCarlo E, Betts F,Sherman P, MendelsohnR, BoskeyAL.FTIRmicrospectroscopicanalysisofhumanosteonal bone.CalcifTissueInt1996;59:480—7.

[12]AudranM,MauryE,BouvardB,LegrandE,BasléMF,Chappard D.Istransiliacbonebiopsyapainfulprocedure?ClinNephrol 2012;77:97—104.

[13]Beebe K. Alcohol/xylene: the unlikely fixa- tive/dehydrant/clearant.JHistotechnol2000;23:45—50.

[14]ChappardD.Technicalaspects:howdowebestpreparebone samplesforproperhistologicalanalysis?In:HeymannD,editor.

Bonecancer:progressionandtherapeuticapproaches.2ndEd.

London:Acad.Press;ElsevierInc.;2014.p.111—20.

[15]DempsterDW,CompstonJE,DreznerMK,KanisJA,Malluche H. Standardizednomenclature, symbols, and units for bone histomorphometry:a2012updateofthereportoftheASBMR HistomorphometryNomenclatureCommittee.JBoneMinerRes 2013;28:2—17.

[16]Gamsjaeger S,MendelsohnR, Boskey AL, Gourion-Arsiquaud S,KlaushoferK,PaschalisEP.Vibrationalspectroscopicimag- ingfortheevaluationofmatrixandmineralchemistry.Curr OsteoporosRep2014;12:454—64.

[17]PaschalisEP.Fouriertransforminfraredimagingofbone.Meth- odsMolMed2012;816:517—25.

[18]MieczkowskaA,BouvardB,ChappardD,MabilleauG.Glucose- dependent insulinotropic polypeptide (GIP) directly affects collagenfibrildiameterandcollagencross-linkinginosteoblast cultures.Bone2015;74C:29—36.

[19]PaschalisEP,Verdelis K,Doty SB,Boskey AL, MendelsohnR, YamauchiM.Spectroscopiccharacterizationofcollagencross- linksinbone.JBoneMinerRes2001;16:1821—8.

[20]BoskeyAL,DiCarloE,PaschalisE,WestP,MendelsohnR.Com- parisonofmineralqualityandquantityiniliaccrestbiopsies fromhigh-andlow-turnoverosteoporosis:anFT-IRmicrospec- troscopicinvestigation.OsteoporosInt2005;16:2031—8.

[21]Ou-YangH,PaschalisEP,MayoWE,BoskeyAL,MendelsohnR.

Infraredmicroscopicimagingofbone:spatialdistributionof CO₃(2-).JBoneMinerRes2001;16:893—900.

[22]Wlodarski KH, Reddi HA. Tumor cells stimulate in vivo periostealboneformation.BoneMiner1987;2:185—92.

[23]Nome R, Hernes E,Bogsrud TV,BjøroT, Fosså SD. Changes inprostate-specificantigen,markersofbonemetabolismand

FTIRanalysisofbonematrixinmetastases 9 bone scans after treatment with radium-223. Scand J Urol

2015;49:211—7.

[24] ChappardD,LaurentJL,BousquetG,RiffatG,OdelinMF.Mean osteoplastic surface in zones of plexiformneo-osteogenesis comparedwith mean osteoplastic surface measured on the fetalbone.BullAssocAnat1978;62:389—99.

[25] Vukmirovic-PopovicS,EscottNG,DuivenvoordenWC.Presence andenzymaticactivityofprostate-specificantigeninarchival prostatecancersamples.OncolRep2008;20:897—903.

[26] ChappardD,LiboubanH,LegrandE,IfrahN,MassonC,BasléMF, etal.Computedmicrotomographyofbonespecimensforrapid analysisofbonechangesassociatedwithmalignancy.AnatRec 2010;293:1125—33.

[27] SchmidtA,BlanchetO,DibM,BasléMF,IfrahN,ChappardD.

Bonechangesinmyelofibrosiswithmyeloidmetaplasia:ahis- tomorphometricandmicrocomputedtomographicstudy.EurJ Haematol2007;78:500—9.

[28] ChappardD.Boneandcancer(metastasisandhematologicdis- orders[Tissuosseuxetpathologiescancéreuses(métastaseset maladieshématologiques)].In:GuillaumeB,AudranM,Chap- pard D, editors. Tissu osseux et biomatériaux en chirurgie dentaire.Paris:QuintessenceInternational;2014.p.209—32.

[29] YavropoulouMP,YovosJG.TheroleoftheWntsignalingpath- wayinosteoblastcommitmentanddifferentiation.Hormones 2007;6:279—94.

[30] ChiechiA,WaningDL,StayrookKR,BuijsJT,GuiseTA,Moham- madKS.RoleofTGF-␤inbreastcancerbonemetastases.Adv BiosciBiotechnol2013;4:15—30.

[31] AkkusO,AdarF,SchafflerMB.Age-relatedchangesinphysico- chemicalpropertiesofmineralcrystalsarerelatedtoimpaired mechanicalfunctionofcorticalbone.Bone2004;34:443—53.

[32] PaschalisEP,MendelsohnR,BoskeyAL.Infraredassessmentof bonequality:areview.ClinOrthopRelRes2011;469:2170—8.

[33] BiX,SterlingJA,MerkelAR,PerrienDS,NymanJS,Mahadevan- JansenA.Prostatecancermetastasesalterbonemineraland matrixcompositionindependentofeffectsonbonearchitec- tureinmice—AquantitativestudyusingmicroCTandRaman spectroscopy.Bone2013;56:454—60.

[34]Currey J. Mechanical properties of vertebrate hard tis- sues. Proc Institut Mechanic Engin Part H. J Engin Med 1998;212:399—411.

[35]Gorski JP.Isallbone thesame?Distinctivedistributionsand properties of non-collagenous matrix proteins in lamellar vs. woven bone imply the existence of different underly- ing osteogenic mechanisms. Crit Rev OralBiol Med 1998;9:

201—23.

[36]MulderL,KoolstraJH,denToonderJM,vanEijdenTM.Rela- tionshipbetweentissuestiffnessanddegreeofmineralization ofdevelopingtrabecularbone.JBiomedMaterRes—PartA 2008;84:508—15.

[37]Brunski JB. In vivo bone response to biomechanical load- ing at the bone/dental-implant interface. Adv Dent Res 1999;13:99—119.

[38]GrizonF,AguadoE,HuréG,BasléMF,ChappardD.Enhanced boneintegrationofimplantswithincreasedsurfaceroughness:

alongtermstudyinthesheep.JDent2002;30:195—203.

[39]Chapurlat RD, Orcel P. Fibrous dysplasia of bone and McCune—Albrightsyndrome.BestPractResClinRheumatol 2008;22:55—69.

[40]DiCaprio MR, Enneking WF. Fibrous dysplasia: pathophysi- ology, evaluation, and treatment. J Bone Joint Surg Am 2005;87:1848—64.

[41]BerrutiA,DogliottiL,BitossiR,FasolisG,GorzegnoG,Bellina M,etal.Incidenceofskeletalcomplicationsinpatientswith bonemetastaticprostatecancerandhormonerefractorydis- ease:predictiveroleofboneresorptionandformationmarkers evaluatedatbaseline.JUrol2000;164:1248—53.

[42]Carroll R, Matfin G. Endocrine and metabolic emergencies:

hypocalcaemia.TherAdvEndocrinolMetab2010;1:29—33.

[43]RosenL,HarlandSJ,OosterlinckW.Broadclinicalactivityof zoledronicacid in osteolytic to osteoblastic bonelesions in patientswithabroadrangeofsolidtumors.AmJClinOncol 2002;25:S19—24.

[44]Saad F. Zoledronic acid significantly reduces pathologic fractures in patients with advanced-stage prostate cancer metastatictobone.ClinProstateCanc2002;1:145—52.