Accretion of Bone Quantity and Quality in the DevelopingMouse Skeleton

Lisa M Miller,1 William Little,2 Anne Schirmer,1 Farhan Sheik,2 Bhavin Busa,2 and Stefan Judex2

ABSTRACT: In this work, we found that bone mineral formation proceeded very rapidly in mice by 1 day ofage, where the degree of mineralization, the tissue mineral density, and the mineral crystallinity reached 36%,51%, and 87% of the adult values, respectively. However, even though significant mineralization had oc-curred, the elastic modulus of 1-day-old bone was only 14% of its adult value, indicating that the intrinsicstiffening of the bone lags considerably behind the initial mineral formation.

Introduction: To meet the mechanical challenges during early development, the skeleton requires the rapidaccretion of bone quality and bone quantity. Here, we describe early bone development in the mouse skeletonand test the hypothesis that specific compositional properties determine the stiffness of the tissue.Materials and Methods: Tibias of female BALB mice were harvested at eight time-points (n � 4 each)distributed between 1 and 40 days of age and subjected to morphometric (�CT), chemical (Fourier transforminfrared microspectroscopy), and mechanical (nanoindentation) analyses. Tibias of 450-day-old mice served asfully mineralized control specimens.Results: Bone growth proceeded very rapidly; at 1 day of age, the degree of mineralization (phosphate/proteinratio), the density of mineralized bone (TMD), and mineral crystallinity had reached 36%, 51%, and 87% ofthe adult (450 days) values, respectively. Spatially, the variability in mineralization across the mid-diaphysiswas very high for the early time-points and declined over time. In contrast to the notable changes in miner-alization, carbonate substitution into the mineral lattice (carbonate/phosphate ratio) and collagen cross-linkingdid not show any significant changes over this time period. Even though significant mineralization hadoccurred, the elastic modulus of 1-day-old bone was only 14% of the adult value and increased to 89% (of itsadult value) after 40 days. Between samples of different time-points, significant positive correlations wereobserved between the elastic modulus and TMD (r2 � 0.84), phosphate/protein ratio (r2 � 0.59), andcrystallinity (r2 � 0.23), whereas collagen cross-linking showed a small but significant negative correlation(r2 � 0.15).Conclusions: These data indicate that specific chemical and morphometric properties modulate bone’s stiff-ness during early growth. The intrinsic stiffening of the bone, however, lags considerably behind the initialmineral formation, emphasizing the importance of bone mineral quality for optimizing matrix integrity.J Bone Miner Res 2007;22:1037–1045. Published online on April 2, 2007; doi: 10.1359/JBMR.070402

Key words: bone, mineralization, elastic modulus, infrared microspectroscopy, µCT, nanoindentation

INTRODUCTION

BMD MEASUREMENTS ARE the most common way to as-sess fracture susceptibility in the clinic. However, it has

become increasingly clear that there is a substantial overlapin the BMD of normal individuals and patients who sustainfractures.(1) These findings suggest that the quality of bone,in addition to bone quantity, is important for defining frac-ture risk.

An important contributor to bone quality is the chemicalmakeup of bone. It is well established that the chemical

composition and mechanical properties of bone changewith age.(2–7) However, the interdependence betweenbone’s specific chemical makeup (e.g., mineral content,composition, and crystallinity; collagen content, structure,and cross-linking) and its corresponding mechanical prop-erties (e.g., strength, toughness, stiffness) is still poorly un-derstood.

The large changes in chemical composition and mechani-cal properties in the developing mouse skeleton offer greatopportunities to advance our understanding of the relation-ships between bone’s chemical, morphometric, and me-chanical properties. Bone mineralization has been studiedin mice during embryonic development(8–10) and senes-The authors state that they have no conflicts of interest.

1National Synchrotron Light Source, Brookhaven National Laboratory, Upton, New York, USA; 2Department of Biomedical Engi-neering, Stony Brook University, Stony Brook, New York, USA.

JOURNAL OF BONE AND MINERAL RESEARCHVolume 22, Number 7, 2007Published online on April 2, 2007; doi: 10.1359/JBMR.070402© 2007 American Society for Bone and Mineral Research

1037

JO609559 1037 1045 July

cence,(6,11) and as a function of disease(12,13) and treat-ment.(14) During the period between birth and adolescence,rapid mineralization accompanies cortical bone growth asorchestrated by osteoblasts and osteoclasts that deposit ma-terial on the periosteal surface while removing tissue fromthe endocortical surface. For the mouse tibia, skeletal pa-rameters such as BMD, total and cortical areas, and thecross-sectional moment of inertia have been shown to reachmaturity before the age of 6 mo,(15,16) but the most dra-matic changes occur between prenatal(10,17) and adoles-cent(18) stages of growth.

The mechanical properties of the mouse skeleton alsochange dramatically in early growth. Embryonal mousebone is highly flexible when unmineralized, but becomestwo orders of magnitude stiffer on endochondral ossifica-tion in the mouse embryo.(10) Comparative studies of dif-ferent strains of mice have shown that both bone geometryand mineral density play a role in determining the mechani-cal properties of the adult mouse skeleton.(19–21)

In this work, we evaluated the chemical, morphometric,and mechanical properties of the mouse tibia during thefirst 40 days of life using Fourier transform infrared mi-crospectroscopy (FTIRM), �CT, and nanoindentation(NI), respectively. The focus of this study was at the mate-rial level, where we tested the hypothesis that specific com-positional properties of bone determine the stiffness of thetissue.

MATERIALS AND METHODS

Animals

All procedures were reviewed and approved by the In-stitutional Animal Care and Use Committee (IACUC).BALB/cByJ mice were chosen for this study because oftheir well-documented skeletal development and commonuse in several models of human bone disease.(9) All datawere collected from female mice to eliminate sex differ-ences in bone size and development. To cover the periodbetween initial postnatal mineralization and early adoles-cence, mice were killed at 1, 2, 4, 7, 14, 21, 30, and 40 daysafter birth (n � 4 each). Mice were weaned at the age of 21days. Adult mice (450 days; n � 5) were killed to serve asa comparison and control.

Sample preparation and handling

A tibia was obtained from each animal and fixed in 70%ethanol for 3 days and subsequently in successive solutionsof 80%, 90%, and 100% ethanol for 2 days each. The lengthof each bone was measured with calipers (for mice olderthan 2 days) or radiographically using a frontal full-length�CT image (for 1- and 2-day-old mice). Each sample wasmounted on a prescored glass slide with putty and cut at themid-diaphysis by repeated passes of a titanium-sapphire la-ser at low power at the Center for Thermal Spray Researchat Stony Brook University. The cut bone was embedded inlow viscosity epoxy resin (Epo-thin; Buehler, Lake Bluff,

IL, USA) and allowed to cure for 24 h. Excess epoxy wasremoved using abrasive paper until the cut bone surfacewas exposed. An additional ∼0.5 mm of bone was removedto eliminate any tissue that was potentially damaged duringlaser cutting. The bone surface was polished with a series offine abrasive paper and four diamond suspensions (particlesize 3, 1, 0.25, and 0.05 �m) for 5 min at each step. Thismethodology has been shown not to affect FTIRM.(22) Em-bedded samples were stored at −20°C between experimentsto minimize any changes in bone’s material properties.(23)

µCT

Samples were �CT scanned (MicroCT40; Scanco Medi-cal) before embedding and cutting. A 3D CT scan was gen-erated for a 0.3-mm-long section of the mid-diaphysis. Thissection was defined at midspan between the growth platesor between the most proximal and distal points of miner-alization. Diaphyseal samples from 1- to 40-day-old micewere scanned at 45 kV and 177 �A. Because of their greaterand more homogenous mineralization levels, 450-day-oldsamples were scanned at 70 kV and 114 �A. For all scans,1000 projections at a total integration time of 900 ms (av-erage of three frames at 300 ms each) resulted in an isotro-pic resolution of 6 �m over a matrix size of 2048 × 2048.Calibration hydroxyapatite (HA) phantoms (Scanco Medi-cal) were used to facilitate the conversion of the linear at-tenuation of a given voxel to mgHA/cm3.

Global thresholding separated bone from soft tissue andvery poorly mineralized osteoid. To enable accurate recon-structions, thresholds were identical for samples from anygiven time-point but were adjusted across time-points.(24)

Using raw images, 12 linear attenuation values (one fromeach 30° sector) were determined for three bones from eachtime-point in the first of three steps. To this end, a visuallyappropriate threshold was determined for the youngest ageand scaled to any given older time-point by multiplying thisthreshold (TH) by the ratio of their respective linear at-tenuation (LA), as recommended by the manufacturer. Forinstance, at 40 days: TH(40 days) � TH(1 day) × LA(40days)/LA(1 day). These initial thresholds were used to cal-culate the tissue mineral densities (TMDs) for all bones. Inthe second step, thresholds were recalculated. Similar to thefirst step, these intermediate thresholds were extrapolatedfrom the selected 1-day threshold by its multiplication withthe ratio of the mean TMDs calculated in the previous step.For instance, at 40 days: TH(40 days) � TH(1 day) ×TMD(40 days)/TMD(1 day). To ensure that raw andthresholded CT images matched at any given age, in thefinal step, thresholds were adjusting by up to 6% for indi-vidual age groups. In hydroxyapatite equivalent units, thefinal threshold amounted to 320, 337, 443, 466, 536, 573,615, 642, and 749 mgHA/cm3 for the nine time-points.

Mineralized cortical bone area, periosteal and endostealareas, transcortical thickness, intracortical porosity as afraction of the area between the defined periosteal and en-docortical surfaces, minimal and maximal second momentsof area, polar moment of area, and TMD were quantitatedwith software provided by the �CT manufacturer. Corticalbone area was quantified as mineralized bone volume, ex-

MILLER ET AL.1038

cluding any volume taken up by porosities, with the lengthas of the segment as referent. For 1- to 4-day-old diaphysesthat primarily consisted of bone fragments, endosteal andperiosteal surfaces were constructed by linking individualsurfaces with smooth lines. The scanner software excludedthe two rows of voxels adjacent to any given surface in itscalculation of mean TMD to avoid edge artifacts.

FTIRM

FTIRM was performed using a Thermo Nicolet Magna860 Step-Scan FTIR spectrometer and Continuum IR mi-croscope at beamline U10B of the National SynchrotronLight Source at Brookhaven National Laboratory. Datawere collected in reflection mode from the surface of thepolished bone with a 15-�m incident beam size (4 cm−1

resolution, 256 scans/pixel). A custom ratiometer was usedto normalize the data to the synchrotron beam current.Spectra were converted to absorbance using a Kramers-Kronig transformation.

For each bone sample, spectra were collected from 16–20points evenly distributed throughout the cortical shell. Forearly time-points, the cortical shell was thin (10–20 �mthick, i.e., one pixel) and incomplete, so points were se-lected where mineralized tissue was present. For the olderanimals, points were selected uniformly around the midcor-tex. In addition, one sample from each time-point was im-aged in its entirety at a step size between 12 (1-, 2-, and 4-dsamples) and 25 �m (all samples >4 days), depending onthe size of the sample. Any spectra that contained absor-bance features at 1508 and 1250 cm−1 from epoxy resinwere discarded as spectra from nonmineralized tissue.

Spectra were analyzed for three characteristic peaks thathave been associated with specific chemical components ofbone tissue. Protein (1600–1700 cm−1, baseline: 1800 cm−1),carbonate (1365–1425 cm−1, baseline: 1800 cm−1), and phos-phate (930–1150 cm−1, baseline: 930–1150 cm−1) contentswere determined by area integration.(25) The ratio ofnonreducible/reducible collagen cross-linking was deter-mined from a peak height ratio of 1660/1690 cm−1 (linearbaseline at 1800 cm−1),(26) and crystallinity was determinedfrom a peak height ratio of 1020/1030 cm−1 (baseline: 930–1150 cm−1).(27)

NI

Samples were rehydrated in PBS buffer solution for 8–10h before data collection.(18,28) Rehydration of bone samplesafter polishing may result in the measurement of intrinsicmaterial properties that are similar to those of fresh andfully hydrated tissue.(29) NI (Triboindenter; Hysitron, Min-neapolis, MN, USA) was performed under trapezoidalloading (1-mN maximum force, 15-s duration, 5-s loadingand unloading segments) using a Berkovich diamond tip(inclination angle 142.3°, radius � 100 nm). From the linearregion of the elastic unloading phase (20–95% by displace-ment), a reduced modulus of elasticity (Er) was calculatedby the standard Oliver-Pharr method, Er � (√II)/(2) * (S)/(√Ac) where S is the contact stiffness and Ac is the contactarea.(30) The tissue elastic modulus (E) was determined by

E � [Ei * Er * (1 − vb2)]/[Ei + Er * (vi

2 − 1)] where Ei is theelastic modulus of the diamond tip (1140GP), �i is Poisson’sratio of the tip (0.07), and �b is Poisson’s ratio for bone(0.3).

All samples were indented at 16–20 points evenly distrib-uted throughout the cortical shell, in the same areas asanalyzed with the FTIR microscope so that direct correla-tions between the mechanical and chemical propertiescould be obtained.

Statistical analysis

A mean ± SD for each animal was calculated for allFTIRM, �CT, and NI data at each time-point. All statisticalanalyses were performed using these results. Linear or non-linear regressions were used to correlate animal age, themorphometric parameters from the �CT data, the chemicalparameters from the FTIRM data, and the elastic modulusfrom the NI data. Regressions used the values of all indi-vidual mice. Correlation strength was reported as thesquare of the Pearson’s correlation coefficient (r2). Statis-tical significance was defined as p < 0.05. Combined ef-fects of chemical or morphometric components on theelastic modulus were assessed by a general linear model(GLM), where the elastic modulus was the dependent vari-able and chemical or morphometric components were co-variates.

To further determine the interdependence of the chemi-cal, morphometric, and mechanical parameters, a correla-tion analysis was performed on the chemical (FTIRM),morphometric (�CT), and mechanical (NI) data. All statis-tical analyses were performed using Prism 3.0 (GraphPadSoftware) or XL Stat Pro v7.5 (Addinsoft).

RESULTS

Bone morphometry and TMD

During the first 40 days of growth, the mid-diaphysealcortical bone area of the mouse skeleton increased 9-fold(Fig. 1), coinciding with increases in the periosteal area(4-fold), the endosteal area (3-fold), cortical thickness (2-fold), maximal second moment of area (31-fold), and an8-fold decrease in intracortical porosity (Table 1). Duringthis period, all morphological parameters increased linearlyexcept for the nonlinear decrease in intracortical porosity(Fig. 1). Although the focus of this study was on the first 40days of growth, it should be mentioned that the linear na-ture of the correlations between the morphometric, chemi-cal, and mechanical parameters with age would havechanged to logarithmic by including the 450-day time-point.

At 40 days, the bone had reached 61% of the corticalarea of adult 450-day-old mice, and 73% of the adult peri-osteal area, 97% of the adult endocortical area, 77% of theadult cortical thickness, 48% of the adult maximal momentof area, and 193% of the adult cortical porosity (Table 1).TMD (mgHA/cm3) of the diaphyseal tibia also increasedsignificantly with age (Fig. 1). At 1 day, the bone already

BONE QUANTITY AND QUALITY IN YOUNG MICE 1039

had accumulated 51% of the adult TMD value. Over thefollowing 40 days, TMD increased linearly (R2 � 0.93) andhad reached 80% of the value for the adult mouse bone by40 days.

Beyond the simple increase in cortical bone volume, acomplex morphological transformation occurred between 1and 40 days of age. During this period, the mineralizedportions of the bone transitioned through at least threedistinct morphological shapes (Fig. 2A). From 1 to 2 days,mineralized tissue existed as distinct vertical tracks ratherthan a complete cortical shell, with mineralized bone occu-pying only one half of the area between the periosteal andendocortical surfaces (Table 1). An extremely thin corticalshell was apparent by 4 days, which had thickened by al-most 2-fold at 7 days and by almost 3-fold at 21 days (Table1). This condensation of tissue was apparent by the >2-folddecrease in the relative amount of cortical porosity betweendays 2 and 7 (Table 1). A second morphological transfor-mation occurred from 14 to 21 days, during which the cor-tical shell developed from a circle of uniform thickness intothe combination of different surface curvatures that char-acterizes the adult tibial cortex; by 21 days, the ratio of themaximum second moment of area to the minimal momentof area, an indicator of how circular a shape is, had deviatedsubstantially from the circular value of 1.0 (Table 1). After21 days, the cortical shell continued to thicken and the dif-

ferentiation between the “flat” medial cortex and therounded lateral cortex became more apparent.

When morphometric results were compared between thedifferent thresholds that were calculated during the three-step process, differences amounted to <10%, with the ex-ception of cortical porosity.

Chemical composition

In the growing mouse, the level of mineralization (phos-phate/protein ratio) was found to increase significantly as afunction of age (Fig. 3A). Mineralization proceeded rapidlyin the very young mouse; at 1 day of age, the level of min-eralization was already 36% of the value for the mature(450 days) mouse (Table 1). By 40 days, the level of min-eralization had increased to 62% of the 450-day value. Formineral crystallinity, cortical bone from a 1-day-old mousehad already matured to 87% of the 450-day-old mouse (Fig.3B). The increase in mineralization (r2 � 0.56) displayed asimilar pattern as the temporal changes in TMD and alsofollowed a linear increase over time. In contrast to the levelof mineralization, no significant differences in carbonatesubstitution into the mineral lattice (carbonate/phosphateratio) or collagen cross-linking as a function of animal agewere observed (data not shown).

Although the level of mineralization was high even in the

FIG. 1. (A) Areal properties, (B) percentintracortical porosity, (C) moments of area,and (D) TMD as a function of animal age.Each data point represents one animal. In D,the dashed line represents the value for amature mouse (450 days).

TABLE 1. �CT, FTIRM, AND NI PARAMETERS REPORTED AS PERCENT OF THE VALUE AT 450 DAYS

DayBonearea

Periostealarea

Endostealarea

Corticalporosity

Corticalthickness

PolMOA Imax Imin TMD Phos/Prot Crystallinity Cross-linking E

1 7% 19% 33% 1580% 32% 2% 2% 2% 51% 36% 87% 112% 14%2 9% 20% 25% 1710% 42% 2% 2% 2% 52% 49% 85% 114% 15%4 8% 26% 59% 1244% 23% 2% 2% 2% 57% 34% 90% 104% 10%7 21% 37% 66% 658% 41% 9% 9% 9% 60% 48% 82% 116% 23%

14 21% 33% 59% 357% 39% 8% 8% 8% 67% 46% 96% 110% 45%21 47% 60% 91% 218% 63% 31% 33% 29% 73% 48% 88% 87% 63%30 73% 86% 118% 161% 79% 67% 70% 64% 77% 65% 84% 80% 72%40 61% 73% 97% 193% 77% 48% 48% 46% 80% 62% 94% 109% 89%

MILLER ET AL.1040

1-day-old animal, the quantity of mineralized bone wassmall and dispersed throughout the cortex (Fig. 2B). Infra-red images of the phosphate/protein ratio as a function ofanimal age emphasized that the level of mineralization at 1day is similar to that of the 4-day-old animal (mostly cyan/green pixels). However, at 1 day, the mineralized bone hadnot formed a spatially continuous and homogeneous corti-

cal shell, which was observed at 4 days of age. From 4 to 40days of age, mineralization increased (more yellow and redpixels). In addition, the level of mineralization was morehomogeneously distributed in the 14- and 40-day-oldbone.

Bone stiffness

The stiffness of the bone matrix, as measured through NIas the elastic modulus, increased significantly as a functionof animal age (r2 � 0.85; Fig. 4). In contrast to the 36%mineralization and 51% TMD that was achieved at 1 day,the elastic modulus was only 14% of the value for the ma-ture mouse at this point (Table 1), indicating that the me-chanical properties of the bone did not mature as quickly asthe chemical properties. However by 40 days, bone’s stiff-ness was statistically indistinguishable from the maturebone.

Interrelationships between chemical, morphometric,and mechanical properties

Cross-correlations were performed to determine the in-terdependence of the chemical, morphometric, and me-chanical parameters considered in this study. The elasticmodulus increased significantly with TMD (p < 0.001; r2 �0.84 linear; r2 � 0.89 quadratic), the phosphate/protein ra-tio (p < 0.001; r2 � 0.59 linear, r2 � 0.63 quadratic), andmineral crystallinity (p � 0.012; r2 � 0.23), whereas thecollagen cross-linking parameter was inversely correlated(p � 0.048; r2 � −0.15) with the elastic modulus (Fig. 5).The carbonate/phosphate ratio did not impact the elasticmodulus (data not shown). The elastic modulus was alsosignificantly correlated with all of the morphometric param-eters, including the periosteal area, endosteal area, corticalbone area, intracortical porosity, cortical thickness, mo-ments of area, and TMD (Table 2). Morphometric param-eters were all significantly correlated with each other, indi-cating a well-orchestrated process during growth. Withinthe chemical parameters, crystallinity was positively corre-lated with the phosphate/protein ratio and collagen cross-linking and negatively correlated with the amount of car-

FIG. 2. (A) �CT and (B) FTIRM phos-phate/protein images of the mouse mid-diaphysis at 1, 4, 14, and 40 days of age. Allimages for each technique are plotted on thesame intensity scale for direct comparison.

FIG. 3. (A) Phosphate/protein ratio and (B) crystallinity as afunction of animal age. Each data point represents one animal.The dashed line represents the value for a mature mouse (450days).

BONE QUANTITY AND QUALITY IN YOUNG MICE 1041

Fig 2 live 4/C

bonate substitution into the mineral lattice (Table 2).Between chemical and morphometric properties, the phos-phate/protein ratio was significantly correlated with allmorphometric properties, collagen cross-linking was corre-lated with all morphological properties (but not TMD),crystallinity was correlated only with TMD, and there wasno significant correlation between the carbonate/phosphateratio and any of the morphometric properties (Table 2).

Because TMD, phosphate/protein, crystallinity, andcross-linking were all correlated with the elastic modulus(Fig. 5), the GLM was used to determine whether thesecorrelations were the result of independent effects. For themorphometric data, the GLM showed that the elasticmodulus was independently determined only by TMD,whereas geometrical parameters were only co-correlated(p � 0.025). For the chemical parameters, the elastic modu-lus was determined only by the level of mineralization (p �0.001) and collagen cross-linking (p � 0.008) when consid-ering other chemical parameters as covariates. Crystallinitywas not found to be an independent factor when consider-ing the level of mineralization (and all other chemical pa-rameters), suggesting that crystallinity’s correlation with theelastic modulus was a secondary effect, likely related to thelevel of mineralization.

DISCUSSION

The hypothesis of this study was that a number of bone’schemical and morphometric properties determine its me-chanical properties. Not surprisingly, bone growth pro-ceeded very rapidly at an early age but it is remarkable thatthe degree of mineralization (phosphate/protein ratio), theTMD, and the mineral crystallinity had reached 36%, 51%,and 87% of the adult (450 days) values, respectively, by 1day of age. Spatially, the variability in mineralization acrossthe mid-diaphysis was very high for the early time-pointsand declined over time. In contrast to the notable changesin TMD and mineralization, carbonate substitution (car-bonate/phosphate ratio) and collagen cross-linking did notshow any significant changes over this time period. Eventhough significant mineralization had occurred by 1 day of

age, the elastic modulus of 1-day-old bone was only 15% ofits adult value but reached 89% of its adult level by 40 daysof age. Significant positive correlations were observed be-tween the elastic modulus and TMD, phosphate/protein ra-tio, and crystallinity, whereas collagen cross-linking showeda small but significant negative correlation. These resultsemphasize the importance of some, but not all, chemicalproperties in determining bone stiffness during rapid corti-cal bone growth and indicate that a number of chemical,morphometric, and mechanical properties follow a verysimilar developmental pattern.

Examination of the morphometric properties with ageshowed that the elastic modulus was independently deter-mined only by TMD, whereas geometrical parameters wereonly co-correlated. Micromechanical models describingBMD and stiffness have shown that mature trabecular bonebehaves as a densely porous material, where its stiffness isproportional to the cube of the density.(31) A similar rela-tionship was found for mature compact bone.(32) Howeverwhen a wider range of trabecular bone densities was con-sidered, a correlation closer to a quadratic function wasfound.(33) In this work, we found that the relationship be-tween elastic modulus and TMD was approximately linearwith an r2 value of 0.82. Fitting the data to a cubic or qua-dratic function improved the r2 value incrementally to 0.89.On closer examination, the lowest and highest TMD valuesdeviated the most from the linear relationship, wherechanges in TMD had a lesser effect on the elastic modulus.

To understand this relationship between TMD and elas-tic modulus, we considered the ultrastructure of bone andthe process of mineralization. Bone mineralization is initi-ated before birth and begins with endochondral ossificationof cartilage. At the nanoscale, electron micrographs of min-eralized turkey tendon have shown that mineralization ofcollagen fibrils occurs in a highly organized fashion, wherecrystals are nucleated in the hole-zones within individualcollagen fibrils.(34) Very little additional growth of the crys-tals occurs with time; the increase in mineral mass resultsfrom the multiplication of the number of crystals, not size ofthe crystals.(35) As mineralization proceeds, crystals beginto extend into the overlap zone of the collagen fibrils so thatall of the available space within the fibrils is eventuallyoccupied by the mineral crystals. With significant mineral-ization, the collagen fibrils expand in volume such that theaxial arrangement of the collagen is altered significantly.(36)

In embryonic mouse bone, the rapid mineralization isaccompanied by an increase in elastic modulus by two or-ders of magnitude over this same time period.(10) Here wefound that, despite this rapid increase in bone’s mechanicalintegrity during embryonal development, at birth, its intrin-sic stiffness lagged behind the chemical properties. This ini-tial lag may be associated with the specific locations of themineral crystals within the collagen fibrils. Electron micro-graphs have shown that, in early mineralization, the overallaccumulation of mineral mass is predominant in the colla-gen hole zones compared with overlap zones, where 64% ofthe crystals were located in the collagen hole-zones.(37) Be-cause the elastic modulus of the mouse bone remained lowuntil the TMD reached ∼65% of its adult value, it is possiblethat mineral accumulation in the hole-zones has little effect

FIG. 4. Elastic modulus as a function of animal age. The dashedline represents the value for a mature mouse (450 days).

MILLER ET AL.1042

on bone’s intrinsic stiffness. However, once the mineralcontent increased to where the overlap zones become min-eralized, the elastic modulus increased rapidly as well.

Nanomechanical models of mineralized collagen fibrilsalso suggest that interfibrillary mineral has a significant im-pact on the mechanical behavior of bone.(38,39) This patternof mineralization is consistent with the pseudo-quadraticrelationship we found between TMD and elastic modulus.At low TMD values (<800 mg HA/cm3), where the hole-zones become mineralized, the elastic modulus remainedunchanged. As TMD increases (800–1300 mg HA/cm3), theoverlap zones mineralized and bone stiffness increased lin-early. Finally, at the highest TMD values (>1300 mg HA/cm3), the overlap zones are mineralized and collagen fibrilsize and orientation change, which also has a smaller effecton elastic modulus.

Not only does tissue density influence bone stiffness, buttissue composition does as well. The phosphate/protein ra-tio and the collagen cross-linking significantly affected theelastic modulus independent of the other compositional pa-rameters. The relationship between elastic modulus andphosphate/protein ratio is approximately linear with an r2

value of 0.59. Fitting the data to a quadratic function im-proves the r2 value to 0.62. Because the phosphate/proteinratio is a measure of tissue mineralization, it is not surpris-ing that its correlation with elastic modulus was similar toTMD. However, the correlation coefficient for the phos-phate/protein ratio was lower than for TMD. This lowercorrelation is likely caused by the inherent differences be-tween TMD and the phosphate/protein ratio. �CT mea-sures the average density of a given volume of bone,whereas FTIR measures bone composition. Thus, FTIR isinsensitive to porosities in bone, unlike �CT . On the otherhand, FTIR is much more sensitive to changes in mineralcomposition (e.g., Ca/P ratio, carbonate content) that areundetected by �CT. For example, the FTIRM resultsshowed a slight increase in mineral crystallinity with agethat was influenced by a decrease in carbonate content.

The relationship between bone mineral crystallinity andelastic modulus was also approximately linear. The highdegree of crystallinity of the 1-day-old bone (87% of itsadult value) supports electron microscopy results that indi-cate bone mineral crystals multiply, but do not grow sub-stantially after formation.(35) Despite this relationship,

FIG. 5. Correlations of the elastic moduluswith (A) tissue mineral density, (B) thephosphate/protein ratio, (C) crystallinity,and (D) collagen cross-linking. Nonlinearcurves in A and B represent fits to a qua-dratic function as described in the text.

TABLE 2. RESULTING CORRELATION MATRIX (R VALUES)

VariablesElastic

modulusPhos/Prot

Carb/Phos

Cross-linking Crystallinity

Periostealarea

Endostealarea

Bonearea

Corticalporosity

Corticalthickness

Tissuedensity

Elastic modulus 1 0.77 −0.22 −0.38 0.48 0.86 0.74 0.88 −0.80 0.90 0.91Phos/Prot 0.77 1 −0.31 −0.15 0.51 0.82 0.54 0.88 −0.57 0.88 0.89Carb/Phos −0.22 −0.31 1 −0.15 −0.43 −0.09 −0.07 −0.10 0.26 −0.10 −0.19Cross-linking −0.38 −0.15 −0.15 1 0.26 −0.46 −0.56 −0.40 0.43 −0.38 −0.34Crystallinity 0.48 0.51 −0.43 0.26 1 0.29 0.12 0.36 −0.38 0.34 0.54Periosteal area 0.86 0.82 −0.09 −0.46 0.29 1 0.90 0.99 −0.77 0.96 0.93Endosteal area 0.74 0.54 −0.07 −0.56 0.12 0.90 1 0.82 −0.79 0.76 0.75Bone area 0.88 0.88 −0.10 −0.40 0.36 0.99 0.82 1 −0.76 0.98 0.96Cortical porosity −0.80 −0.57 0.26 0.43 −0.34 −0.77 −0.79 −0.76 1 −0.73 −0.81Cortical thickness 0.90 0.88 −0.09 −0.38 0.34 0.96 0.76 0.98 −0.73 1 0.93Tissue density 0.91 0.89 −0.19 −0.34 0.54 0.93 0.75 0.96 −0.81 0.93 1

Values in bold are significantly different from 0 with a significance level � � 0.05.

BONE QUANTITY AND QUALITY IN YOUNG MICE 1043

crystallinity was not an independent factor when consider-ing the level of mineralization (and all other chemical pa-rameters), suggesting that crystallinity’s correlation withelastic modulus is a secondary effect, likely related to thelevel of mineralization.

Not only does mineral content affect bone’s mechanicalproperties, but a number of recent studies have shown thatalterations in collagen structure and cross-linking may havesimilar effects.(40–44) One of the initial steps in collagenfibril formation and mineralization is the process of colla-gen cross-linking.(40) Intermolecular reducible cross-linksstabilize the young, unmineralized fibril, which serves as atemplate for mineralization.(45) In time, some of these labilecross-links mature into nonreducible pyridinoline.(26,44,46)

In the 1-day-old mouse, we found the ratio of nonreducible/reducible cross-links to be similar to that of the adultmouse. Because the 1-day-old mouse tibia was significantlymineralized, it is likely that much of the cross-linking matu-ration (observable with FTIRM) occurred before the startof this study. However, even though the collagen cross-linking ratio did not significantly change with age, its spe-cific value significantly influenced the intrinsic stiffness ofthe bone.

In summary, there is considerable bone quantity and den-sity present in the mouse diaphysis at birth. Despite the fastaccretion of bone quantity during early development, theprocess of bone mineralization and stiffening is slow com-pared with cortical surface modeling in the young adultskeleton.(18) The spatial and temporal correlations confirmthe relevance of tissue density, composition, crystallinity,and collagen cross-linking. Considering that the stiffeningof the bone lagged behind the mineralization process untilday 1 of age, but, unlike the chemical properties, had almostreached the stiffness of adult bone by day 40 of age, thissuggests that there are additional quality factors that modu-late bone’s elastic modulus. With clear evidence that bone’schemical properties and microstructure play an importantrole in defining the micromechanical properties of the skel-eton during growth, a better mechanistic understanding ofthe underlying processes may enable the diagnosis, preven-tion, and treatment of poor bone quality.

ACKNOWLEDGMENTS

We thank Theodore Feldman, Meghan Ruppel, RandySmith, and Drs Andres Laib and Jon Longtin for experttechnical support. This work was supported by NASANAG 9-1499 (SJ), the Whitaker Foundation RG-02-0564(SJ), and a SUNY-BNL Seed Grant (SJ, LM). Use of theNational Synchrotron Light Source, Brookhaven NationalLaboratory, was supported by the U.S. Department of En-ergy, Office of Science, Office of Basic Energy Sciences,under Contract DE-AC02 98CH10886.

REFERENCES

1. Cummings SR 1985 Are patients with hip fractures more os-teoporotic? Review of the evidence. Am J Med 78:487–494.

2. Rey C, Renugopalakrishnan V, Collins B, Glimcher MJ 1991Fourier transform infrared spectroscopic study of the carbon-ate ions in bone mineral during aging. Calcif Tissue Int 49:251–258.

3. Rey C, Shimizu M, Collins B, Glimcher MJ 1990 Resolution-enhanced Fourier transform infrared spectroscopy study of theenvironment of phosphate ions in the early deposits of a solidphase of calcium-phosphate in bone and enamel, and theirevolution with age. I: Investigations in the upsilon 4 PO4 do-main. Calcif Tissue Int 46:384–394.

4. Legros R, Balmain N, Bonel G 1987 Age-related changes inmineral of rat and bovine cortical bone. Calcif Tissue Int41:137–144.

5. Akkus O, Polyakova-Akkus A, Adar F, Schaffler MB 2003Aging of microstructural compartments in human compactbone. J Bone Miner Res 18:1012–1019.

6. Ferguson VL, Ayers RA, Bateman TA, Simske SJ 2003 Bonedevelopment and age-related bone loss in male C57BL/6Jmice. Bone 33:387–398.

7. Akkus O, Adar F, Schaffler MB 2004 Age-related changes inphysicochemical properties of mineral crystals are related toimpaired mechanical function of cortical bone. Bone 34:443–453.

8. Tarnowski CP, Ignelzi MA Jr, Morris MD 2002 Mineralizationof developing mouse calvaria as revealed by Raman mi-crospectroscopy. J Bone Miner Res 17:1118–1126.

9. Horton WA 2003 Skeletal development: Insights from target-ing the mouse genome. Lancet 362:560–569.

10. Tanck E, Van Donkelaar CC, Jepsen KJ, Goldstein SA, Wein-ans H, Burger EH, Huiskes R 2004 The mechanical conse-quences of mineralization in embryonic bone. Bone 35:186–190.

11. Silva MJ, Brodt MD, Fan Z, Rho JY 2004 Nanoindentationand whole-bone bending estimates of material properties inbones from the senescence accelerated mouse SAMP6. J Bio-mech 37:1639–1646.

12. Camacho NP, Rimnac CM, Meyer RA Jr, Doty S, Boskey AL1995 Effect of abnormal mineralization on the mechanical be-havior of X-linked hypophosphatemic mice femora. Bone17:271–278.

13. Camacho NP, Landis WJ, Boskey AL 1996 Mineral changes ina mouse model of osteogenesis imperfecta detected by Fouriertransform infrared microscopy. Connect Tissue Res 35:259–265.

14. Misof BM, Roschger P, Baldini T, Raggio CL, Zraick V, RootL, Boskey AL, Klaushofer K, Fratzl P, Camacho NP 2005 Dif-ferential effects of alendronate treatment on bone from grow-ing osteogenesis imperfecta and wild-type mouse. Bone36:150–158.

15. Richman C, Kutilek S, Miyakoshi N, Srivastava AK, BeamerWG, Donahue LR, Rosen CJ, Wergedal JE, Baylink DJ, Mo-han S 2001 Postnatal and pubertal skeletal changes contributepredominantly to the differences in peak bone density betweenC3H/HeJ and C57BL/6J mice. J Bone Miner Res 16:386–397.

16. Somerville JM, Aspden RM, Armour KE, Armour KJ, ReidDM 2004 Growth of C57BL/6 mice and the material and me-chanical properties of cortical bone from the tibia. Calcif Tis-sue Int 74:469–475.

17. Price C, Herman BC, Lufkin T, Goldman HM, Jepsen KJ 2005Genetic variation in bone growth patterns defines adult mousebone fragility. J Bone Miner Res 20:1983–1991.

18. Busa B, Miller LM, Rubin CT, Qin YX, Judex S 2005 Rapidestablishment of chemical and mechanical properties duringlamellar bone formation. Calcif Tissue Int 77:386–394.

19. Brodt MD, Ellis CB, Silva MJ 1999 Growing C57Bl/6 miceincrease whole bone mechanical properties by increasing geo-metric and material properties. J Bone Miner Res 14:2159–2166.

20. Jepsen KJ, Pennington DE, Lee YL, Warman M, Nadeau J2001 Bone brittleness varies with genetic background in A/Jand C57BL/6J inbred mice. J Bone Miner Res 16:1854–1862.

MILLER ET AL.1044

21. Akhter MP, Fan Z, Rho JY 2004 Bone intrinsic material prop-erties in three inbred mouse strains. Calcif Tissue Int 75:416–420.

22. Aparicio S, Doty SB, Camacho NP, Paschalis EP, Spevak L,Mendelsohn R, Boskey AL 2002 Optimal methods for process-ing mineralized tissues for Fourier transform infrared mi-crospectroscopy. Calcif Tissue Int 70:422–429.

23. Callaghan JP, McGill SM 1995 Frozen storage increases theultimate compressive load of porcine vertebrae. J Orthop Res13:809–812.

24. Halloran BP, Ferguson VL, Simske SJ, Burghardt A, VentonLL, Majumdar S 2002 Changes in bone structure and mass withadvancing age in the male C57BL/6J mouse. J Bone Miner Res17:1044–1050.

25. Miller LM, Carlson CS, Carr GL, Chance MR 1998 A Methodfor Examining the Chemical Basis for Bone Disease: Synchro-tron Infrared Microspectroscopy. Cell Mol Biol 44:117–127.

26. Paschalis EP, Verdelis K, Doty SB, Boskey AL, MendelsohnR, Yamauchi M 2001 Spectroscopic characterization of colla-gen cross-links in bone. J Bone Miner Res 16:1821–1828.

27. Paschalis EP, Betts F, DiCarlo E, Mendelsohn R, Boskey AL1997 FTIR microspectroscopic analysis of normal human cor-tical and trabecular bone. Calcif Tissue Int 61:480–486.

28. Broz JJ, Simske SJ, Greenberg AR, Luttges MW 1993 Effectsof rehydration state on the flexural properties of whole mouselong bones. J Biomech Eng 115:447–449.

29. Hoffler CE, Guo XE, Zysset PK, Goldstein SA 2005 An ap-plication of nanoindentation technique to measure bone tissueLamellae properties. J Biomech Eng 127:1046–1053.

30. Oliver WC, Pharr GM 1992 An improved technique for deter-mining hardness and elastic modulus using load and displace-ment sensing indentation experiments. J Mater Res 7:1564–1583.

31. Carter DR, Hayes WC 1977 The compressive behavior of boneas a two-phase porous structure. J Bone Joint Surg Am 59:954–962.

32. Currey JD 1988 The effect of porosity and mineral content onthe Young’s modulus of elasticity of compact bone. J Biomech21:131–139.

33. Hodgskinson R, Currey JD 1992 Young Modulus, Density andMaterial Properties in Cancellous Bone over a Large DensityRange. J Mater Sci Mater Med 3:377–381.

34. Landis WJ, Song MJ, Leith A, McEwen L, McEwen BF 1993Mineral and organic matrix interaction in normally calcifyingtendon visualized in three dimensions by high-voltage electronmicroscopic tomography and graphic image reconstruction. JStruct Biol 110:39–54.

35. Glimcher MJ 1987 The nature of the mineral component ofbone and the mechanism of calcification. Instr Course Lect36:49–69.

36. Prostak KS, Lees S 1996 Visualization of crystal-matrix struc-ture. In situ demineralization of mineralized turkey leg tendonand bone. Calcif Tissue Int 59:474–479.

37. McEwen BF, Song MJ, Landis WJ 1991 Quantitative determi-nation of the mineral distribution in different collagen zones ofcalcifying tendon using high voltage electron microscopic to-mography. J Comput Assist Microsc 3:201–210.

38. Gupta HS, Seto J, Wagermaier W, Zaslansky P, Boesecke P,Fratzl P 2006 Cooperative deformation of mineral and collagenin bone at the nanoscale. Proc Natl Acad Sci USA 103:17741–17746.

39. Gupta HS, Wagermaier W, Zickler GA, Raz-Ben Aroush D,Funari SS, Roschger P, Wagner HD, Fratzl P 2005 Nanoscaledeformation mechanisms in bone. Nano Lett 5:2108–2111.

40. Knott L, Bailey AJ 1998 Collagen cross-links in mineralizingtissues: A review of their chemistry, function, and clinical rel-evance. Bone 22:181–187.

41. Masse PG, Rimnac CM, Yamauchi M, Coburn SP, Rucker RB,Howell DS, Boskey AL 1996 Pyridoxine deficiency affects bio-mechanical properties of chick tibial bone. Bone 18:567–574.

42. Vashishth D, Gibson GJ, Khoury JI, Schaffler MB, Kimura J,Fyhrie DP 2001 Influence of nonenzymatic glycation on bio-mechanical properties of cortical bone. Bone 28:195–201.

43. Wang X, Shen X, Li X, Agrawal CM 2002 Age-related changesin the collagen network and toughness of bone. Bone 31:1–7.

44. Paschalis EP, Shane E, Lyritis G, Skarantavos G, MendelsohnR, Boskey AL 2004 Bone fragility and collagen cross-links. JBone Miner Res 19:2000–2004.

45. Otsubo K, Katz EP, Mechanic GL, Yamauchi M 1992 Cross-linking connectivity in bone collagen fibrils: The COOH-terminal locus of free aldehyde. Biochemistry 31:396–402.

46. Eyre DR, Paz MA, Gallop PM 1984 Cross-linking in collagenand elastin. Annu Rev Biochem 53:717–748.

Address reprint requests to:Lisa M Miller, PhD

National Synchrotron Light SourceBrookhaven National Laboratory

75 Brookhaven AvenueUpton, NY 11973-5000, USA

E-mail: lmiller@bnl.gov

Received in original form September 6, 2006; revised form January17, 2007; accepted March 23, 2007.

BONE QUANTITY AND QUALITY IN YOUNG MICE 1045