Bone quality changes associated with aging and disease: a review

Abstract

Bone quality encompasses all the characteristics of bone that, in addition to density, contribute to its resistance to fracture. In this review, we consider changes in architecture, porosity and composition, including collagen structure, mineral composition and crystal size. These factors all are known to vary with tissue and animal ages, and health status. Bone morphology and presence of micro-cracks which also contribute to bone quality, will not be discussed in this review. Correlations with mechanical performance for collagen cross-linking, crystallinity and carbonate content are contrasted with mineral content. Age dependent changes in humans and rodents are discussed in terms of using rodent models of disease. Examples are osteoporosis, osteomalacia, osteogenesis imperfecta, and osteopetrosis, in both humans and animal models. Each of these conditions, along with aging, are associated with increased fracture risk for distinct reasons.

Measuring bone quality

Bone quality is generally defined as a collection of properties that contribute to fracture risk in addition to bone mineral density (BMD).^1,2,3 This collection of variables, summarized in Figure 1, consists of architecture and geometry, turnover, cortical porosity, damage, and composition (extent of mineralization, mineral stoichiometry, collagen, other matrix constituents, and water content). These measures, in addition to BMD, explain bone strength. Numerous techniques exist to measure these quality variables, reviewed elsewhere,⁴ along with a few newer techniques referenced later. Architecture and geometry can be measured by micro-, nano-, or peripheral-computerized tomography and by nuclear magnetic resonance imaging.^5–7 Bone turnover is generally measured by histomorphometry⁸ although it can be approximated from urinary turnover-markers.⁹ Macroscopic cortical porosity has been measured by microcomputed tomography (micro-CT) and MRI;¹⁰ however, newer methods exist based on Raman spectroscopy^11–13 and short-wave infrared Raman spectroscopy¹⁴ that have also been applied to measure water content. Damage has been assessed by basic Fuschin staining¹⁵ and more recently by synchrotron radiation transmission x-ray microscopy with an x-ray negative stain.¹⁶ Compositional measurements include classic methods and newer techniques reviewed elsewhere.⁴ NMR was similarly used to measure water content. Backscattered electron imaging and BMDD measurements along with vibrational spectroscopic imaging methods, provide insight into spatially resolved compositional properties. In this review, we focus primarily on spectroscopic techniques, with limited illustrations from other measurements.

Figure 1.

Schematic illustrating bone quality variables that contribute to fracture risk.

The majority of the results presented in this review were produced using spatially-resolved spectroscopic imaging techniques. Readers are referred to several recent reviews^{4,17,18,19,20} for discussion of spectroscopic principles, sample preparation methods, data analysis and limitations of these techniques. These techniques vary greatly in resolution; in general, the better the spatial resolution, the worse the signal-to-noise ratio (Table 1).

Table 1.

Spatial Resolution of Techniques used to Characterize Bone Composition and Architecture

Technique	Spatial resolution	Sample preparation	Signal/noise
AFM-IR	20–50 nm	Non-decalcified 300 nm sections	Low
Backscattered electron imaging	0.2 µm	Non-decalcified 100–300 µm sections	High
FT-IR imaging	7–25 µm	Non-decalcified 1–4 µm sections	High
Microcomputed tomography	In vivo < in vitro < high resolution	Minimal	High
NMR	5–10 µm	In vitro and in vivo measurements	Depends on methods used
Raman Imaging	0.5–1 µm	No dehydration needed; limited by specimen chamber	Weak
SEM	10 nm	Thick sections	Noisy
TEM	0.2 nm	Thin sections
XTEM	1–2 µm	Thin sections

Several features of bone composition can be derived from these spectroscopic data sets. Because vibrations that are strong in Raman spectra tend to be weaker in Fourier transform infrared spectroscopy (FTIR) and vice versa, these techniques are often viewed as complimentary. Generally, in either technique, complex spectra may be deconvoluted by second derivative spectroscopy (less observer dependent), curve fitting or other deconvolution algorithms, factor analysis and other multidimensional methods (e.g., Principal component analysis). Many bone quality parameters assessed by vibrational spectroscopic imaging have been validated by comparison to independent analytical techniques using purified model compounds. The major variables reported for FTIR Imaging (FTIRI) are mineral-to-matrix ratio (mineral content per organic matrix), carbonate-to-phosphate, also called carbonate-to-mineral ratio (extent of carbonate substitution in the apatite lattice), crystallinity (crystal size and perfection), collagen cross-link ratio (related to enzymatic cross-links) and acid phosphate substitution (inversely related to mineral maturity) (Fig. 2). In addition, aggregates of DNA can be imaged, and sugar and lipid contents can be estimated. In Raman, proteoglycan content and porosity or water content similarly can be estimated. Values for these measurements may be reported in areas between fluorescent labels (to mark newly formed bone), adjacent to specific morphologic domains, in entire trabeculae or cortices, or some combination of these sites. Because the characteristics and kinetics of the bone constituents and their changes with disease differ between the periosteal, osteonal and endosteal compartments of cortical bone, measurements also should report the specific compartment examined²¹. A recent study²² showed sampling five random sites throughout the human cortex or analyzing the entire section did not affect average values, although the spread of the data (line width at half maximum of the pixel distribution or heterogeneity) was quite different.

Figure 2.

FTIR spectrum obtained from a single pixel (6.25 × 6.25 µm²) of an FTIR image. Formulae included for key parameters including mineral/matrix and carbonate/mineral (or carbonate/phosphate) from shaded peak areas and crystallinity, collagen crosslink ratio and acid phosphate substitution from peak height intensities.

Bone quality changes during development

The first study of bone composition based on vibrational spectroscopic imaging used Raman analysis to examine mouse calvaria parietal bone from embryonic day 13.5 (6 days before birth) to 6 months of age.²³ Over this time period, mineral content increased gradually from being undetectable (embryonic day 13.5–14.5) to detectable levels of carbonate-containing hydroxyapatite mineral at embryonic day 15.5. In these mice, mineral-to-matrix ratio increased gradually and continuously until post-natal day 3, again increasing at 6 months, whereas carbonate-to-phosphate ratio was constant from embryonic day 15.5 to post-natal day 3, then increased gradually.

Using FTIR, microcomputed tomography and mechanical testing, female BALB mouse tibias were evaluated at eight time periods, between 1 and 40 days of age, then compared to tibias of 450-day-old mice.²⁴ During the first 40 days, bone mineral density increased, as did polar moment of inertia and elastic modulus; however, cortical porosity decreased. In these mice tibias, mineral-to-matrix ratio increased from 12% of the 450-day value at 1 day to 65% of that value by day 30. In contrast, collagen crosslinking decreased from 112% (day 1) to 80% (day 30). in this mouse model, crystallinity remained essentially unchanged. In a similar study, male C57BL/6 mice, with and without exercise, were evaluated with co-localized Raman and nanoindentation using linear regressions and non-linear multidimensional visualization.²⁵ Based on univariate analysis, material properties of 4- and 5-month-old specimens were not significantly different but for a marginal increase in collagen cross-linking ratio and a marginal decrease in hardness at 4 months. Mineral-to-matrix ratio, carbonate-to-phosphate ratio and crystallinity were significantly greater in 19-month-old (skeletally mature, old) bones compared to the values in skeletally mature young bones (4- and 5-month-old combined). Crosslink ratio and nanoindentation measurements were not significantly different between the two age groups. Exercise had no significant effect on measures by either modality. In linear regressions of single Raman metrics with mechanical properties across all age groups, plasticity index was significantly and positively correlated to both carbonate-to-phosphate and mineral-to-matrix ratios. Bone modulus positively correlated with crystallinity. Bone modulus, hardness and creep viscosity negatively correlated with cross-linking ratio. In the 4–5 month age group, mineral-to-matrix ratio positively contributed to the modulus, while cross-link ratio had a significant negative contribution. Cross-link ratio was the only Raman metric that significantly correlated to hardness (negative effect). Carbonate-to-phosphate ratio positively contributed to plasticity index. In the 19-month-old bones, only mineral-to-matrix ratio contributed to mechanical properties. Unfortunately, similar studies have not been reported in other mouse models, or with other bones, except for those that served as controls for transgenic animals.²⁵

In a study²⁶ examining areas that mineralized over a period of 386 days compared to interstitial areas mineralized over 550 days, fluorochrome-labeled rabbit osteons were reported to rapidly mineralize between days 1 and 18, reaching 67% of interstitial bone levels. This increase was followed by a slower, more progressive accumulation of mineral up to day 350. By day 351, a plateau was reached. Carbonate-to-protein ratio similarly increased rapidly during the first 18 days, reaching 73% of interstitial bone levels and plateauing by 315 days.

Non-human primates, such as baboons, provide insight into Haversian remodeling changes bone tissue with age. In an FTIRI, Raman and mechanical study of baboons from newborn to 32 years of age, ^27,28 consistent systematic variations in bone properties were found as a function of tissue age in osteons. The patterns observed were independent of animal age and positively correlated with bone tissue elastic behavior measured by nanoindentation. The mineral-to-matrix ratio was correlated with the animal age in both old (interstitial) and newly-formed bone tissue; carbonate-to-phosphate ratio and crystallinity increased with animal and osteonal age and porosity, then reached a plateau.

In humans, loss of cancellous bone begins during the third decade of life in both men and women and is accelerated with menopause^29,30 later on, cortical bone loss occurs after menopause or sex steroid-deficiency in aging men and is associated with increased porosity.³¹ Age-related changes in human trabecular bone architecture include decreased trabecular numbers, thickness, and connectivity with decreases in thickness being larger in men than in women;³² women losing most bone by decreases in trabecular number and men by trabecular thinning.³⁰ While acquiring human biopsies for age-dependent studies is difficult, they are sometimes available as controls from other studies. We have examined iliac crest biopsies from individuals with juvenile osteoporosis³³ and post-menopausal women with fragility fractures. ^22,34 The female non-fractured controls from these two studies are shown as a function of age for some of the FTIRI parameters for cortical and cancellous bone in Figure 3A and 3B. Seen here and in general agreement with the findings in other species, ^27,28 with increasing chronologic age, mineral-to-matrix ratio, carbonate-to-phosphate ratio and collagen cross-link ratio increased, crystallinity increased and then reached a plateau while acid phosphate substitution decreased. A study comparing newly formed bone in healthy aging and postmenopausal osteoporosis showed that compositional changes depended on subject age, tissue age and health³⁵. These results in healthy iliac crest biopsies suggested that the kinetics of maturation could be altered with advancing age. Both subject age and tissue age must be taken into account when examining how diseases affect these parameters.

Figure 3.

Correlations between age and FTIRI variable for (A) cortical and (B) cancellous bone were calculated using data for healthy women using GraphPad Prism (version 7.0). Only significant correlations are presented. In some cases, specific parameters, e.g. acid phosphate substitution and carbonate/phosphate, were not measured in all studies due to instrument limitations. Solid lines represent best-fit correlations and dotted lines 95% confidence limits.

Bone quality changes in disease

Bone is a heterogeneous tissue that is constantly in flux, reflected by the activities of the three major bone cell types: osteoblast, osteocyte and osteoclast. When the interactions among these cell types are disrupted due to aging, injury and environmental changes, primary and secondary bone diseases may develop. In this review, we will consider a few examples of these diseases to demonstrate the various changes that affect the tissues’ mechanical performance.

Osteoporosis

Osteoporosis, associated with an increased risk of non-traumatic (fragility) fractures, occurs in both women and in men. High resolution NMR, for example, detailed differences between male and female osteoporosis³⁶ with elderly women having lower apparent bone volume fraction (BV/TV) and trabecular thickness (Tb.Th) than elderly men; elderly osteopenic women (based on BMD) had lower BV/TV and Tb.Th than elderly normal women. In a study comparing mechanical strength of biopsies obtained from the femoral neck of hip fracture patients during replacement surgery, ultimate stress of trabecular bone was significantly higher in men than in women.³⁷ Few studies exist, however, comparing bone quality in men and women with osteoporosis. Existing studies show that vertebral bone samples of men have higher values of BV/TV and Tb.Th compared to women³⁸ and cortical porosity measured by high resolution peripheral quantitative CT (HR-pQCT) is higher in radius and tibia of women than in men.³⁹

In contrast, there are numerous studies of bone quality in osteoporotic women, many (not included here, but recently reviewed⁴⁰) consider the effects of pharmaceutical interventions on bone quality measurements. Bone mineral density distribution (BMDD) based on calcium levels, is lower in iliac crest biopsies⁴¹ and in vertebral bodies⁴²of women with fractures compared to historic controls; tissue from vertebrae with acute compression fractures had a larger variation in matrix mineralization depending on the stage of repair. Similarly, BMDD of cancellous bone in premenopausal women with idiopathic osteoporosis was decreased to lower values than age-matched controls.⁴³

In age-matched, fracture-free controls, FTIR microscopy (FTIRM) found the bone in iliac crest biopsies of individuals with fractures had greater crystallinity than in controls.⁴⁴ In a study of 54 iliac crest biopsies (32 from individuals with fractures, 22 without) who had significantly different spine but not hip BMDs and ranged in age from 30 to 83 years, using FTIRI, the metrics associated with fracture risk in a linear regression model were increased cortical and cancellous collagen maturity, increased cortical mineral/matrix ratio and cancellous crystallinity.³⁴ When matched for both BMD and age, a study of 60 pairs of iliac crest biopsies from women with and without fractures found most of these predictors (mineral-to-matrix ratio, collagen maturity, crystallinity) which increase with age were no longer significant, but the carbonate-to-phosphate ratio in both cancellous and cortical bone, and increased heterogeneity of collagen maturity for cancellous bone were significant.²²

Raman analysis also showed femoral trabecular bone and iliac crest cortical bone in women with fractures had a higher carbonate-to -amide I area ratio and carbonate-to-phosphate ratio, respectively⁴⁵. Mineral-to-matrix ratio (based on Raman) and indentation modulus were reduced in cancellous bone of females with osteoporosis.⁴⁶ Collagen pyridinoline content was elevated in osteoporotic cases as was lipid content, most other Raman changes were associated with aging.³⁵ MicroCT microstructural changes in osteoporotic bone as compared with age-matched controls include decreased bone volume, trabecular number, and connectivity density and increased of trabecular separation^47,48; however, these changes are dependent on resolution of the image.⁴⁹ Increased cortical porosity in osteoporotic tissue is also noted by microCT.⁵⁰

There are fewer studies of quality changes in osteoporotic men, however, an early investigation of 108 men (mean age 52 years) with lumbar osteopenia, based on spinal radiographs found there was at least one vertebral crush fracture in 62 patients, and none in the other. Trabecular separation (Tb.Sp) was increased in the fracture group; whereas trabecular number (Tb.N) decreased.⁵¹ Logistic regression showed that spine BMD, BV/TV and all architectural parameters were significant predictors of multiple vertebral fractures in men. Another investigation based on BMDD and histomorphometry measurements of 25 otherwise healthy males with fragility fractures contrasted with non-fractured historical controls reported a paucity of osteoblasts and osteoclasts on most bone surfaces, a shift to lower mineralization densities for cancellous bone values.⁵²

In male rats with osteoporosis induced by orchiectomy (ORX) analyzed by histomorphometry at 2, 4, 8, and 16 weeks post-ORX, bone mineral content (BMC) was reduced at 16 weeks. Trabecular bone volume was significantly decreased from the 4th week.⁵³ Male rats with osteoporosis caused by spinal cord injury had decreased mineral-to-matrix ratio measured by Raman spectroscopy in both femora and humeri compared to a non-injured control, with no changes in carbonate substitution.⁵⁴

Secondary Osteoporosis

In addition to osteoporosis associated with aging, discussed above, secondary causes can lead to increased fracture risk. This secondary osteoporosis may be induced by glucocorticoids (steroid-induced osteoporosis), thiazide diuretics, opioids, anti-viral therapies and numerous other drugs routinely used in management of non-skeletal conditions⁵⁵. Conditions such as hemophilia⁵⁶ and prolonged bed rest can also result in secondary osteoporosis. We have reviewed the changes in bone quality caused by these drugs elsewhere⁴⁰; here we present two examples, the most common form of secondary osteoporosis, that induced by steroids⁵⁷ and that associated with treatment for chronic obstructive pulmonary disease (COPD).

Treatment with steroids results in marked increases in fracture incidence without significant changes in bone quantity (BMD) and thus provides an indication of the impact of changes in bone quality. Osteoporotic fractures occur in 30–50% of patients on chronic glucocorticoid therapy.⁵⁸ Bones of prednisolone-treated mice had the number of osteocyte lacunae increased; the elastic modulus around the lacunae decreased and an area of hypomineralized bone surrounded the lacunae contrasted with placebo-treated controls.⁵⁹ In the treated-mice vertebrae trabecular BV/TV and cortical thickness Ct.Th decreased. Histomorphometry confirmed the decrease in bone volume fraction and showed reduced trabecular thickness, increased trabecular separation and decreased mineral apposition rate along with increased osteocyte apoptosis. There was also reduced bone formation and increased bone resorption. Raman spectroscopy showed a reduced mineral-to-matrix ratio. In a related mouse model of Cushing’s syndrome, which produces excessive endogenous steroids and develops osteoporosis, using in situ X-ray nanomechanical imaging, synchrotron micro-computed tomography and scanning electron microscopy investigations, a recent study found that the steroid-enriched animals, as compared to wildtype, lacked a uniform cortex, with their posterior cortex being thinner compared to the anterior.⁶⁰ The affected mice cortices had numerous localized cement lines surrounding low mineralized tissue near cavities, in contrast with uniform density in the WT. BMDD data showed the mean calcium content was reduced in endosteal areas, and the tissue heterogeneity was decreased in these areas in the affected mice. WT bone contained a homogeneous network of condensed canals and osteocyte lacunae across the cortical bone while the affected mice had most of these spaces replaced with cavities. Tissue-level stiffness was decreased, maximum strain increased and breaking stress was reduced in bones of the affected mice compared to WT. Collagen fibril deformation was also affected in this model.

Few studies of glucocorticoid-induced osteoporosis in humans include bone quality measurements. In one study⁶¹ men taking glucocorticoids were subdivided into those with and without vertebral fractures and high resolution CT data combined with finite element methods was used to predict fracture risk; results surpassed those based on BMD. HR-pQCT was also used to characterize changes in bone architecture in women with systemic lupus erythematosus (SLE) treated long-term with glucocorticoids. Cortical thinning and increased cortical porosity were the features of longitudinal microstructural deterioration in treated individuals.⁶²

In addition to lupus and rheumatoid arthritis, asthma and other conditions in which glucocorticoids are used for extended periods, fracture risk is increased in individuals with chronic obstructive pulmonary disease (COPD) taking inhaled steroids. While life-style features leading to COPD also affect bone turnover and fracture risk,⁶³ there is a single study⁶⁴ comparing bone quality in newly formed and older bone of individuals with COPD who did and did not sustain fragility fractures based on Raman spectroscopy. Interestingly, no differences were noted between the glucocorticoid-treated individuals with COPD and those who were not treated with glucocorticoids. Those individuals with COPD who had fragility fractures, however, had reduced nanoporosity and elevated mineral-to-matrix ratios and pyridinoline values relative to those without fragility fractures in areas formed two weeks prior to the biopsy. Unfortunately, there are not yet similarly detailed studies on the other agents that cause secondary osteoporosis.

Osteomalacia

Osteomalacia in adults and rickets in children describes conditions in which bones and calcified cartilage are soft (insufficiently mineralized). Generally associated with deficiencies in vitamin D, its metabolites or its receptor, osteomalacia and rickets may also reflect problems with calcium⁶⁵ or phosphate⁶⁶ handling. Osteomalacia may be diagnosed as osteoporosis because of an increased fracture incidence, but is generally characterized by defective bone mineralization, hypophosphatemia or hypocalcemia. Bone quality in rodents with rickets and osteomalacia was characterized by x-ray diffraction in 1991.⁶⁷

Later, in a small study, we used FTIR microscopy to study human tissues comparing iliac crest biopsies from seven women without apparent bone disease and eleven women diagnosed with osteomalacia.⁶⁸ No significant differences in cortical parameters were noted between the two groups of biopsies. The mineral-to-matrix ratio was lower in trabecular regions than in controls. Mineral crystallinity tended to be decreased in osteomalacic trabecular bone, pointing to a sub-optimal mineralization at the bone surface. Another study based on degree of mineralization (BMDD) in 13 individuals reported both lower microhardness and BMDD than controls.⁶⁹ Microhardness of the non-mineralized osteoid tissue was 3-fold lower than the total microhardness in the adjacent calcified matrix located in its vicinity. Similarly, histomorphometry of biopsies from adults with hypophosphatasia (due to inactivating mutations in alkaline phosphatase) showed decreased osteoid mineralization and an increased number of osteoblasts, compared both to controls and other individuals with osteomalacia, based on histomorphometry.⁷⁰

X-ray diffraction investigations of mouse models of vitamin D-resistant hypophosphatemic rickets (HYP), showed the bone tissue contained larger crystals in the less mineralized areas at early⁶⁷ stages of development; this result was confirmed for older animals by FTIR microscopy. Pregnant HYP mice were found to have increased porosity by nano-CT.⁷¹ These pregnant HYP mice maintained calcium and vitamin D levels by increasing PTH and activating osteoclasts. Raman analyses of these mice showed increased carbonate-to-phosphate ratio in both the baseline and lactating HYP mouse bones.⁷² Bones in another model with increased vitamin D activity, a knockout of the Ca-transporter TRPV5 (transient receptor potential, vanilloid 5) Ca(2+) channel, normally expressed in kidney and osteoclasts, had decreased bone thickness by microCT.⁷³ These studies are a few examples of osteomalacia and rickets in which bone quality is known to be affected by the improper handling of phosphate and calcium transport. Readers are referred to recent papers on FGFR1, FGF23, PTH, and KLOTHO for more details of other related systems.^74,75

Osteogenesis Imperfecta

Osteogenesis imperfecta, also known as brittle bone disease, is a group of genetic abnormalities resulting in the inappropriate synthesis, folding, modification and transport of type I collagen, and hence affects all tissues containing this collagen, especially bones and teeth.⁷⁶ These mutations cause the bones of individuals with OI to fracture, with typical brittle behavior (Fig. 4), such that after the bone begins to yield, there is minimal post-yield deflection. Bone quality in mouse models of OI has been shown to be abnormal by almost all the methodologies mentioned in this review, including x-ray diffraction,⁷⁷ infrared imaging,^{78,79, 80,81}, BMDD measurements,^82,83 Raman spectroscopy,^84,85,86,87 atomic force microscopy (AFM),⁸⁷ second harmonic generation⁸⁸ and mechanical testing.^85,89,90 Of interest, most models, and the majority of the human forms of OI are hypermineralized, perhaps, because of the deficiency in type I collagen. Table 2 summarizes bone quality in several well-characterized mouse models studied by one or more of these techniques. Data on bone quality exist for the mild to severe forms of OI but not yet for models of the perinatal lethal form (Type II OI). Bones from humans with OI have been characterized less frequently; limited human data exist for type II OI based on similar analytical methods (Table 3).^94–97 The mouse and human data are in agreement, typically finding hypermineralization, improper mineral alignment with collagen, and altered collagen fibril properties. The brittle behavior appears to be both a function of the abnormal collagen and related changes in the matrix leading to abnormal mineralization patterns.

Figure 4.

Typical stress-strain curves from 3-point-bending tests of male and female aga2/+ mice and their wild-type controls. The aga2/+ mice curves end abruptly, showing no post-yield behavior, a characteristic of brittle bone tissue.

Table 2.

Bone Quality Studies in Mouse Models of Osteogenesis Imperfecta

OI Model	Defect	Techniques Used	Characteristic
Type I	Autosomal dominant	Mechanical tests89,91 pCT89 Raman89 SAXS92 Microhardness92 qBEI92	Reduced mechanical properties91 Sex-dependent composition89 Reduced hydroxyproline content89 Increased hardness92
+/oim
Type III	Autosomal dominant; Qualitative-splicing mutation; glycine substitutions	XRD77 FTIRI78,79,80,81,90 SHG88 Raman84,89 Mechanical tests90	Crystals smaller77,80 Hypermineralized78,79,80,81 Decreased CO3/PO478,79,80 Increased collagen x-links78,79,80 Crystals not well aligned with collagen84; sex-dependent composition89; reduced hydroxyproline content89; reduced material properties90 Poorly aligned collagen 81 Decreased Ca content Brittle90
oim/oim
G610C (Amish)
Type IV	Autosomal recessive; Qualitative	Raman86 microCT86 AFM87	Reduced cortical thickness86 Less fracture resistant up to 6 mo. of age; mineral/matrix reduced at 6 mo86; Collagen d- spacing altered87
Brtl
Type VI	Autosomal recessive	Micro-CT, FTIRI, histomorphometry93 SAXS and BMDD82	Reduced trabecular bone volume; accumulation of unmineralized bone; increased mineral/matrix 93 increased Ca content, altered particle alignment with collagen82
PEDF−/−
Fro/fro		FTIRI78 uCT78	Decreased Min/Mat78; decreased cross-linking78
Type VII  Crtap−/−	Autosomal recessive Defective 3′-OH- prolylase complex	Micro-CT85,81 Raman 85 Mechanical tests 85 FTIRI81 Backscattered electron imaging 83	Hypermineralization; increased mineral/matrix; decreased crystal size; sex-dependent differences in material properties81

Table 3.

Bone Quality Changes in Human Osteogenesis Imperfecta (OI) Bone Tissue

OI Type	Mutation	Technique	Observation vs. Control
OI Type I	qualitative and quantitative combined type I collagen mutations	qBEI, SAXS, BMDD90 Raman (newly formed tissue)91	No change in particle width; OI tissue denser packing of particles 90 Relative GAG content reduced, nano-porosity reduced, PYD content increased in quantitative group Crystallite length reduced – both91
OI Type II	perinatal lethal – type I collagen mutations	X-ray diffraction92	Small crystals 92
OI Type VI	mutation in SERPINF1 leading to loss-of- function of pigment epithelium-derived factor (PEDF)	BMDD, SAXS	Highly mineralized matrix with low mineral content 93
OI Type VII	hypomorphic mutations with CRTAP expression	BMDD	Increased mineralization81
OI Type VIII	null mutations in P3H1, encoding prolyl 3- hydroxylase 1.	Histomorphometry, qBEI, BMDD	Decreased cortical width; thin trabeculae; patches of increased osteoid; hypermineralization 94

Osteopetrosis

Failure of bone to be remodeled results in retention of calcified cartilage along with characteristics of older bone, and frequent fractures in the condition known as osteopetrosis. As reviewed elsewhere, osteopetrosis is generally associated with a defect in one or more osteoclastic resorption mechanisms.⁹⁸ Similar to most of the conditions mentioned in this review, these abnormalities result in altered bone quality and increased fracture risk. The condition may be caused by genetic abnormalities in regulators of osteoclast activity or by excessive suppression of bone turnover.⁹⁹ Bone quality in osteopetrosis was characterized in the toothless rat and the osteopetrotic mouse by x-ray diffraction^100,101 and FTIRI;^101,102 the latter showing microhardness was increased in osteopetrotic bone, the former demonstrated the persistence of small crystals. The c-src knockout mice, which also develop osteopetrosis, had increased numbers of microcracks; however, adult animals with this knockout were not mechanically weaker than control mice when tested in three-point bending.¹⁰³ Bone tissue of another knockout mouse with osteopetrosis, Fos^-/-, had a more homogeneous matrix based on BMDD due to the persistence of calcified cartilage.¹⁰⁴ Mice with carbonic anhydrase deficiency, also critical for osteoclast-mediated remodeling, provide another model of osteopetrosis. These animals have smaller bones, and the normalized cortical bone volume was similar to that in wildtype bones; however, they had significant metaphyseal widening of the tibial plateau. Trabecular BV/TV was increased almost 50% relative to WT. Histomorphometry showed significant decreases in bone formation rate, and increased number of osteoclasts.¹⁰⁵ In humans, a case report of two non-related male and female individuals with osteopetrosis, HR-pQCT found increased density, increased cortical thickness, high trabecular number and thickness, increased BV/TV. Relative to unaffected controls, the tissue was extremely heterogeneous, with islets of dense bone interposed with areas of normal density.¹⁰⁶ Thus, the fractures in osteopetrotic bone seem to be related to the presence of calcified cartilage, a failure to correct micro-cracks, and, similar to OI, but present for different reasons, smaller crystals.

Conclusions

Age-dependent changes in bone quality in animals and humans and the alterations due to bone diseases have been discussed. As emphasized in this review, rodents are the most commonly-used animals to investigate aging and bone disease. Our ability to manipulate the mouse genome and associated molecular signaling pathways has advanced our understanding of the regulation of bone mass. However, as reviewed elsewhere¹⁰⁷, key differences exist in the rodent skeleton when compared to humans. In particular, rodents lack osteonal remodeling, continue longitudinal bone growth after sexual maturity, and do not undergo a true menopause. Nevertheless, advantages and the similarities in age-related bone loss make rodents a good model to study changes in bone quality.

Alterations in bone quality and tissue mechanical properties at both the micro- and macro- levels have been correlated in a variety of studies, as reviewed recently.¹⁰⁸ Hence understanding how bone quality parameters are modified by both aging and disease should lead to improved treatments to modify bone tissue mechanical properties associated with these conditions, and a reduction in weakened bones and fractures. We have noted how both increases and decreases in the average size of the mineral crystals can lead to fractures; and, how decreased and increased cross-linking of collagen can have similar effects. Changes in micro-architecture and chemical composition of the matrix and mineral are important. Hopefully this review has provided the reader with insight into the importance of understanding these variables and their alterations in some rare and common conditions.