Osteoporosis International

, 20:1461

Is bone quality associated with collagen age?


    • Nordic Bioscience
  • K. Henriksen
    • Nordic Bioscience
  • I. Byrjalsen
    • Nordic Bioscience
  • P. Qvist
    • Nordic Bioscience
  • S. H. Madsen
    • Nordic Bioscience
  • P. Garnero
    • CCBR-Synarc
    • INSERM
  • M. A. Karsdal
    • Nordic Bioscience

DOI: 10.1007/s00198-009-0904-3

Cite this article as:
Leeming, D.J., Henriksen, K., Byrjalsen, I. et al. Osteoporos Int (2009) 20: 1461. doi:10.1007/s00198-009-0904-3


The World Health Organization defines osteoporosis as a systemic disease characterized by decreased bone tissue mass and microarchitectural deterioration, resulting in increased fracture risk. Since this statement, a significant amount of data has been generated showing that these two factors do not cover all risks for fracture. Other independent clinical factors, such as age, as well as aspects related to qualitative changes in bone tissue, are believed to play an important role. The term “bone quality” encompasses a variety of parameters, including the extent of mineralization, the number and distribution of microfractures, the extent of osteocyte apoptosis, and changes in collagen properties. The major mechanism controlling these qualitative factors is bone remodeling, which is tightly regulated by the osteoclast/osteoblast activity. We focus on the relationship between bone remodeling and changes in collagen properties, especially the extent of one posttranslational modification. In vivo, measurements of the ratio between native and isomerized C-telopeptides of type I collagen provides an index of bone matrix age. Current preclinical and clinical studies suggests that this urinary ratio provides information about bone strength and fracture risk independent of bone mineral density and that it responds differently according to the type of therapy regulating bone turnover.


AntiresorptivesBone qualityIsomerizationOsteoporosisType I collagen


In the description of osteoporosis presented in 1990 at the Consensus Development Conference, the disease was defined as a decrease in bone mineral density (BMD) and a microarchitectural deterioration of bone tissue leading to an increase in bone fragility and susceptibility to fracture [1]. Bone mineral density values at least 2.5 standard deviations below the mean levels in young adults are used to identify persons with osteoporosis [2]. Although numerous studies have shown that measurement of BMD alone is not sensitive enough to identify the majority of women who will sustain a fracture, the definition of osteoporosis has not changed [38]. Recently, a model combining BMD with clinical risk factors was proposed to improve the assessment of fracture probability [9]. This model, however, did not include factors related to bone remodeling and bone matrix quality.

Bone mineral density is currently the most important determinant of fracture risk [10, 11]. Yet, several studies have shown that up to 50% of persons who experience fracture have a BMD value above the level of 2.5 standard deviations below the reference-population mean, or a T score ≥2.5 [3, 4, 12]. Older persons can have up to a tenfold increased 10-year fracture risk in comparison with younger individuals with the same BMD [13]: For example, an 80-year-old person with the same BMD as a 50-year-old one has a fivefold higher risk of hip fracture [13]. Thus, in individuals with comparable BMD, fracture risks are not the same. In addition, more than 50% of all incident fractures occur in women with osteopenia, as defined by a BMD T score ≥2.5 and ≤1 [14]; at-risk women in this group will not be detected by applying the World Health Organization BMD definition of osteoporosis.

The bisphosphonate alendronate increases hip and spine BMD, leading to a lowering of the risk of vertebral fracture [5], nevertheless a meta-analysis showed that change in spine BMD accounted for only 16% of the vertebral fracture reduction [6]. Studies of the selective estrogen receptor modulator (SERM) raloxifene demonstrated that an increase in femoral neck BMD accounted for only 4% of vertebral fracture risk reduction [7]. Calcitonin treatment results in only a 1.2% increase in BMD of the lumbar spine, but it is associated with a significant 33% reduction in vertebral fractures [15]. Finally, changes in BMD associated with teriparatide (parathyroid hormone) account for less than 40% of its effect in reducing vertebral fracture risk [16]. Together, these data demonstrate that changes in BMD with osteoporosis treatments only partially explain fracture risk reductions and that additional independent factors contribute to the clinical efficacy of these therapies.

Bone quality has been extensively discussed in the literature [1721]. However, the definition of bone quality remains imprecise and in vivo assessment of its various components is nearly impossible to perform, as such assessment relies mainly on invasive techniques, which consequently are not generally applicable in clinical trials. In this review, we use the term “bone quality” as the quality of the bone matrix thus comprising mineralization, posttranslational changes in the collagen molecule that occur with age, and microscopic tissue damage which all are parameters affected by bone remodeling. Recent developments in the field of biochemical marker research have indicated that age-related changes in bone matrix molecules are associated with bone fracture resistance independent of BMD and may thus allow for the monitoring of some aspects of bone matrix quality in a manner that is noninvasive [17, 18, 2229]. In this review, we emphasize the likely connection between bone matrix quality and bone remodeling, highlighting the importance of bone matrix age. Because determination of the age-related α to β isomerization in the C-telopeptide of type I collagen (αα/ββ CTX ratio) is currently the only method available for indexing mean bone age noninvasive [22, 23, 27], we discuss in detail the association between this index, bone matrix quality, and bone remodeling.

What is bone matrix quality?

Bone matrix is composed of mineral and organic matrix, mostly type I collagen. Bone matrix quality thus comprises the organization and extent of mineralization, structural changes in type I collagen molecules, and microscopic tissue damage occurring in this tissue. Central events regulating these properties of bone matrix quality are controlled by the rate of bone remodeling [21, 3033]. The mineralization of the bone matrix is dependent on the age of the bones, with older bones being more mineralized because of alterations in bone-remodeling rates [34]. The mode of action needed to increase the bone matrix strength is dependent on the existing level of mineralization of a given bone. If the mineralization is low initially, an increase in mineralization would most likely benefit the bone strength in contrast to an already highly mineralized bone where an increase in mineralization most likely would decrease the strength of the bone [3436]. The specific effects of changes in bone mineralization are unknown, although a study has indicated that microcracks preferentially occur in the highly mineralized areas of bones, i.e., the old bone [37]. Therefore, monitoring of the changes in bone mineralization, along with other parameters related to bone turnover, is essential to fully characterize bone matrix quality.

Also, microcracks accumulate in bones with slow remodeling, i.e., when the age of the bones increases [3840]. This accumulation is associated with loss of structural properties such as bone stiffness and energy absorption [41, 42]. Although we are aware of the possible effects of number of microcracks and degree of mineralization on bone matrix quality, further discussion of these properties is beyond the scope of this review in which we focus on the effect of posttranslational changes on bone matrix quality.

Posttranslational changes with age

Proteins in bone matrix, including type I collagen, undergo several nonenzymatic and enzymatic posttranslational modifications with time. The rate at which these transformations occur and eventually accumulate in bone tissue is controlled by the level of bone remodeling [31, 43]. This continuous remodeling of bone involves the function of osteoclasts, which act to achieve a coordinated and balanced resorption of aged bone, and of osteoblasts, responsible for sufficient formation of new bone, occurring in a localized, coordinated, and sequential manner referred to as coupling [44]. Secondary to inhibition of bone resorption, bone formation is decreased to a large degree by most antiresorptives [44]. This leads to the generation of fewer new bone proteins, and as bone formation is inhibited, the age profile of the bone also increases. The result is accumulation of age-related posttranslational modifications, such as isomerization (αα/ββ CTX ratio) and formation of advanced glycation end products (AGEs) and crosslinks. These posttranslational modifications have been investigated in bone samples and systemically in urine samples, as they may be important determinants of bone strength and matrix quality [17, 18, 22, 25, 45, 46]. Crosslinks, such as pyridinoline (PYD) and deoxypyridinoline (DPD), occur during the maturation of the bone collagen matrix (Fig. 1c). AGEs are presented as protein–protein crosslinks (Fig. 1d) and the concentration of the AGE pentosidine in trabecular bone of the vertebrae has been shown to contribute negatively to mechanical properties and thus quality of the bone matrix [22, 4749]. Because AGE, PYD, and DPD are not as bone specific as the release of CTX, which is generated by cathepsin K [50], and that the accurate measurement of AGEs in body fluid remains challenging, we focus on the urinary ratio of αα/ββ CTX, which currently seems to be the most reliable marker to provide a noninvasive assessment of bone matrix quality.
Fig. 1

Age-related changes in collagen type I: a intramolecular crosslinks and location of the CTX I epitope in the C-terminal end of the α1-chain (black square), b enzymatic isomerization of the aspartic acid (Asp) in the CTX epitope, c intermolecular crosslinks mediated by lysyl oxidase, and d advanced glycation end products in lysine, hydroxylysine, or arginine

Proteins containing an aspartate (D), asparagine (N), glutamate (E), or glutamine (Q) residue linked to a low molecular weight amino acid, such as glycine (G), can undergo a spontaneous nonenzymatic isomerization [51]. This isomerization introduces a kink in the conformation of the molecule, as the peptide backbone is redirected from the α-carboxyl group in the native newly synthesized form to the side chain β-carboxyl [52]. The alpha 1 chain of type I collagen undergoes a β-isomerization in the DG motif within the CTX sequence (1207EKAHDDGR1214) of its C-terminal telopeptide (Fig. 1a, b). This transformation has been shown to occur in vitro when bone matrix is incubated at 37°C and has also been documented in vivo in humans [26, 52]. The CTX neoepitope is released during osteoclastic bone resorption by the action of cathepsin K (Fig. 2a) and can be detected by an immunoassay. The native form ααCTX, composed of two crosslinked αCTX peptides (one from each of the two alpha I chains of a collagen molecule), originates from newly synthesized collagen type I, whereas ββCTX is released from aged collagen type I. Thus, the ratio between αα and ββ CTX gives an indication of the bone matrix age, maturation, and possibly quality (Fig. 2b) [51]. As biochemical markers are measured systemically, the ratio between αα and ββ CTX represents the sum of all sites where bone is being remodeled. It has been shown that the mean age of bone assessed by αα/ββ CTX is lower in bone that is highly remodeled (trabecular bone) versus higher in less remodeled bone (cortical bone) [53]. Ex vivo experiments using bovine bone tissue or human vertebral bodies have also shown an association between the extent of type I collagen isomerization and mechanical properties, independent of BMD [25].
Fig. 2

Schematic overview of the bone collagen age profile measured as the ratio between ααCTX and ββCTX. a The osteoclast-specific Cat K-generated CTX epitope located in the C-terminal telopeptide of intact collagen type I exists in two isoforms: the α and β forms. The endogenous age profile is reflected in the isomer composition. b ααCTX and ββCTX fragments reflecting the endogenous bone collagen age are released from bone and can be assessed by enzyme-linked immunosorbent assay. The ratio between ααCTX and ββCTX in urine samples reflects the mean bone collagen age, which may be regarded as an index of bone matrix quality. c Theoretical bone age profiles in healthy individuals, patients with bone remodeling disorders, and patients during antiresorptive treatment

Bone remodeling rates control aging of the bone matrix

To maintain optimal quality of bone tissue, continuous remodeling of the bone matrix is essential, and the rate of remodeling controls the age of the bone matrix [17, 21, 22, 3033, 43]. The important contribution of bone remodeling as an independent determinant of fracture risk in both untreated and antiresorptive-treated subjects has been demonstrated using systemic biochemical markers [7, 8, 54, 55]. Naturally occurring, rare genetic disorders of bone turnover also have revealed the effects of disrupted bone remodeling on fracture risk and bone matrix quality as changes in bone remodeling cause local alterations in bone tissue age (Fig. 2c). Osteopetrosis and pycnodysostosis are characterized by absent or low osteoclastic bone resorption, high bone mass, and poor bone matrix quality, leading to increased fracture risk [56, 57]. Conversely, in sclerotic diseases (e.g., sclerosteosis and Van Buchem disease), which result primarily from increased bone formation, the phenotype is associated with improved bone strength and lower fracture risk [5860]. Similar findings have been reported in patients with autosomal dominant osteopetrosis type I; a high bone mass phenotype [6163] associated with increased bone strength [63, 64]. In Paget’s disease patients have a locally high remodeling rate with increased fracture risk thus impaired bone matrix quality [65]. These rare genetic disorders suggest that the expected association between high bone mass and improved bone strength is seen only when a certain level of osteoclastic bone resorption is maintained.

In summary, alteration of bone remodeling, particularly bone resorption, leads to pathological conditions characterized by alterations in bone matrix quality parameters, such as mineralization, and posttranslational modifications all suggested to be associated with bone quality. As a consequence, it increased fracture risk. This underlines the need for monitoring of bone remodeling and changes in bone matrix quality when assessing the effects of novel interventions on fracture risk.

Effects of osteoporotic treatments on bone matrix quality

Treatments for osteoporosis are targeted at reducing fracture risk by improving bone strength. Most available treatments target the resorptive capacity of the osteoclasts. Because bone formation is tightly coupled to bone resorption [44, 66, 67], most antiresorptive drugs decrease overall bone turnover [55]. In addition to eliminating the pathological excess of bone resorption, some antiresorptive treatments also alter targeted bone remodeling [44]. Targeted remodeling is the process ensuring that osteoclasts remove bone matrix where microcracks have occurred. The process involves osteocyte apoptosis in response to microdamage, which then in turn leads to activation of osteoclastic bone resorption, leading to removal of the damaged matrix, and finally new bone formation by the osteoblasts [29]. The outcome of such treatments would be an increase in the age of the bone matrix, eventually leading to accumulation of biochemical alterations in the collagen matrix as well as other aged-related factors [22, 23, 26, 46]. Such effects have raised major concerns within the field of bone research, especially with respect to the massive suppression of bone remodeling, as seen with potent osteoporotic drugs [20, 27, 68]. Correlation of the change in overall bone turnover, assessed by the αα/ββ CTX ratio, shows that bisphosphonates, which markedly reduce bone resorption and bone turnover, induce a decrease in the αα/ββ CTX ratio, indicating an increase in bone tissue age (Fig. 3). Drugs with more moderate effects on bone remodeling, such as SERM, hormone replacement therapy (HRT), and calcitonin, do not or vaguely modify bone tissue age [23, 27].
Fig. 3

Change in αα/ββ CTX ratio correlate to change in spinal BMD. BMD data have been adapted from phase III studies: PROOF study for calcitonin [15]; MORE study for SERM [73]; Women’s HOPE study for HRT [71, 72]; FIT study for alendronate [77, 78]; BONE study for ibandronate [76]. The αα/ββ CTX data have been adapted from two other treatment studies [23, 27]

The different antiresorptive therapies show evidence of varying impact on bone turnover and BMD, although their efficacy for the prevention of fracture, particularly vertebral fracture risk, is similar. SERM, HRT, and calcitonin exhibit a 25–50% reduction in bone turnover and 1–3% BMD increase (Fig. 4a, b), whereas potent bisphosphonates decrease bone resorption up to 70% of pretreatment levels leading to approximately 6–7% increase in BMD [6975]. Nevertheless, the reduction in fracture risk is similar among all these treatments [7680], and SERM, HRT, and calcitonin demonstrate antifracture efficacy approaching that obtained with the more potent antiresorptives [81]. Thus, data indicate that changes in BMD during antiresorptive therapy do not directly translate to proportional decrease in fracture risk. We believe that changes in bone collagen maturation (αα/ββ CTX), which is not captured by measurements of BMD, may also be involved. The bone matrix age decreases as bone turnover is decreased during antiresorptive treatments [23, 27] (Fig. 4c), indicating that bone matrix quality is compromised when using potent drugs such as bisphosphonates due to accumulation of aged bone matrix.
Fig. 4

Effect of various antiresorptives on a BMDspine increase and vertebral fracture risk reduction, b systemic bone turnover (bone resorption) and vertebral fracture risk reduction, c systemic αα/ββ CTX ratio and systemic bone turnover reduction (bone resorption). BMD, bone turnover, and fracture risk data have been adapted from phase III studies: PROOF study for calcitonin [15], MORE study for SERM [73, 97], Women’s HOPE study for HRT [71, 72], FIT study for alendronate [77, 78], and BONE study for ibandronate [76]. The αα/ββ CTX data have been adapted from two separate treatment studies [23, 27]

Preclinical data

A recent study in skeletally mature female beagle dogs compared treatment with clinically relevant doses of alendronate, risendronate, raloxifene, or vehicle for 1 year [45]. The bisphosphonates decreased the αα/ββ CTX ratio by 29–56% in vertebral bone, compared with vehicle. In contrast, raloxifene did not change the age of the collagen matrix. Interestingly, the rate of bone turnover was significantly correlated with the concentration of the αα/ββ CTX ratio as well as other age-related protein modifications with a higher rate of bone remodeling associated with lower bone collagen maturation. This finding indicates that administration of bisphosphonates leads to a more aged collagen profile in vertebral trabecular bone, compared with bone from SERM dogs which probably was due to the higher suppression of bone remodeling by the bisphosphonates. Other studies of the effect of SERM on bone matrix quality parameters confirm the absence of increased aging in both cortical and trabecular bone from femurs [82, 83]. Several preclinical histomorphometric investigations have also proven that bisphosphonates reduce bone turnover to a much further extent than SERM and HRT [84, 85]. In summary, preclinical data highlight the important finding that clinically relevant doses of drugs that markedly suppress bone remodeling lead to increased aging of the bone matrix.

Clinical data

A 2-year study of healthy postmenopausal women evaluated the effect of antiresorptive treatment on the systemic αα/ββ CTX ratio as a measure of bone age [23]. Participants were treated with bisphosphonates, raloxifene, or HRT up to 24 months. The bisphosphonates alendronate and ibandronate induced a significantly higher bone age profile than for SERM and HRT (38–52% vs. 3–15% reduction in αα/ββ CTX, respectively). These data indicate that the interventions have different effects on bone age, probably because of their different effects on bone remodeling. However, another study reported no change in the αα/ββ CTX ratio in patients treated with alendronate for 2 years [86], possibly because of a lesser reduction in bone remodeling. In a similar manner, the effect of oral calcitonin on the αα/ββ CTX ratio was assessed in postmenopausal women after 1 and 3 months of therapy [27]. Bone resorption was reduced by 30% but the αα/ββ CTX ratio was unchanged, suggesting that calcitonin does not affect the age profile of bone matrix. This is most likely due to the fact that calcitonin decreases bone resorption but either does not reduce bone formation, or shows a markedly lower suppression of bone formation than other antiresorptive agents [8789]. These data indicate that calcitonin does not eliminate the signaling between osteoclasts and osteoblasts which is referred to as “uncoupling” [90, 91]. Thus, bone formation is not largely decreased as seen with other drugs and the bone remodeling events remain.

The retrospective risedronate and alendronate cohort study included women with osteoporosis who received either risedronate or alendronate [92] carried out for comparison of fracture risk reduction in the two groups. In the risedronate group, the risk of nonvertebral and vertebral fractures was 18% and 43% lower, respectively, than in the alendronate group. A trend was observed as early as 3 months after the initiation of therapy; at 6 months, this difference was significant. Similar results were found in a meta-analysis of six trials of risedronate or alendronate therapy [93] in which the relative risk reduction for nonvertebral fractures was greater for risedronate than for alendronate. Although there are methodological limitations in the design of these analyses, the results suggest that risedronate may be more effective than alendronate in decreasing fracture risk, possibly because it reduces bone turnover to a lesser extent than alendronate and thereby preserves the bone quality properties of the matrix by a lesser increase in bone tissue age [77, 78, 9496].

Concluding remarks

Bone remodeling is a key event in the maintenance of bone tissue mass, age, and quality. Assessment of the different components of bone matrix quality is currently hindered by the lack of specific, precise, and, importantly, noninvasive tests [23, 27]. However, the preliminary data discussed here suggest that the use of the urinary αα/ββ CTX ratio can provide valuable information on the age, and thus the quality of bone matrix and on its changes during treatment. The relationship between changes in and absolute levels of the αα/ββ CTX ratio, on the one hand, and the incidence of fracture in women receiving osteoporotic treatments, on the other hand, should be investigated, as associations may differ from those observed in untreated persons. Monitoring changes in the αα/ββ CTX ratio may then become a useful biological tool to assess the potential detrimental effects of some therapies, as there has been a concern that sustained long-term suppression of bone turnover may lead to increased bone fragility by compromising some aspects of bone matrix quality. It is interesting to note the renewed interest in drugs that modulate bone remodeling toward a more “steady-state” instead of leading to over suppression. It seems that a vast reduction in bone turnover in order to achieve gain in BMD compromises the quality of bone.

We propose that there is a relationship between the αα/ββ CTX ratio and the strength of the bone matrix; however, additional data are still needed to demonstrate the causality of this relationship. Whether the αα/ββ CTX is a better marker than other urinary markers for evaluation of fracture risk reduction has yet to be evaluated in clinical studies with fractures as the primary endpoint. However, it has been demonstrated that the ratio between the two biochemical markers αα CTX and ββ CTX provides us with a different picture than observing one of the markers alone [23, 27]. It provides a mean ratio of the bone matrix age that is resorbed in contrast to other bone remodeling markers that rather are index’s of the mean bone turnover, which will affect the bone matrix age.

Data discussed here indicate that BMD alone does not explain fracture risk but that we should rather consider the effect of increased BMD in combination with the bone matrix quality to obtain a more precise prediction of who will fracture (Fig. 5a). The amount of bone mass that a patient has at a given time point in combination with knowledge about the age of the bone matrix could provide a better measure of fracture risk. We speculate that the optimal future treatments for bone loss require drugs that increase BMD while maintaining the rate of bone turnover thus bone age to preserve or increase the quality of bone during treatment as illustrated in Fig. 5b in contrast to current drugs that increase BMD at the expense of increasing the bone age due to decrease in the bone turnover. For the efficient development of novel osteoporosis drugs, monitoring of changes in BMD and bone remodeling as well as bone matrix quality should be considered, as these three factors are likely to provide important and complementary information on bone strength and fracture risk and to predict long-term safety.
Fig. 5

Association between BMD, bone age (αα/ββ CTX), bone matrix quality (Bone Q) and fracture risk (Fx). a Impact of BMD and bone matrix quality on fracture risk. By considering the two effects, it may be possible to improve prediction of Fx (BMD × Bone Q = Fx). With current treatments, it is speculated that when BMD goes up (+), the quality of bone goes down (−). b Comparison of current and possible future treatments for bone turnover disorders and their effect on BMD, bone age, and bone matrix quality. Current treatments increase BMD while lowering the bone age resulting in lowering bone matrix quality. Optimal future treatments should still increase BMD, however, without changing the bone age resulting in increase in bone matrix quality


We acknowledge the funding from the Danish “Ministry of Science, Technology and Science”.

Conflicts of interest

Leeming DJ, Henriksen K, Byrjalsen I, Qvist P, Madsen SH, and Karsdal MA are employees of Nordic Bioscience.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2009