BoneKEy-Osteovision | Meeting Reports

Meeting report from the 25th annual meeting of the American society for bone and mineral research

---

DOI:10.1138/20030112

Genetics of bone and mineral disorders Robert F. Klein, University of Oregon Health Sciences Center and Portland VA Medical Center, Portland, OR, USA

Now that sequencing of the human genome is essentially complete, the genetic underpinnings of a wide variety of bone and mineral disorders are steadily being revealed. Although little is known about the specific genes responsible for isolated hypoparathyroidism (IH), GCMB (the human ortholog of the glial cells missing gene in Drosophila) is an attractive candidate. GCMB protein is expressed predominantly, if not exclusively, in parathyroid cells and seems to be a switch for normal parathyroid gland development. Recently, a patient was described with IH that was the result of a homozygous intragenic deletion (i.e., a “knockout”) of GCMB. To follow-up on this provocative observation, the same group has now examined the prevalence of GCMB mutations in 25 patients with early-onset IH (17 affected subjects from 10 families and eight de novo patients) (). Variant nucleotide changes resulting in heterozygous missense mutations were identified at five different sites within exon 5 of GCMB. Subsequent in vitro transactivation studies indicated that two of these mutations (G203S and N315D) resulted in a 70% to 80% reduction in transactivation potential. Given the number of allelic variants of GCMB that seem to exist, further studies are underway to determine whether certain compound heterozygotes may be clinically relevant to the pathogenesis of IH.

The molecular defects responsible for the three most common forms of pseudohypoparathyroidism (PHP) — PHP type Ia (PHP Ia), pseudopseudohypoparathyroidism (PPHP), and PHP type Ib (PHP Ib) — seem to reside in GNAS1, the gene encoding the α-subunit of the stimulatory GTP binding protein (Gs). However, the development of one disorder or another depends on the sex of the parent transmitting the genetic defect. For example, PHP Ib develops only if the genetic defect is transmitted from a female carrier. Although most mammalian genes are expressed from both alleles, a small group of special genes is imprinted; thus, only one of the parental alleles is actually expressed in target cells. This epigenetic process involves regulation at a number of different stages of development and is very complex. One of the major players in the imprinting process is DNA methylation of the 5′ region of genes. Methylation targets one of the two parental alleles of imprinted genes for silencing via transcriptional repression. In studies of this epigenetic behavior, a heterozygous 3-kb deletion has been identified (), residing 280 kb upstream of the GNAS1 locus and present only in affected members and unaffected carriers. For each affected patient, the deletion was inherited maternally and associated with loss of methylation at GNAS1 exon A/B. In contrast, when the deletion was inherited paternally, methylation remained intact and Gsα function was normal (unaffected carrier status). This group hypothesized that the identified 3-kb deletion disrupts a cis-acting regulatory element necessary for methylation imprinting control of GNAS1. Future studies should help elucidate the specific regulatory element upstream of GNAS1 responsible for its imprinting control.

Mutations in specific regulatory elements may also explain some of the phenotypic differences between van Buchem disease and sclerosteosis, which are both caused by reduced levels of the SOST gene product, sclerostin. In contrast to patients with sclerosteosis who have inactivating mutations in the SOST gene itself, patients with van Buchem disease exhibit a milder, sclerosteosis-like phenotype with no SOST gene mutations, but instead have a 52-kb deletion (Vbdel) located ~ 35 kb downstream of the SOST transcript (). To examine the implications of this deletion, transgenic animals carrying either the wild-type SOST chromosomal region or the van Buchem SOST gene chromosomal region with the 52-kb deletion were constructed. Embryonic expression of the human SOST transgene was observed in both wild type and Vbdel mice. However, mice carrying the Vbdel deletion failed to express human SOST transcript in adult mineralized tissue. Analysis of the Vbdel region identified eight evolutionarily conserved elements, three of which exhibit enhancer activity in vitro. Because sclerostin functions as a bone morphogenetic protein (BMP) antagonist, further study of the Vbdel regulatory region may provide important insight into the temporal and/or spatial regulation of bone acquisition and maintenance.

In one of the more eagerly anticipated presentations, a group at deCODE Genetics reported their initial findings from a genome-wide linkage study of bone density and osteoporotic fractures in the Icelandic population (). This work will shortly be published in the new public-access journal PLoS Biology. Using affected-only, allele-sharing multipoint linkage analysis, a locus on chromosome 20p12 with significant evidence of linkage (LOD score = 4.6) was identified. Using linkage disequilibrium mapping, the investigators found a significant association between osteoporosis and haplotypes in the BMP-2 gene. A follow-up study in a much smaller Danish population confirmed the association between osteoporosis and BMP-2. Because deCODE Genetics has assembled one of the world's largest and most comprehensive collections of population data on genealogy, genotypes, and phenotypes, one can anticipate the identification of other candidate osteoporosis genes in the near future.

Estrogen plays a pivotal role in adolescent growth, sexual maturation, and skeletal development. In an interesting study, Eriksson et al. () examined the impact of a functional polymorphism (Val158Met) in the catechol-O-methyltransferase (COMT) gene that decreases the metabolism of circulating estrogen. Young females (age range, 10-12 years; Tanner stage, 1-3) with the reduced activity genotype exhibited higher estradiol levels; increased height; and greater total, femoral, and spinal bone mass and reached a later Tanner stage than did those possessing the COMT genotype with increased activity. It is not yet clear whether the association between this COMT polymorphism and skeletal development is a direct effect of increased estradiol on the skeleton or an indirect effect of the known impact of increased estrogen on the growth hormone/insulin-like growth factor 1 (IGF-1) axis to accelerate pubertal development and longitudinal bone growth. Nevertheless, this study clearly demonstrates that genetic mutations can exert pleiomorphic effects that may be overlooked if not specifically considered.

Osteoporosis-pseudoglioma (OPPG) syndrome is a rare autosomal recessive disorder characterized by severe juvenile-onset osteoporosis and congenital or early-onset blindness. The gene responsible was recently identified as the low-density lipoprotein receptor-related family member, low-density lipoprotein receptor-related protein (LRP5) on chromosome 11q11-12. Wnt-mediated signaling via LRP5 affects bone accrual during growth and is important for the establishment of peak bone mass. Clinical and genetic studies were presented of a conservative Mennonite family with eight children, three of whom had OPPG (). Sequence analysis of the LRP5 gene in this family revealed a novel exon 6 allelic variant (Trp-425-X). The parents were both heterozygous for the variant allele. Anecdotally, the authors reported that bisphosphonate therapy, which was initiated in two of the children (at ages 5 and 9 years), improved bone mineral density (BMD) and may have also reduced fracture occurrence.

In the past few years, various groups have successfully employed linkage analysis in various mouse models to identify chromosomal regions (or quantitative trait loci [QTL]) that are associated with BMD. The remarkable denouement of one QTL story was presented at this meeting (). A congenic strain had been created in which the interval containing a chromosome 11 QTL for high bone mass from C57BL/6 was transferred onto a DBA2 genetic background. Genome-wide microarray expression profiling identified a 20-fold reduction in expression of Alox15 in the high BMD congenic strain. Alox15 encodes a 12/15-lipoxygenase of hitherto unknown function. Happily, the gene had already been knocked out, and this allowed experiments in which crossbreeding of Alox15 knockout mice could be used to rescue mice from the low bone mass phenotype associated with the DBA2 allele of Alox15. Even more happily, a drug that inhibits Alox15 was available (because Alox15 was previously a putative target for treatment of cardiovascular disease), and this drug was used to show that pharmacological inhibition of 12/15-lipoxygenase increased BMD and improved bone strength in a murine model of osteoporosis. The results implicate a new pathway in the control of bone mass and simultaneously pinpoint a new class of drugs to treat osteoporosis. The final act of this story was played at a breathtaking pace that illustrates the awesome combined power of genomic analysis, microarray expression studies, knockout technology, and chemical biology to elucidate a complex disease. —GJS

Biomechanicists are quick to remind us that the load-bearing capacity of a skeletal element is determined by both its intrinsic material properties and the total amount (size) and spatial distribution (shape) of the bone tissue. Consequently, BMD studies may be an incomplete approach for effectively detecting the various genetic pathways that confer optimal skeletal mechanical strength. Using a novel panel of chromosome substitution strains (CSS) derived from a cross between the A/J and C57Bl/6J progenitor inbred strains, Ryan et al. () have begun to map the chromosomal location of QTLs influencing femoral biomechanical properties, such as ultimate failure load, post-yield deflection, and work to failure, which together describe most, if not all, aspects of the failure process. These investigators now plan to develop a hierarchical paradigm relating the mechanical properties associated with bone failure in specific CSS with readily measurable phenotypic traits (e.g., bone area, mineral content, moment of inertia, etc.) that exhibit strong heritability. Such a systems biology approach for resolving the global phenotype of skeletal fragility into relevant intermediate traits is likely to provide important mechanistic insight into the regulation of skeletal strength.

Murine researchers clearly benefit from the wide variety of genetic resources available for experimental study. These resources are taken advantage of in ongoing studies to identify genes responsible for differences in circulating IGF-1 levels between C3H/HeJ (C3H) and C57BL/6J (B6) inbred mice (). A congenic mouse carrying a portion of the C3H chromosome 6 exhibits 20% lower serum IGF-1 than does a B6 background mouse. However, attempts to generate higher resolution mapping of the QTL with interval-specific congenic sublines failed in 584 consecutive attempts. Elegant fluorescent in situ hybridization studies on chromosome spreads from the congenic and progenitor strains revealed an inversion on chromosome 6 within the introgressed region that was responsible for the observed lack of genetic recombination. To circumvent this roadblock, the investigators simply turned to an alternate cross between DBA/2J and B6 strains that exhibits the same chromosome 6 IGF QTL. Microsatellite analysis has demonstrated normal recombination rates between D2 and B6 chromosomes throughout the QTL region. Consequently, the group plans to pursue a fine-mapping strategy in congenic mice carrying different portions of the D2 chromosome 6 on a B6 background.

Another experimental resource available to mouse researchers, gene knockout technology, has proven to be a very useful method for discovering how genes function to regulate mammalian physiology. A group at Lexicon Genetics () set the goal of generating 5000 mouse lines with targeted gene-inactivating mutations by using a combination of gene-trapping and homologous recombination technologies. It is their hope that a large-scale reverse genetics approach in the mouse will identify the best targets for drug development. Thus far, the investigators have generated 1205 knockout lines, 905 of which have undergone skeletal phenotyping with DXA and/or microCT. As a test of principle, the high throughput screens successfully identified LRP5 knockouts with low BMD and klotho knockouts with high BMD. The researchers also described the identification of a novel gene coding for an enzyme that has been observed to influence bone mass. Genetic ablation of this enzyme resulted in increased trabecular bone mass. Unfortunately, to the chagrin of many in the audience, the identity of this gene was withheld, pending completion of the patenting process.

Surprises about the osteoclast Roberta Faccio, Washington University, St. Louis, MO, USA, and University of Bari, Bari, Italy Ranking the importance of RANKL

Despite the progress achieved in understanding how the receptor activator of NF-κB ligand/receptor activator of NF-κB (RANKL/RANK) pathway regulates osteoclast differentiation and function, the mechanism by which RANKL (but not other cytokines that activate NF-κB) specifically induces terminal differentiation of osteoclasts has been incompletely understood. New information in this regard has emerged using genome-wide screening.

NFATc1 is the transcription factor most effectively induced by RANKL in preosteoclasts and Raw264.7 cells (). Upregulation of NFATc1 occurs within 24 hours of RANKL stimulation and is dependent on the binding of c-Fos to the NFAT promoter (). NFAT activation is regulated by calcineurin, a Ca2+/calmodulin-dependent protein phosphatase and a downstream target of intracellular Ca2+ signaling (). During sustained calcium elevation, calcineurin dephosphorylates NFAT, thus promoting its nuclear translocation, an event blocked by cyclosporine A, which arrests calcineurin activity and inhibits osteoclast differentiation (). In preosteoclasts, NFATc1 moves to the nucleus in response to RANKL, wherein it associates with the AP1 complex and binds to the TRAP and calcitonin receptor promoters (). Of interest, ectopic expression of a constitutively active, calcineurin-independent NFATc1 mutant is sufficient to induce formation of mature osteoclasts capable of resorbing mineralized substrate in the absence of RANKL ().

At the 2003 ASBMR meeting, a number of insights into how NFATc1 promotes osteoclast differentiation were forthcoming. Evidence of collaboration between the JNK/c-Jun pathway and NFAT was provided, using a transgenic mouse in which dominant negative (DN)-c-Jun was driven by the TRAP promoter (). This mutant is osteopetrotic because of a paucity of osteoclasts. Moreover, adenoviral expression of DN-JNK or DN-c-Jun in osteoclast precursors, as well as treatment with a specific JNK inhibitor, blocks osteoclastogenesis while inducing apoptosis. Taken together, these findings argue that activation of JNK/c-Jun is essential for RANKL-mediated osteoclastogenesis.

The inhibitory effects of DN-c-Jun on osteoclastogenesis are dependent on the inability of transgenic cells to express NFAT following RANKL stimulation. Of interest, although overexpression of NFAT in wild-type (WT) cells induces osteoclast formation, even without RANKL, such is not the case when cells express DN-c-Jun, which suggests that collaboration between c-Jun and NFAT regulates osteoclastogenesis (). Similar results were described by Galson et al., who expressed a nuclear form of NFAT into c-Fos(-/-) precursors, eventuating in normal osteoclastogenesis (). In contrast, rescue of c-Fos null osteoclastogenesis with this truncated and more potent form of NFAT still requires RANKL stimulation. Therefore, despite the fact that NFAT expression induces osteoclastogenesis in WT cells independent of RANKL, this cytokine activates intracellular signals which, in the presence of NFAT, compensate for lack of c-Fos.

When NF-κB signaling is eliminated (in a p50/p52 double knockout mouse), RANKL-mediated osteoclastogenesis is blunted. However, the derived splenocytes can be induced to form osteoclasts if c-Fos is overexpressed and the cells are treated with interleukin-1 (IL-1) or tumor necrosis factor (TNF) (). These in vitro data demonstrate that cytokines can compensate for the lack of RANKL/RANK-mediated signals in the presence of c-Fos, which underscores the central role of c-Fos as a downstream mediator of osteoclast recruitment. This insight is reinforced by the fact that c-Fos is highly expressed in osteoclasts in inflamed joints, but not in cells adjacent to normal metaphysis ().

Fra-2, another Fos family member, also seems to regulate bone remodeling. Although Fra-2 transgenic mice have increased bone formation and volume, those lacking Fra-2 die postnatally with severe osteoporosis (). It is not clear if Fra-2 is required only for osteoblast differentiation, which is arrested in its absence, or if the protein also impacts osteoclasts. Thus, the question as to why deletion of Fra-2 leads to defective osteoclast differentiation in vitro, but decreased bone mass in vivo, needs to be answered.

As discussed by Yongwon Choi in a State-of-the-Art Lecture, the bone marrow milieu of cytokines and growth factors may compensate for the absence of specific intracellular signaling proteins (). For example, the diversity of the in vivo phenotype of TRAF6(-/-) mice generated in three different laboratories may be explained by differences in the age of the animals studied. A novel (yet controversial) hypothesis postulates the presence of physiological bone age-associated factors (BAAFs). These putative molecules may contribute to osteoclast differentiation and may be responsible for the presence of osteoclasts in TRAF6(-/-) older mice and their absence in younger animals. To this end, Choi et al. rescued the arrested in vitro osteoclastogenesis of TRAF6(-/-) cells, but not their function, by adding transforming growth factor-β (TGF-β) and tumor necrosis factor-α (TNF-α) in the presence of macrophage colony-stimulating factor (M-CSF) (). Therefore, TGF-β and TNF-α are possible candidate BAAFs, but further confirmation regarding age-dependent expression of these factors is required.

Osteoclasts and immunobiology

T lymphocytes participate in the bone loss that characterizes conditions like inflammatory arthritis and estrogen deprivation. It seems that T-cell production of RANKL is central to this phenomenon. At this meeting, the NFAT family of transcription factors was shown to regulate the RANKL gene in CD4+ cells following T-cell receptor activation (). On the other hand, a study of nude mice reconstituted with different T-cell subpopulations provides evidence that CD4+ lymphocytes are primary TNF-α-producing cells and major players in ovariectomized-induced osteoporosis (). Therefore, the ability of T cells to contribute to the bone loss attending pathologic conditions is dependent on the multiple proosteoclastogenic cytokines produced in loco.

IL-7 may be the key cytokine in estrogen deficiency-induced T-cell activation (). At this meeting, it was shown that IL-7 controls T-cell activation by enhancing the antigen-presenting cell (APC) activity of macrophages (). This event reflects upregulation of the APC-regulating gene CTIIA in response to increased production of interferon-γ. TGF-β1, which blocks T-cell activation via downregulation of CTIIA expression and APC activity, adds to the complexity of T-cell regulation during estrogen deficiency (). In agreement with this hypothesis, increased circulating TGF-β prevents ovariectomy-induced bone loss. Moreover, T cells from dominant negative TGF-β receptor II mice, transplanted into sham nude mice, are no longer sensitive to the presence of estrogen and sustain bone loss equal to that observed in WT ovariectomized mice. Thus, TGF-β renders T cells insensitive to lack of estrogen and blunts their capacity to induce bone loss in vivo.

Alternatively, TGF-β directly promotes osteoclast and osteoblast differentiation by upregulating TGF-β inducible early gene (TIEG) in precursors of both cells (). Furthermore, the growth factor stimulates Cox2-dependent prostaglandin production by breast cancer cells, which in turn, leads to enhanced osteoclastogenesis that promotes the progression of osteolytic metastasis (). Thus, the impact of TGF-β on bone resorption is both complex and interesting.

From cytoskeleton to skeleton

The capacity of osteoclasts to degrade bone requires a highly organized cytoskeleton. These cells contain podosomes, unique adhesive structures consisting of a central core of actin filaments surrounded by several cytoskeletal proteins, which rapidly form and disassemble (). Although the mechanism of podosome development is unclear, more is known about how it disassembles.

The Arp2/3 complex regulates growth of branched actin filaments at the leading edge of membrane extensions () and, in mature osteoclasts, colocalizes with the actin core of the podosome and the actin ring (). In Raw cells, Arp2-SiRNA inhibits osteoclast differentiation and, in the mature polykaryon, induces actin ring disruption (). Osteopontin, which promotes podosome reorganization, stimulates association of PtdIns P2, WASP, and Arp2/3 in murine osteoclasts. This event is dependent on activation of Rho GTPases (). The importance of small GTPases in regulating the actin ring is buttressed by the fact that overexpression of DN-Rac alters the osteoclast cytoskeleton and the cell's ability to resorb bone (). Because these data come from overexpression studies, their physiological relevance is yet to be established. In this regard, examination of mice deleted of specific members of the Vav family of guanosine exchange factors, which are responsible for Rho and Rac activation, indicates that Vav3 is uniquely expressed in abundance by osteoclasts and is central to αvβ3 integrin- and M-CSF-mediated organization of the cell's cytoskeleton and resorptive capacity ().

Finally, information regarding the means by which c-Src exerts its fundamental impact on the osteoclast cytoskeleton was presented. Overexpression of Chk, the Csk homolog in osteoclasts that inactivates the kinase activity of c-Src by phosphorylating SrcY527, disrupts the actin ring and arrests bone resorption. Hence, Chk, which is specifically expressed by osteoclasts, is a potential therapeutic target ().

Surprise, surprise

Osteoporosis is one of the leading causes of morbidity in the elderly, and understanding the mechanisms by which osteoclasts work and/or the factors that regulate their maturation and activity is critical for its cure. New evidence details unexpected links between osteoclasts and proteins not previously considered functional in the skeleton. The NFAT family of transcription factors, known to control T-cell differentiation and peripheral vascular development (); the Dap12 receptor, involved in transducing activation signals for an extended array of receptors in NK cells, granulocytes, monocytes/macrophages, and dendritic cells (); and the Vav family of guanosine exchange factors, involved in actin dynamics related to T- cell receptor engagement (), all participate in osteoclast biology (,,,,). Perhaps even more surprisingly, two pituitary hormones, thyroid stimulating hormone and oxytocin, the latter known for its effects on lactation and parturition, are negative and positive regulators, respectively, of bone remodeling (). Thus, the discoveries reported at the 2003 ASBMR meeting leave us with much to do.

New twists in osteoblast function Gerard Karsenty, Baylor College of Medicine, Houston, TX, USA

This brief and necessarily incomplete review highlights some topics about which progress was reported at the 25th Annual Meeting of the ASBMR. Transcriptional control of osteoblast differentiation was again a major focus at the meeting in the field of osteoblast biology. Attention was focused primarily on regulation of Runx2 function and identification of novel regulators of osteoblast differentiation. Three regulators of Runx2 function were reported, Twist-1, Twist-2 and Stat-1. Twist-1, a basic helix-loop-helix (bHLH) protein, and its homologue Twist-2 () inhibit osteoblast differentiation in vitro and interact with the Runx2 DNA binding domain. Surprisingly, this interaction occurs not through the bHLH domain of Twist-1 and Twist-2, as initially anticipated, but through the last 20 amino acids of these two proteins, a novel and unique domain called the “Twist box.” The importance of the Twist box was confirmed by loss-of-function experiments in vivo. This inhibition of osteoblast differentiation occurs early in development, at a time when Runx2 is expressed, but during which osteoblast differentiation does not occur. Of interest, the expression of Twist-1 is downregulated by one member of the Wnt family, Wnt-11 in a cell line, which suggests the existence of an additional regulatory loop during osteoblast differentiation () — Wnt proteins favoring osteoblast differentiation by downregulating the expression of an inhibitor of osteoblast differentiation.

Acting much later than the Twist protein, Stat-1 is also a novel regulator of Runx2 function. Stat-1-deficient mice show a progressive increase in bone mass as they become adult. This increase is the result of an increase in bone formation parameters, whereas bone resorption is not affected in Stat-1-deficient mice (). This study (), soon to be published, confirms the finding of another group who recently reported that Stat-1 acts as an inhibitor of Runx2 by retaining it in the cytoplasm (). In another study of Runx2 function, it was reported that inactivation of the predominant form of Runx2 in bone, Runx2-II, revealed a surprisingly mild skeletal phenotype. However, there was no indication of whether the Runx2-I isoform was overexpressed in these mutant mice. If it was, the observed mild phenotypic abnormalities would be explained ().

Information about a novel transcriptional regulator of osteoblast differentiation, Fra-2, was presented at the meeting. Fra-2 is the most recent member of the AP-1 family shown to be involved in skeletal development (). On one hand, overexpression of Fra-2 results in increased bone mass caused by an increase in bone formation parameters, although the number of osteoblasts remains unchanged. On the other hand, Fra-2-deficient mice, at time of death, have a severe reduction in bone mass. The number of osteoblasts is normal in vivo, although there seems to be a defect of differentiation of these cells as Osteocalcin expression is decreased in vivo and as the growth of nodule of mineralization is altered in vitro. The osteoblast is not the only cell type affected by Fra-2 inactivation. Indeed, osteoclasts are bigger in vivo, yet all culture experiments revealed a noncell-autonomous defect in osteoclast differentiation and fusion. The molecular bases of these various abnormalities remain unclear.

Several studies addressed the emerging topic of Wnt signaling in bone formation. A first study described in detail several classes of mutations in the beta propeller 1 of low-density lipoprotein receptor-related protein 5 (LRP5), a putative coreceptor for Wnt protein (). This study found a surprising difference between LRP5 and LRP6 (i.e., a similar mutation had different consequences when introduced in LRP5 and LRP6). Yet, a poster from the same group showed evidence that LRP6(+/-) mice have low bone mineral density, suggesting that Lrp6, which is expressed in the osteoblast, could also be involved in the Wnt-dependent regulation of bone formation (). No histologic parameters were reported, and no comparison between Lrp5(+/-) and Lrp6(+/-) mice was presented. In light of these results, analysis of Lrp5(+/-)/Lrp6(+/-) mice will be of great interest.

Several studies addressed the role of the canonical Wnt signaling pathway and specifically of its molecular node, β-catenin. Inhibition of β-catenin degradation by an inhibitor of glycogen synthase kinase 3β results in increased bone formation in mouse calvaria (). Moreover, mice in which a constitutively active form of β-catenin was “knocked in” exclusively in osteoblasts had increased bone formation in vivo. Surprisingly, this mutation had a second consequence: it increased expression of Osteoprotegerin, and as a result, markedly decreased osteoclast differentiation, indicating that the canonical Wnt signaling pathway may have two effects — enhancing bone formation while at the same time decreasing bone resorption (). Another study showed that bone morphogenetic protein 2 (BMP2) signaling and β-catenin have a synergistic effect to increase calvaria bone formation in vivo and an additional effect on extracellular matrix mineralization in vivo. A dominant negative form of the Tcf protein partially inhibited the interaction between these two pathways. Of interest, an anti-Smad4 antibody could immunoprecipitate β-catenin ().

Another area of research whose representation has increased at the meeting recently is bone formation by the central and peripheral nervous systems. One study performed in rats showed that β2 adrenergic agonists have a deleterious effect on bone mass (). In the discussion following this presentation, it was reported that children receiving β2 adrenergic agonists alone as a treatment for asthma had increased incidence of bone fracture (). In addition to this sympathetic regulation, two other studies focused on neurons of the hypothalamic arcuate nuclei. One study showed that inactivation of the genes encoding receptors for neuropeptide Y (NPY), a major gene product of arcuate neurons, leads to increased bone mass (). The other study showed that although arcuate neurons exert a modest antiosteogenic function, compared with ventromedial hypothalamic neurons, inactivation of NPY does not affect bone mass (). The difference between the results of these two studies may be explained by the decrease in serum leptin level in NPY receptor-deficient mice and/or the fact that other ligands, such as peptide YY or pancreatic peptide, can bind to these receptors.

Phosphatonin comes of age Gordon J. Strewler, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA, USA

After the ASBMR meeting a year ago, there was little doubt that fibroblast growth factor 23 (FGF23) induces renal phosphate excretion and is thereby a physiologically important regulator of phosphate homeostasis. FGF23 is a potent phosphatonin that is secreted by tumors and causes phosphate wasting in tumor-induced osteomalacia (TIO) (). Mutations that block the degradation of FGF23 cause autosomal dominant hypophosphatemic rickets (ADHR) (). At last year's ASBMR meeting, knockout of the FGF23 gene was shown to produce hyperphosphatemia and markedly increased 1,25-dihydroxyvitamin D levels (), a phenotype that is the mirror image of renal phosphate-wasting disorders. This work, yet to be published, makes clear that FGF23 is a phosphate-regulating hormone.

At this year's ASBMR meeting, there were two important additions to the portrait of FGF23 as a phosphate regulator. First, neutralizing antibodies were used to show that blocking FGF23 action in the mouse increases serum phosphate and 1,25-dihydroxyvitamin D (). Combined use of amino- and carboxyl-terminal neutralizing antibodies had a greater effect than either antibody alone. Second, manipulation of dietary phosphate intake in healthy humans changed serum levels of FGF23 in the expected directions, confirming a report at last year's meeting () and identifying a phosphate-sensing mechanism that allows FGF23 to respond to modest changes in dietary phosphate intake in humans (). To be clarified are the identity and site of the putative phosphate sensor, the mechanism of coupling to FGF23, and the physiologically relevant site of FGF23 secretion.

Although X-linked hypophosphatemia (XLH) is the paradigm of inherited phosphate wasting, the relationship of XLH to the phosphate-wasting disorders caused by FGF23 has been up to this time uncertain. Causative mutations in XLH inactivate PHEX, a metalloprotease that is associated with the plasma membrane and expressed most strongly in bone. One laboratory has reported that PHEX cleaves FGF23 (). This finding would be consistent with a scenario in which inactivating mutations in PHEX prevent degradation of FGF23 (as do FGF23 mutations in ADHR) and thereby allow the accumulation of this phosphatonin. However, other workers have had difficulty showing that FGF23 is a target for PHEX (). Two pieces of evidence at this meeting tie XLH to FGF23 in a definitive way (). First, FGF23 levels are extremely high in the Hyp mouse, an animal model of XLH. Second, neutralization of FGF23 activity in Hyp mice with antibodies reverses all manifestations of the Hyp phenotype, producing marked hyperphosphatemia, increased 1,25-dihydroxyvitamin D, and after four weeks of treatment, healing of rickets and osteomalacia.

The exact relationship between PHEX and FGF23 remains uncertain, however. The expression of FGF23 is markedly increased in bones of Hyp mice, most consistent with a complex epigenetic relationship between FGF23 and PHEX, in which PHEX acts locally to increase secretion of FGF23 from bone cells (). However, the serum level of FGF23 is increased in less than one-half of XLH patients and is normal in the remainder (), in contrast to the Hyp mouse. This raises the possibility that, in humans, PHEX also acts downstream of FGF23. It will be important to determine the chemical nature of circulating FGF23 in XLH patients and the minimal FGF23 sequence that is required for biological activity. It is not inconceivable that PHEX, although it does not cleave intact FGF23, acts by inactivating a biologically active fragment of FGF23. Alternatively, PHEX could activate a second protease that attacks FGF23, either directly or indirectly.

A fourth hypophosphatemic disorder was also linked to FGF23 at this meeting. In confirmation of results published last month (), it was reported that FGF23 is the cause of renal phosphate wasting in the McCune-Albright syndrome (). The source of FGF23 in the McCune-Albright syndrome seems to be fibroblasts from fibrous dysplastic lesions ().

Two other phosphatonins were also the subjects of presentations at the meeting. Like FGF23, soluble frizzled-related protein 4 (sFRP4) was identified in tumors that cause TIO. As recently published (), sFRP4, a wnt antagonist, is a potent phosphatonin and decreases renal phosphate reabsorption without a compensatory increase in 1,25-dihydroxyvitamin D (), consistent with expectations in phosphate-wasting disorders. With some TIO patients having normal FGF23 levels (), there is room for a second phosphatonin, and sFRP4 is a good candidate. Matrix extracellular phosphoglycoprotein (MEPE), a third protein found in TIO, is a less attractive candidate. MEPE has phosphaturic effects, albeit at high concentrations (), but MEPE knockout mice have normal phosphate metabolism ().

PTH vs. PTHrP and calcitonin vs. CGRP Gordon J. Strewler, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA, USA

Little by little, the relationships of parathyroid hormone (PTH) and parathyroid hormone-related protein (PTHrP) as regulators of bone remodeling become clear. At the 2003 ASBMR meeting, it was shown that mice having a deletion of the PTH gene and also heterozygous for the PTHrP gene [PTH(-/-)/PTHrP(+/-)] have severe osteoporosis in comparison with PTH(-/-) mice. These mice have a major impairment in osteoblast function (): osteoblast surface is decreased and osteoblast formation is impaired (reduced osteoblast colony formation in culture), at the same time that osteoblast apoptosis is increased. The mice develop osteoporosis despite a decrease in receptor activator of NF-κB ligand (RANKL) and a proportionate decrease in osteoclast number. Thus, PTH and PTHrP may function in the same pathway to regulate osteoblast formation and survival, and the subtle osteoporotic phenotype previously observed in PTHrP(+/-) () is apparently rescued in part by the presence of PTH. PTHrP would provide for paracrine signaling from mature osteoblasts, which express the gene, to regulate bone formation, but the physiological circuitry for local regulation of osteoblast function remains to be found. It is presumably this shared pathway that is activated by anabolic regimens of PTH in the treatment of osteoporosis.

Along these lines, it was reported that marrow cells in osteitis fibrosa cystica can become osteoblasts and form bone. During a seven-day continuous infusion of PTH, much of the proliferating population of marrow fibroblasts, but only 5% of osteoblasts on the bone surface, can be labeled with [3H]-thymidine (). One week after administration of PTH and [3H]-thymidine is discontinued, however, large numbers of osteoblasts and osteocytes now bear the tritium label, indicating that they originated from labeled fibroblastic marrow precursors. After 28 days, bone mass, bone stiffness, and torque to failure are increased. Thus, fibroblast-like marrow cells in osteitis fibrosa, like similar cells in bones of mice with a constitutively active PTH receptor (), are not true fibroblasts, but early osteoblast precursors. It seems that PTH can induce increased numbers of osteoblasts by differentiating lining cell precursors () or fibroblast-like precursors, or by inhibiting osteoblast apoptosis (). It may well be that conversion of fibrous cells to osteoblasts is the explanation for the “hungry bone syndrome” after parathyroidectomy.

The adapter protein Na/H exchanger regulator factor (NHERF) binds the PTH1 receptor (PTH1R) in kidney cells and switches its signaling properties (). NHERF also has a role in altering the specificity of PTH peptides for internalization of PTH1R (). In cells that express NHERF (renal proximal tubule and SaOS2 cells), only the PTH agonist PTH(1-34) induces internalization of the receptor, but in cells that do not express NHERF (distal convoluted tubule and ROS 17/2.8 cells), either PTH(1-34) or the antagonist PTH(7-34) can induce receptor internalization. Introduction of NHERF converts ROS 17/2.8 cells to a state where PTH1 receptor internalization is agonist-specific. This action of NHERF requires that cells have an intact cytoskeleton and that NHERF possesses the ERM domain, which binds to actin-associated proteins (ezrin, radixin, and moesin).

Whether one develops pseudohypothyroidism depends on the sex of the parent who transmitted the trait. Isolated resistance to PTH, pseudohypoparathyroidism 1b (PHP1b), is a maternally transmitted trait linked to the GNAS1 gene. GNAS1 encodes the PTH receptor coupling protein Gsα, as well as several other transcripts. It seems that in PHP1b undermethylation of GNAS1 prevents the expression of the maternal allele, the only allele that is normally expressed in the proximal renal tubule (). The methylation defect in PHP1b could not be ascribed to an inherited mutation of the GNAS1 gene, however, so the genetic locus of the imprinting defect remains to be found. This complex story was previously retold in a BoneKEy Commentary.

The mutation that causes PHP1b has now been located (), and remarkably, it is 280 kb upstream of the GNAS1 locus. In 12 kindreds and four sporadic cases, an identical 3-kb deletion was identified within the gene for syntaxin-16, a SNARE protein involved in protein sorting in the Golgi. But several lines of evidence argue that syntaxin-16 transcripts are not directly involved in PHP1b. First, the mutation is always maternally inherited, but the syntaxin-16 gene is not an imprinted locus. Second, the mutation always acts in cis. Third, an identical mutation is found in multiple kindreds, suggesting that more than a simple loss of syntaxin-16 function is involved. Rather, it seems highly likely that the deleted region represents an imprinting control region for GNAS1. This work may ultimately close the book on the molecular pathogenesis of PHP1b, but it will simultaneously open a new chapter in the imprinting story. Furthermore, one continues to wonder how an isolated proximal tubule defect in PTH action results in hypocalcemia, in the face of multiple compensatory mechanisms that remain intact.

The calcitonin system continues to provide surprises. It was recently reported that deletion of the calcitonin gene leads expectedly to a high bone mass phenotype (). But the other calcitonin gene product, calcitonin gene-related peptide (CGRP), was also deleted in that particular mouse. Was it the loss of calcitonin or of CGRP that caused the bone phenotype? At this year's meeting, a mouse was presented in which only CGRP had been deleted from the calcitonin gene by selective targeting (). This mouse had normal calcitonin levels and no bone phenotype. Barring trivial explanations, such as strain differences in bone mass, it thus seems that calcitonin deficiency per se is what affects bone mass.

Calcitonin also has a role in bone during lactation (). Mice in which calcitonin and CGRP are deleted have a markedly exaggerated loss of bone mineral during lactation and transient hypercalcemia, suggesting strongly that calcitonin (and/or CGRP) protects against excessive bone loss during normal lactation. Now, the mice generated in Abstract 1052 can be used to determine whether it is indeed calcitonin or CGRP that is responsible for protection.

Defining fracture risk using biochemical markers and novel imaging techniques Aubrey Blumsohn, University of Sheffield, UK, and Yebin Jiang, University of California, San Francisco, CA, USA

Apart from bone mineral density (BMD), many factors contribute to fracture risk. Several recent studies have also drawn attention to the fact that change in BMD with antiresorptive therapy explains less than one-fourth of therapeutic fracture risk reduction (,,,). It therefore seems likely that other radiographic techniques and laboratory investigations might be useful for capturing other facets of risk or for therapeutic risk reduction.

Evidence was presented suggesting that the early change in bone resorption in response to risedronate therapy explained about 33% of vertebral fracture risk reduction in the Hip Intervention Program (HIP) trial (). Change in urinary N-telopeptide fragment of type I collagen (uNTX) was significantly associated with time to vertebral fracture in patients receiving risedronate (5 mg) (Cox regression, P = 0.003), but change in a marker of bone formation (PINP) was not predictive of risk reduction in these patients (P = 0.142). In a combined analysis of patients receiving risedronate in the Vertebral Efficacy with Risedronate Therapy (VERT) and HIP trials (group 1, low BMD), early reduction in uNTX (3 to 6 months) explained 33% of vertebral fracture risk reduction over 3 years, compared with measured change in BMD, which explained 26% of such risk reduction (). Measured change in hip BMD and in uNTX/Cr explained about one-half of vertebral fracture risk reduction. The relatively poor predictive performance of BMD change might be attributed to the imprecision of serial densitometry, at least to some extent.

The effect of bone turnover rate prior to antiresorptive therapy on the efficacy of therapy is also poorly understood. In the Fracture Intervention Trial (FIT), reduction in nonspine fractures seems to be greater in osteoporotic women with high baseline PINP (), but not in those with high bone alkaline phosphatase (ALP). Treatment efficacy for spine and hip fractures, however, was not associated with baseline bone turnover.

Debate continues about the use of markers of bone turnover to predict fracture in untreated patients. One study () examined the ability of prevalent fracture and markers of bone turnover to predict all-incident fracture over 10 years in osteopenic women in the OFELY study. Prior fracture was a powerful predictor of future fracture in this cohort (all fractures, n = 54/322; hazard ratio [HR], 2.7; P < 0.002). However, combined assessment using prevalent fracture and serum bone ALP did not increase prediction (HR, 2.7; P < 0.002), and addition of serum βCTX impaired prediction (HR, 1.8; P = 0.03). In another long-term study, urinary pyridinoline was predictive of both vertebral and nonvertebral fracture over 10 years (). Three large prospective studies have previously examined the ability of markers of bone turnover to predict hip fracture, and the ASBMR meeting saw the addition of a fourth study (). This study cohort consisted of 5212 women (age, 75 years or older), and 157 women sustained hip fracture over three-year median observation. Although serum βCTX was significantly predictive of hip fracture (odds ratio [OR], 1.3 per SD decrease; confidence interval [CI], 1.1-1.5), this was not independent of either hip BMD (OR, 2.5/SD; CI, 2.0-3.1) or body weight (OR, 1.85/SD; CI, 1.48-2.32). Other studies have demonstrated the substantial importance of body weight as a predictor of vertebral () and hip () fracture. These studies reemphasize the importance of showing that new technologies can provide useful incremental information if age, prior fracture, clinically evident risk factors, and BMD are known.

A plethora of enhanced radiographic techniques also holds promise for adding to information provided by traditional densitometry. Enhanced variants of conventional dual-energy x-ray absorptiometry (DXA) may be useful. Geometric indices, such as section modulus (Z, bending strength) at the narrowest point in the femoral neck, can be derived from two-dimensional (2D) DXA scans (hip structure analysis). A study of long-term care residents showed age-related changes in hip structure using this technique and that femoral geometry is predictive of hip fracture in this group (). Another study () showed that although hip BMD declines in the decade following attainment of peak bone mass in premenopausal women, Z increases. In the femoral neck, this increase was attributed to prior calcium supplementation (). Radiographic texture analysis of conventional calcaneal DXA images may also enhance the information content of these measurements (). Three-dimensional (3D) reconstruction of skeletal geometry is also possible using two perpendicular DXA scans, which may allow better estimation of failure load (). Complete quantitative computed tomography-based 3D reconstruction and finite element modeling of the skeleton is possible () and can be used to predict the response to particular loading conditions.

Microcomputed tomography (microCT) can also reproducibly quantify 3D microarchitecture of trabecular and cortical bone in iliac crest biopsy specimens (). One study () using bone cores from the proximal femur showed that structural parameters from 2D bone core radiographs correlate well with bone failure load and with 3D microCT measurements, indicating that inexpensive proximal femoral trabecular structural analysis from hip radiographs may also be useful. It is possible to use mathematical techniques to identify plate- or rod-like structures in a 3D image (), and osteoporotic patients can be shown to have a lower volume of rod and branching components.

Imaging techniques may also provide useful information about chemical composition of skeletal tissue. One study () used epi-illumination Raman microspectroscopic imaging to examine carbonate/phosphate and phosphate/amide I ratios in trabecular bone specimens. It may be possible to use such techniques in vivo.

Although the path from research to clinical practice may be a steep one for several of these techniques, risk assessment based on enhanced imaging techniques seems likely to be exciting.

The material and structural properties of bone and bone strength Ego Seeman, University of Melbourne, Melbourne, Australia

Bone mass, bone mineral content, bone mineral density, apparent density, true density, quality and strength. Our field is full of vague terminology; bone mass means bone tissue volume, but is sloppy shorthand for bone mineral mass, bone mineral content (BMC), or bone mineral density (BMD). Bone mineral content is the amount of bone mineral in an undefined volume of a skeletal region, and BMD is BMC per unit projected area of a region measured using bone densitometry. True density is territory where angels should fear to tread because “true density” may refer to the amount of mineral per unit volume of (i) bone tissue, including marrow and tissue void spaces, as measured by quantitative computed tomography (QCT); (ii) tissue alone, excluding voids; or (iii) volume of bone mineral only, excluding tissue and voids. Implicit in the word “true” is the flawed notion that this measure, often obtained using QCT, is the “real thing” — that is, “true” because it is volumetric and thus promises to be a better predictor than other expressions of bone mineral mass. A full discussion of this nomenclature is long overdue, but such a manuscript remains in search of an author. My brief is to welcome bone “quality” to the usual list of suspect terminologies.

Bone mineral density + bone quality = bone strength, the National Institutes of Health consensus equation, is an attempt to focus investigators on the structural, material, and biomechanical basis of bone strength (). This is a good thing, but “quality” and “strength” are just as vague as “density.” This equation does little to advance our understanding, and indeed instead impedes it, because tissue mineral content (TMC), a material property, is contained in both the first and second terms. An alternative might be to follow the approach taken in the biomechanical literature: structural and material properties of bone = bone strength.

The lack of communication with biomechanical engineers has resulted in a lack of discussion concerning the vagaries of bone “strength.” What is bone “strength?” What is “minimal” trauma? What are the abnormalities in the material and structural determinants of peak, yield, and ultimate stress (load per unit area) and strain (change in length/original length) during compressive, tensile, and torsional loading? What is bone toughness — the ability of bone to absorb energy imposed on it by deforming without fracture?

Things are improving. We are getting the message, but not clearly enough. Several abstracts at the ASBMR meeting this year addressed the structural and material properties of bone during aging, in disease, and in response to drug therapy, but the number of studies was painfully few. In part, the sparsity reflects the lack of tools available to accurately measure material properties — TMC, microdamage burden, and osteocyte density; properties of collagen, such as its cross-linking; or macro- and microstructure (e.g., bone size and shape, cortical thickness, intracortical porosity, trabecular number, thickness, and connectivity).

Material properties of bone

Tissue Mineral Content, Bone Remodeling, and Bone Strength

Nature chooses the amount of material for the function needed. A most important material property of bone is the degree of mineralization of its tissue. The greater the TMC or ash density, the greater the stiffness and peak stress the bone will tolerate. More is not necessarily better: 100% mineralized bone is brittle and will not “give” during impact loading — that is, bone toughness declines.

A loss of bone toughness is a potential concern with prolonged bisphosphonate therapy. Suppression of bone remodeling allows more time for secondary mineralization to occur, and thus, TMC and homogeneity of bone increase. Increased tissue mineral homogeneity may allow microdamage propagation, which causes microcracks to increase in length. Bone remodeling removes microdamage. Reduced microdamage removal plus increased microdamage production leads to increased microdamage burden, which in turn, may reduce bone toughness. At this time, animal experiments using high doses of alendronate, risedronate, and incadronate indeed confirm that this occurs and that reduced bone toughness is found in some (but not in all) studies (,,,). No data exist in human subjects given bisphosphonates to confirm or refute the presence of microdamage in bone biopsy specimens, nor is there evidence that bone fragility is increased with bisphosphonates. Nevertheless, this issue should not be ignored.

In postmenopausal women, high bone remodeling may reduce TMC and thus reduce material stiffness. Treatment may restore TMC and material strength, but is it possible that bisphosphonates may increase TMC to the point where material stiffness is increased — predisposing to brittleness (loss of toughness) — in women with normal or higher TMC? There is evidence that some patients with fractures have high degrees of TMC (). Would tailoring treatment to patients with low TMC be more efficacious?

Boivin et al. () reported that zoledronate, like several other bisphosphonates, increases TMC. The extent of increase in secondary mineralization may vary according to the type of bisphosphonate used and the duration of therapy. Recent studies suggest that raloxifene also increases TMC, but to an extent that is no greater than that found in calcium-supplemented controls (). Strontium ranelate (SR) also does not seem to increase bone TMC, and Le Geros et al. () reported that one-year treatment with SR had no effect on crystal size or on composition of bone in nonhuman primates.

Bone remodeling rate

At some time before menopause, bone balance in the basic multicellular unit (BMU) becomes negative because of a reduction in the volume of bone formed in each BMU. The negative bone balance within each BMU is the basis of bone loss and structural damage. The same loads on bone are imposed on a structure diminished in cross-sectional area (CSA), and thus the load per unit area increases, predisposing to buckling, microdamage, and ultimately, fracture.

The rate of bone remodeling itself is regarded as a property of bone that determines its strength. Higher remodeling rates reduce material stiffness by replacing older, more densely mineralized bone with younger, less densely mineralized bone. Recker et al. () reported that activation frequency increases with advancing years after menopause — 0.15, 0.29, and 0.44 at age 49, 55, and 60 years, respectively. This observation is interesting because it suggests that rather than being high initially, there is a progressive increase in remodeling rate, which perhaps relates to progressively worsening estrogen deficiency. The age at which remodeling rate reaches a peak, and then perhaps is compounded by secondary hyperparathyroidism, is not clear.

In the face of remodeling imbalance in the BMU, it is likely that interindividual differences in the rate of bone remodeling, rather than interindividual differences in the degree of BMU imbalance, produce the structural damage. Nevertheless, there is likely to be heterogeneity — that is, some individuals may have a high remodeling rate with little differences in BMU imbalance, whereas other individuals may have slower rates of bone remodeling with large differences in BMU imbalance. Furthermore, individuals with BMU imbalance caused by resorptive deficits may suffer greater loss of strength than do those with BMU imbalance caused by reduced bone formation at the cellular level because greater resorptive depth is more likely to produce complete loss of structural elements. If this is correct, should antiresorptive therapy be used in individuals with greater remodeling rates and greater resorptive removal of bone and should anabolic therapy be used in individuals with reduced bone formation at the cellular level? Indeed, in patients with high resorptive removal of bone, could parathyroid hormone be deleterious or at least less effective? What is the relative contribution of the decline in material stiffness to increased bone fragility?

Osteocytes and microdamage

The osteocyte is the most abundant, yet most neglected, cell in bone. After completion of bone remodeling, osteoblasts become lining cells, undergo apoptosis, or become osteocytes housed in their lacunae in the less dense lamellae of the osteons in cortical bone. These cells connect with each other and lining cells, probably to serve as sensors of deformation. Nicolella et al. () reported that osteocyte deformation is responsible for the transduction of mechanical signals. The degree of cell deformation increased as a function of fluid shear stress and correlated with the increase in prostaglandin E2 and dendrite formation. At the end of 10 minutes of flow, cell branching occurred. Bone fragility may have been partly the result of osteocytic cell death. In this situation, perhaps failure of osteoocytic signaling of increasing microstrain may fail to initiate compensatory bone formation leading to microdamage. Collishaw et al. () reported that ovariectomy increased apoptosis of osteocytes and that the selective estrogen receptor modulator LY117018 and estrogen prevented such apoptosis.

Oiu et al. () examined whether the lacunae containing osteocytes may be stress risers — that is, sites that concentrate stress, which may then initiate or predispose to microdamage. (A useful analogy is eddy currents created around a rock protruding from the surface of a gently flowing stream.) If this is correct, the density of lacunae should be higher to adjust to regions with microdamage. The opposite was reported. The investigators found that lacunar density was 16% lower in the vicinity of microdamage. This finding contrasts with that of Reilly (), who reported greater microdamage around lacunae.

Bone collagen

The extent of cross-linking of collagen is likely to be a determinant of bone toughness, the ability of bone to deform during loading without fracturing. Garnero et al. () studied bovine fetal bone incubated at increasing duration to induce cross-linking of collagen without altering TMC. The investigators reported a decline in the elastic strength of bone consistent with the notion that increased collagen cross-linking is a determinant of bone strength. Keaveny et al. () reported that in specimens from 23 men and 19 women, energy to failure correlated with collagen content and inversely correlated with the extent of pentosidine and cortical porosity. Whether these observations were adjusted for the amount of TMC is not clear.

Silva et al. () examined the material basis for low bone strength in the SAMP6 mouse. Both SAMP6 and SAMR1 controls had a decline in ultimate force and ultimate stress with demineralization of bone specimens, but both had a similar decrease in stiffness and energy to failure. There was a decrease in hydroxyproline and collagen per unit dry weight. The authors inferred that bone collagen abnormalities contributed to reduced bone strength. High levels of nonenzymatic cross-linking give rise to more brittle (less tough) bones.

Macro- and Microarchitecture

Cortical and trabecular bone

Strength and lightness are achieved by architectural design. Long bones are weight bearing and should not bend too much (i.e., stiffness is favored over flexibility). Bone tissue is fashioned into long bones with a medullary canal and a cortex of mineralized tissue placed distant from the central long axis, a geometric feature that confers greater resistance to bending than the same unit area placed near the long axis, because the resistance to bending (the moment of inertia) is a function of the fourth power of the distance from this long axis.

The absolute and relative movements of the periosteal and endosteal envelopes determine the CSA, diameter of long bone, mass and thickness of cortical bone, and distance that cortical mass is placed from the neutral axis of the bone. The main sex difference in bending strength is achieved not by a thicker cortex in males, but by placement of the cortex further from the neutral axis in males than in females. This greater radial displacement of the cortex also produces a larger CSA upon which compressive loads can be distributed.

In contrast, vertebral bodies are structures consisting of an open porous spongiosa, a mineralized interconnecting honeycomb of plates that functions like springs able to store energy by deforming in compression. This structural adaptation achieves lightness by its porous network, with strength in tolerating greater peak strains than does cortical bone while sacrificing peak stresses (load/area), compared with cortical bone; the trabecular structure withstands larger deformations to facilitate flexion, extension, and rotation of the whole vertebral skeleton of the upper body.

The irregular external shape of the femoral neck, long bones, and vertebrae defy accurate measurement in vivo. The femoral neck is fairly circular in children, but during advancing age, the structure is more oval, with longer diameter “seen” by the densitometer while the shorter depth in the anteroposterior direction is not measured using densitometry. The assumption of circularity of the femoral neck overestimates external volume of the bone and therefore underestimates its apparent volumetric BMD.

There is no single thickness of the cortical shell of a long bone. Just as the diameter of the long bone varies from millimeter to millimeter along its whole length, the thickness of the cortex varies at every point in the anterior, posterior, lateral, and medial direction and at every level from proximal to distal. Thus, a single “cut” by QCT or densitometry at the midpoint of the femoral neck hardly provides an estimate of the whole bone volume or of the cortical thickness of the bone, and estimation of static measures of strength, such as the section modulus derived from these single cuts, should be viewed cautiously.

Problems with vertebral dimensions are similarly challenging. The vertebral body is not a cube. It has no single height, width, or depth. Its CSA is ellipsoid. Estimates of apparent volumetric BMD assuming a cube overestimate the volume of the vertebral body. Derivation of the volumetric density using posteroanterior scanning includes the BMC of posterior processes; thus, inaccuracies result when volumetric BMD measurements are derived using projected BMC by posteroanterior scanning divided by an estimated volume (derived from the projected area). The inaccuracy is sex specific because females have more BMD of the whole vertebrae in the posterior processes than do males. Whether the proportions of bone mineral distributed in the vertebral body and posterior processes differ by race is unknown.

The important contributions of Tom Beck in developing estimates of bone diameter, cortical thickness, medullary area, section modulus, and buckling ratio have given us a door into bone structure and the biomechanical determinants of bone strength by using the output of the densitometer to derive these measures. A great many studies of structural changes in skeletal growth, aging, sex and racial differences in bone structure, and the effect of drug therapy and exercise have applied these methods to obtain a better understanding of fracture risk.

This is a step in the right direction (i.e., toward bone strength). However, derived estimates of structure and strength need to be validated in vitro, and prospective studies are needed to test whether the calculated measures of strength using biomechanical principles predict fractures. The assumption that TMC or “true” mineral density of bone tissue is a constant needs validation, particularly during growth and parathyroid hormone administration when the tissue mineral density of the newly deposited bone may be changing. The CSA derived at the “narrow neck” is not a conventional CSA, but a “collapsed” CSA. If increased, it is not clear whether the cause would be periosteal or endocortical apposition.

Thus, it is difficult to evaluate the veracity of the geometric and structural properties of many studies presented at the ASBMR. For example, Hillier et al. () reported that derived hip structural variables predict fracture, but that they do not improve the prediction for hip fracture above BMD alone. Rivadeneira et al. () reported an association between insulin-like growth factor 1 gene promoter polymorphism and hip bone geometry and risk of hip fracture. The authors suggest that this polymorphism influences the distribution of bone mass in a sex-specific pattern in males.

Uusi-Rasi et al. () reported altered sensitivity to loading in patients with osteoporosis and fractures. Bone strength to load is expressed as the section modulus divided by the fat-free mass (measured by bioelectrical impedance). In approximately 2800 women followed for almost four years, bending strength declined and buckling ratio increased faster in hip fracture subjects than in fracture-free subjects. The authors inferred that mechanosensitivity is impaired in women with osteoporotic fracture, and because this was found in individuals with a family history of hip fracture, suggested that this may be an inherited trait.

MacKelvie et al. () reported that an eight-month exercise program in prepubertal boys resulted in increased diameter of the femoral neck and increased cortical thickness measures of bone strength using the hip structural analysis program. Price et al. () reported the results of a 10-year follow-up in patients exposed to calcium supplementation. Section modulus increased, whereas BMD decreased.

Yates et al. () reported that hip structural analysis is a predictor of incident hip fracture. However, Duboeuf et al. () reported that parameters of hip geometry predicted fracture risk in elderly women, but that these parameters, which were related to BMD, were no better than BMD as predictors. The investigators also reported that after adjustment for BMD, none of the odds ratios remained significant. Le Bras et al. () examined whether geometric parameters calculated from three-dimensional reconstruction added to areal BMD improved fracture rate prediction. The fracture rate correlated best with femoral neck areal BMD, and adding geometric parameters improved the estimate (i.e., R2 increased from 62% to 73%).

The periosteum

The periosteum is the most neglected surface of bone. Larger bones in young men have a greater CSA upon which larger muscles exert the same load per unit area — that is, stress on bone CSA is no different in young males than in young females because of the scaling in nature. Fragility fractures are uncommon in young adults because loads are well below the ability of bone to withstand them. Structural failure emerges during aging, in part because periosteal bone formation fails to completely offset the fragility produced by bone loss and architectural destruction that proceeds on the endosteal surfaces of the bone. One reason that men fracture less commonly than women could be because men have greater periosteal apposition than do women. Concern has been raised that the femoral neck lacks a periosteum, but there is now evidence from Bliziotes et al. () that periosteal bone remodeling is occurring, at least in nonhuman primates.

The challenge is to measure each of the material and structural determinants of bone strength and to define the material and structural abnormalities that account for changes in bone strength in compression, tension, and torsion. In doing so, it may be possible to more accurately identify women likely to sustain fracture, target specific treatment to those most likely to benefit, and avoid treating those less likely to respond. There is progress.

See Recommended Readings(,,,,,,)

Secreted factors and regulation of bone mass Pamela G. Robey, National Institute of Dental Research, NIH, Bethesda, MD, USA

Although the critical role of members of the transforming growth factor-β (TGF-β) superfamily in skeletal development and maintenance has long been recognized, the manner in which they are precisely regulated and the signaling pathways that they elicit continue to be a major focus of ASBMR annual meetings. Based on several presentations, it is clear that bone morphogenetic protein (BMP) activity, at least, must be tightly regulated through interaction with competitive antagonists. Too much or too little activity can have profound effects on the skeleton. BMP-4 expression under control of the bone-specific type I collagen promoter causes perinatal lethality, and by examining embryos, it was determined that although bone was formed, it was undervascularized and characterized by an increased number of osteoclasts, suggesting that persistent expression of BMP-4 may induce osteopenia. In contrast, expression of Noggin, a BMP antagonist under control of the same promoter, prevents hypertrophy of cartilage rudiments and subsequent bone formation leading to perinatal lethality (). However, not all BMP antagonists exhibit the same methods of action. Overexpression of the antagonist gremlin, controlled by the osteocalcin promoter that becomes active postnatally, results in severe dwarfism. Histomorphometric analysis indicates decreased trabecular number of normal thickness, but increased cortical thickness, composed largely of woven bone. Osteoblast number was found to be reduced, as opposed to what had been described previously in transgenic mice overexpressing Noggin under the same promoter, in which trabecular number was also decreased, but in which osteoblast number was unchanged ().

Determination of how different BMP antagonists elicit varying responses in osteogenic cells is in its infancy. However, some new clues have been provided by studies on sclerostin, another member of the DAN family that antagonizes BMP activity. Previously, it was determined that humans with a rare high bone density disease are homozygous for null mutations in the SOST gene that codes for sclerostin, and which is highly expressed in osteocytes. Proof of principle was provided by the demonstration that transgenic mice, in which SOST was deleted, presented a progressive increase in bone density in all skeletal elements as a function of age, mimicking the course of the human disease (). In contrast, when SOST was overexpressed in mice under control of the osteocalcin promoter, the mice displayed some developmental abnormalities (e.g., smaller size at birth, extra floating ribs, or kinky tails caused by fusion of vertebrae). Trabecular number, cortical thickness, and bone mineral density were reduced in vertebrae (). Similar results, in which the human SOST gene was overexpressed in mice, were obtained by another group (). Histomorphometric analysis was not presented, so it is not clear how this form of osteopenia relates to that of Gremlin or Noggin transgenic animals. Based on the pattern of expression of sclerostin in osteoblasts and osteocytes, experiments were conducted to determine if sclerostin is involved in wnt signaling, using a Tcf/Lef-induced luciferase assay as a readout of wnt binding to the low-density lipoprotein receptor-related protein 5/frizzled receptor complex. When sclerostin was expressed, activation of the construct was markedly reduced ().

Intracellular circuitry: It's a (s)mad-(s)mad world!

The complexity of the intracellular signaling that is initiated by binding to BMP and TGF-β receptors continues to be revealed through molecular approaches to systematically eliminate the players that have been identified to date (in either transgenic animals or through antisense inhibition of translation). Based on its inhibitory role in mediating signaling from BMP and TGF-β receptors, transgenic mice overexpressing Smad6, specifically in chondrocytes, were generated. Although embryonic development was for the most part normal, with appropriate formation of the proliferative zone and expression of Sox9 and type II collagen, longitudinal growth was compromised after birth, type X collagen expression was diminished, and the primary spongiosa was decreased in size, suggesting that excess Smad6 signaling prevents the final stage of chondrocyte maturation required for proper deposition of bone (). Menin, mutations of which lead to multiple endocrine neoplasia, has also been found to play a role in osteoblastic differentiation, and a number of presentations focused on the interaction of menin with Smads in mediating BMP and TGF-β-induced signaling. Menin was found to coimmunoprecipitate not only with Smad 1/5, but also with Runx2, after BMP treatment. However, antagonism of osteoblastic differentiation by menin antisense only influenced alkaline phosphatase and Runx2 expression in osteoprogenitor cells and not in more mature osteoblastic cells, suggesting a stage-specific effect (). Although studies looking at the role of menin in TGF-β-induced signaling used pituitary cells, it was also found that menin interacts directly with Smad3 and that deletion of menin residues 101-195 reduced TGF-β-induced cell proliferation (). Further studies are needed to determine if menin has the same influence on TGF-β-induced changes in osteoblastic cells.

Regulation of transcription

New information on the control of gene transcription by new or unexpected partnership, and by isoforms of well-known transcription factors, were presented, several of which are highlighted here. Insulin-like growth factor binding protein 6 (IGFBP-6), which resides in the nucleus, has been known to inhibit osteoblastic differentiation (). The yeast two-hybrid system identified LMP-1, a transcription factor recently identified as a regulator of osteoblastic differentiation, as the partner of BP-6, suggesting that the inhibitory action of BP-6 is via sequestration of LMP-1 in the nucleus. Although the importance of estrogen and TGF-β in regulating bone homeostasis is well understood, cross-talk between their signaling pathways is not. Estrogen receptor (ER)-α was found to interact with Smad4 via the DNA-binding domain of Smad4, and in doing so, prevented Smad4 binding to the promoter of genes (PAI and Smad7) known to be upregulated by TGF-β (). Further characterization of the cross-talk between these two pathways may provide a basis for understanding how the opposing effects of TGF-β on stimulating proliferation of both osteoblastic and osteoclastic progenitors is modulated. The role of another binding partner of ER-α, steroid receptor coactivator-1 (SRC-1), which enhances transcriptional activities of steroid receptors, was studied by deleting its activity in transgenic mice. These mice developed a high turnover form of osteopenia, which was only moderately overcome by administration of sex steroids. Of interest, SRC-1 deficiency had no effect on glucocorticoid-induced osteopenia, suggesting that SRC-1 is essential for the anabolic action of sex hormones, but not for the catabolic action of glucocorticoid, at least (). In another study, the role of inducible cAMP early repressor (ICER) isoforms in modulating cAMP-induced genes was investigated in transgenic animals by enforcing expression with type I collagen promoters. After stimulation of cAMP production, calvarial cells isolated from transgenic animals expressed higher levels of the cAMP-inducible gene, Cox-2, but lower levels of other cAMP-inducible genes, c-FOS and interleukin-6, than did wild-type cells, suggesting that ICER isoforms interact with both positive and negative regulators of these genes (). In another study, it was determined that transgenic mice that overexpress the Δ2ΔFosB isoform, which can heterodimerize and bind to DNA, but which lacks the ability to activate transcription, are osteosclerotic and have decreased adipose tissue, similar to that found in Δ2FosB transgenic animals. The authors suggest that these naturally occurring isoforms may mediate their proosteoblastic activity by positive interactions with other members of the AP-1 family, Runx2 and Smads, and by inhibiting the action of adipogenic transcription factors ().

All of this represents a smorgasbord of new information, but we still have a long way to go to fully understand how osteogenic cells tick.

Osteoporosis: A disease of affluence or of disadvantage and deprivation? Mark A. Kotowicz, University of Melbourne, Melbourne, Australia

A substantial worldwide increase in hip fracture number has been predicted, based on an assumption of stable incidence rates and projected changes in population demographics associated with an aging population (). Hospital discharge data and self-reported race/ethnicity classifications from California indicate a significant decrease in hip fracture incidence in non-Hispanic white and Asian women from 1983 to 2000. In contrast, fracture rates in Hispanic American women and men (which were originally lower) increased significantly, while remaining stable between the sexes and in other ethic groups (). As in Asian populations, this apparent increase in fracture incidence in Hispanic Americans is likely related to lifestyle factors, including reduced physical activity, smoking, alcohol consumption, and dietary factors (). Mexican Americans have the highest prevalence of metabolic syndrome, associated with increasing obesity and decreasing physical activity, in the U.S. population (). Despite higher bone mineral density (BMD) at the femoral neck and lumbar spine, subjects with type 2 diabetes in the Rotterdam Study had increased adjusted nonvertebral fracture risk (hazard ratio [HR], 1.4 [1.0, 1.8]). After adjusting for confounders and BMD, increased risk was associated with hypoglycemic agent use (HR, 1.7 [1.2, 2.5]) (). Increased fracture risk (relative risk [RR], 1.6 [1.1, 2.5]), despite higher BMD in black and white participants (2979 men and women; mean age, 74 ± 3 years) with type 2 diabetes, was also reported from the Health, Aging, and Body Composition Study. Diabetic black women had an almost twofold higher fracture incidence than did nondiabetic black women ().

Age- and sex-adjusted incidence rates for osteoporosis-related fractures were linked to socioeconomic deprivation in North Glasgow residents. Furthermore, fracture cases in the lowest socioeconomic strata were less likely to accept diagnostic evaluation or secondary prevention therapies, even though such interventions are available without direct cost under the National Health Scheme (). The influence of social deprivation on fracture incidence is likely to be multifactorial. Osteoporosis and/or fracture risk factors are likely to be more prevalent in lower socioeconomic strata. Low calcium and low protein intake seem to act synergistically in elderly men and women, producing greater bone loss at the proximal femur (). Findings of a 10-year prospective, population-based study suggest that factors associated with frailty and debility, such as reduced serum albumin, low total triiodothyronine, and limited physical ability (e.g., ability to run for a bus), are significant predictors of vertebral (but not nonvertebral) fracture, even after adjusting for lumbar spine or hip BMD ().

Lifestyle factors operating during adult life, however, seem to be poor predictors of long-term hip fracture risk. Over 50 years of follow-up, only smoking in middle-aged men predicted hip fracture (). In contrast, experience of severe deprivation in childhood, related to internment in concentration camps, is associated with a threefold higher prevalence of self-reported hip fracture. This effect seems largely attributable to increased fracture risk (odds ratio [OR], 13.1) in individuals aged 16 to 17 years or younger at the onset of World War II (). On the other hand, skeletal benefits of adolescent/childhood exposures may persist into old age. A small longitudinal study reported that former athletes maintained higher total body and femoral neck BMD than did controls five years after cessation of sporting activity, and in a case-control study, former soccer and ice hockey players (n = 400) had reduced prevalence of self-reported fractures after 50 years (). These data support the hypothesis that exposures occurring before attainment of skeletal maturity may have a profound effect on subsequent bone fragility ().

Additional support for early life influences on bone strength comes from the association between umbilical vein calcium concentrations, maternal vitamin D status during late pregnancy, and whole body bone mineral content (WBBMC) at age nine years (). Lower maternal social class and educational attainment, smoking during pregnancy, reduced prepregnant weight and height, and late pregnancy fat stores were associated with lower childhood WBBMC (), suggesting that transgenerational effects related to maternal exposures may be operative in the association between low socioeconomic status and bone fragility.

Vitamin D level may also influence BMD in young adults. In 200 young men (age range, 18.6-20.6 years), hypovitaminosis D, defined as serum 25-hydroxyvitamin D </= 37.5 nmol/L, was almost universal in winter (94.6%), and summer prevalence approached 30%. A positive correlation between serum 25-hydroxyvitamin D and lumbar spine, femoral neck, trochanter, and total hip BMD was observed. Vitamin D level correlated with BMC, but not with scan area (). Influence of 25-hydroxyvitamin D concentration on lower limb musculoskeletal function was reported using cross-sectional data from the National Health and Nutritional Examination Survey III (NHANES III) (n = 4100; adults older than age 60 years). Individuals in the lowest quintiles for 25-hydroxyvitamin D had decreased walking speed and longer sit-to-stand times. Greatest improvement in these functional measures occurred when 25-hydroxyvitamin D level increased from 22.5 nmol/L to 94 nmol/L (). Results of the Osteoporosis Population-Based Risk Assessment study (1044 women; mean age, 75 years) showed that vitamin D level was correlated with fall-associated gait variables, Romberg test score, self-reported activity level, but not with muscle strength. Vitamin D-deficient women (4% of the sample) were at increased risk of falling after one year of follow-up (OR, 2.09 [1.03, 4.25]). Vitamin D deficiency was a predictor of hip fracture (OR, 4.10 [1.35, 12.46]) and of having more than two low-trauma fractures (OR, 3.85 [1.09, 13.64]). Vitamin D status was not a significant predictor of falls or fracture if gait or balance variables were included in models (). Seasonal periodicities of serum 25-hydroxyvitamin D and parathyroid hormone (PTH) levels were documented in south-eastern Australia and corresponded to a seasonal pattern of serum CTx occurring approximately one month after the peak in serum PTH. Furthermore, seasonal periodicity of hip and Colles' fractures occurred 0.5 to 3 months after the peak in serum PTH, and for Colles' fracture, coincided with peak serum CTx (). Thus, vitamin D insufficiency may be associated with failure to achieve optimal BMD, increased bone fragility related to higher bone turnover and PTH concentrations, and increased propensity to fall because of lower limb musculoskeletal impairment.

Poor dietary folate intake associated with alcohol abuse or poor nutrition can elevate plasma homocysteine levels () and may have relevance for fracture risk in lower socioeconomic groups. Plasma homocysteine concentrations for men (>/= 15 µmol/L) and women (>/= 13.8 µmol/L) in the highest quartile in the Framingham Study were associated with increased risk of hip fracture (men: HR, 3.84 [1.38, 10.7]; women: HR, 1.92 [1.18, 3.10]) (). Similar estimates were observed for hip fracture risk associated with plasma homocysteine concentrations for individuals in the upper quartile in the Amsterdam Longitudinal Aging Study (men: RR, 3.3 [1.1, 1.9]; women: RR, 1.8 [1.2, 4.1]), and significant positive association with serum osteocalcin and deoxypyridinoline, and negative association with broadband ultrasound attenuation, were also observed (). Mechanisms behind the increase in bone fragility related to elevated plasma homocysteine concentrations remain to be elucidated.

More on vitamin A

The status of vitamin A intake as a risk factor for osteoporosis and fracture remains unclear. Analysis of 22-year follow-up data (µ = 13.7 ± 6.3 years) from the NHANES I epidemiologic follow-up study (3219 men; age, 40-74 years; 51 incident hip fractures), suggests that the relationship between serum vitamin A level and hip fracture risk may be biphasic. RR for hip fracture was 3.8 (1.4, 10.6); 2.8 (1.0, 8.1); 1.0; 0.7 (0.2, 2.8;, and 3.3 (1.2, 9.5) (95% confidence interval), from lowest to highest quintile, respectively (). Total hip BMD seemed to be positively associated with vitamin D intake in 5995 men (mean ± SD daily intake, 10,597 ± 8567 international units [IUs]) in the Osteoporotic Fractures in Men study (), whereas vitamin A intake in 100 postmenopausal women (average daily intake, 10,034 ± 4352 IUs) was significantly associated with decreased BMD at the proximal femur, forearm, and total body ().

Other epidemiological studies

In a study of 2131 women (age, 80 years or younger), sleep patterns were assessed for 72 hours using a piezoelectric sensor to detect fine movement. Excessive self-reported daytime sleepiness increased risk of self-reported falls, ascertained prospectively using a postcard system (RR, 2.6 [1.3, 5.4]), and sleep disturbance (i.e., reduced efficiency, increasing number of wake episodes, or total sleep time less than 4 hours) increased risk of falls by 50% to 80% ().

After mean follow-up of 10.4 years, hip structure analysis (HSA) from proximal femur DXA scan in 7646 participants in the Study of Osteoporotic Fractures predicted hip fracture. Combining HSA parameters and BMD did not improve the models, and receiver operating characteristic analysis suggested that BMD and HSA are superimposable in sensitivity and specificity for fracture prediction, although BMD was a better predictor than HSA for individuals with BMD below -2.5 SD. Correlation between BMD and HSA may account for these observations. However, among 908 women who did not report any falls during follow-up, BMD was not a significant predictor of fracture, and combined HSA parameters were able to explain 8% of the variance, with the HR for buckling ratio increasing from 1.6 to 3.7. HSA may have a role in fracture prediction in women with BMD greater than -2.5 SD ().

Osteoporosis treatment Ian R. Reid, University of Auckland, Auckland, New Zealand Vitamin D

The therapeutics of osteoporosis was a significant focus of the meeting, with the presentation of new data relating to existing agents and clinical data on novel compounds. In the former category, a study on the treatment of vitamin D deficiency found that daily administration of 5,000 to 10,000 international units of vitamin D for three months is necessary to normalize serum 25-hydroxyvitamin D levels in patients with vitamin D deficiency (). The long-term continuation of such doses is likely to be associated with vitamin D intoxication and is therefore not appropriate, but the study does suggest that the low-dose daily regimens in widespread use are likely to take a prolonged period of time to restore normal vitamin D status.

Bisphosphonates

Additional follow-up data were presented from a multinational study of risedronate, in which 83 women with postmenopausal osteoporosis received risedronate (5 mg/day) for the entire seven-year study period (). The decline in vertebral and nonvertebral fracture rates, which occurred in the placebo-controlled phase of the study, was maintained to study end, suggesting sustained antifracture efficacy from this intervention.

Data from a new study of oral clodronate were presented (). The study involved randomization of 5592 women (age, 75 years or older) to clodronate (800 mg/day) or placebo. This was a population-based sample, and subjects were not selected for clinical or other evidence of osteoporosis. Nineteen percent of the cohort was osteoporotic, based on total hip bone mineral density (BMD) at trial entry. The number of clinical fractures was significantly reduced by 20% over three years, and the number of osteoporotic fractures was significantly reduced by 23%. Hip fracture incidence did not differ between groups. Antifracture efficacy was independent of body mass index (BMI), age, baseline bone density, and prior fracture history. This result significantly expands the indications for the use of bisphosphonates, because up to this time, antifracture efficacy had only been demonstrated in patients with baseline evidence of osteoporosis. These findings are similar to those recently published from the Women's Health Initiative showing antifracture efficacy of estrogen to be independent of baseline fracture risk ().

Data were also presented suggesting that the beneficial effect of bisphosphonates might extend to cartilage. Garnero et al. presented data from a trial comparing the efficacy of alendronate and risedronate in the treatment of postmenopausal osteoporosis, which reported decreases in the C-telopeptide of type II collagen, a marker of articular cartilage turnover (). Both alendronate and risedronate caused significant decreases in this marker, although the decreases were substantially less than those seen in the C-telopeptide of the type I collagen. However, these results do raise the possibility that bisphosphonates may be chondroprotective. This would need to be explored with more direct measures of osteoarthritic activity.

Hormone replacement therapy

A new analysis of the Women's Health Initiative () recalculated the global index (a measure of the overall risk-benefit profile) according to baseline risk of osteoporotic fractures. This demonstrated that even in high-risk subjects, intervention was still associated with an excess of adverse events. It should be noted that this global index only counts hip fractures and that the inclusion of other fracture types may change this outcome.

A strategy that might circumvent the adverse outcome of conventional hormone replacement therapy is intervention with a much lower dose of estrogen. This was explored in the Ultra Low-dose Transdermal Estrogen Replacement Assessment (ULTRA) study (), in which 417 women (age, older than 60 years) were randomized to transdermal estradiol (12.5 µg/day) or placebo for two years. The intervention reduced serum osteocalcin by 16% and increased spine and hip BMD by 2.4% and 1.5%, respectively, at two years. This finding opens up a whole new range of therapeutic possibilities, although there is no immediate prospect of a determination being made of the nonbone risks and benefits of this intervention.

Parathyroid hormone

Much new trial data were presented with respect to parathyroid hormone (PTH) treatment of osteoporosis. Further analyses from the Lilly Phase 3 study, which had previously reported increases in BMD associated with PTH, demonstrated that there were also increases in cross-sectional area and cortical thickness and a reduction in buckling ratio in the proximal femur (). Such geometric changes would be expected to contribute to increases in bone strength, which suggests that stimulation of periosteal growth caused by PTH may contribute significantly to its antifracture efficacy.

Lindsay et al. () presented results from quadruple tetracycline-labeled bone biopsies in patients starting PTH that demonstrated increased mineral apposition rates and bone formation rates at cancellous and endosteal surfaces in those taking PTH. Twenty-five percent of posttreatment bone formation occurred on surfaces that showed no evidence of prior resorption, a phenomenon that seems to be unique to PTH treatment.

McClung et al. () reported an 18-month randomized, controlled trial comparing PTH with alendronate in postmenopausal women with osteoporosis. As expected, PTH produced greater increases in bone density at most sites than did alendronate, and its use was also associated with greater decreases in pain than was alendronate use.

Several studies addressed the important question of the interaction between PTH and antiresorptive therapies. The Parathyroid Hormone and Alendronate for Osteoporosis (PaTH) study randomized 238 postmenopausal women with BMD T score < -2 to PTH, alendronate, or a combination of the therapies. Bone turnover markers were almost doubled with PTH, and with alendronate, more than halved. With combination therapy, formation markers increased at one month, but subsequently both formation and resorption markers paralleled changes seen with alendronate alone. PTH produced greater changes at the spine than did alendronate, whereas the reverse tended to be the case in the hip. Combination therapy resulted in changes that were intermediate between those found with the respective monotherapies. This study has now been published in full ().

Ettinger et al. () studied the effect of 18 months of PTH treatment on women who had just stopped taking either alendronate or raloxifene. The responses of bone markers and BMD in the raloxifene group were similar to those seen in treatment-naive patients, whereas the responses were generally less in women who had previously been taking alendronate. This implies a blunting of the anabolic effect of PTH in individuals in whom residual suppression of bone turnover from previous bisphosphonate use has occurred.

Cosman et al. () also studied patients established on alendronate therapy. Patients continued taking alendronate, adding either nothing, daily PTH injections, or cyclic PTH injections (three months on, three months off). In the two PTH groups, increases in spine BMD were comparable and were about 2% greater than increases seen in patients continuing on alendronate alone.

None of these studies is entirely comparable, but collectively they do suggest that PTH continues to have an anabolic effect in the presence of bisphosphonate treatment; however, the effect is somewhat attenuated. The relevance of these changes in markers and BMD to fracture prevention is, of course, unknown.

Novel therapies

The role of simvastatin was examined in 82 postmenopausal women with T scores < -1 (). At one year, there were no changes in bone turnover markers or in BMD of the spine or total body. In contrast, in the forearm, BMD was 1.5% higher in the statin group. It is unclear why the forearm scans performed differently from those at other sites, but the extent of the changes is not sufficient to regard oral simvastatin as a useful intervention in osteoporosis.

There is extensive laboratory evidence of nitric oxide action on both bone formation and bone resorption, and some epidemiological data suggest that the use of nitrate medications is associated with lower fracture incidence. Jamal et al. () reported on a 12-week study of biochemical markers. Treatment with isosorbide mononitrate (5 or 20 mg/day) reduced NTX by 36% to 45%, but increased bone-specific alkaline phosphatase by 16% to 23%. These changes accord with the in vitro actions of nitric oxide and suggest that this widely available and inexpensive medication may have a role in the management of osteoporosis.

Workers from Amgen presented the first clinical data with a monoclonal antibody to the osteoclastogenic regulatory factor, receptor activator of NF-κB ligand (RANKL) (). A single subcutaneous dose of this compound reduced bone resorption markers by as much as 80%, and in the upper dose range, the effect persisted beyond six months. In the 49 women studied, the agent seemed to be very well tolerated, which opens up the possibility of a novel class of antiresorptive agents for investigation in the management of osteoporosis and, possibly, in the management of other bone diseases.

---

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.