| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
CLINICAL REVIEW 165 |
Department of Medicine (B.F.), University of Heidelberg, D-69117 Heidelberg, Germany; and Bone Research Program, ANZAC Research Institute, and Department of Endocrinology, Concord Hospital Medical Centre (C.R.D., M.J.S.), University of Sydney, Sydney 2139, New South Wales, Australia
Address all correspondence and requests for reprints to: Prof. Markus J. Seibel, M.D., Ph.D., F.R.A.C.P., Bone Research Program, ANZAC Research Institute, and Department of Endocrinology and Metabolism, Concord Hospital Medical Centre, The University of Sydney, Sydney, New South Wales 2139, Australia. E-mail: mjs{at}anzac.edu.au.
 |
Abstract |
|---|
 Top Abstract Mechanisms of Metastatic Bone...Markers of Bone Remodeling...Future DevelopmentsReferences |
|---|
In patients with established bone metastases, most markers of bone remodeling are abnormal compared with healthy controls or cancer patients without bone lesions. Although bone markers may have a potential as diagnostic tools in cancer patients, the available data do not allow final conclusions regarding the accuracy and validity of any of the presently used markers in the diagnosis of bone metastases.
As regards monitoring of anticancer therapy, most markers of bone remodeling respond to active treatments. These indices therefore may have the potential to be used in the monitoring of antitumor therapies. However, most if not all of the available evidence on the use of bone markers in monitoring anticancer therapy is observational, and it remains unclear whether they have any beneficial effects on overall outcome. The same is true for their prognostic value, although evidence suggests that suppressed levels of bone formation or high rates of bone resorption are independent predictors of poor survival.
|
Mechanisms of Metastatic Bone Disease |
|---|
|
TopAbstractMechanisms of Metastatic Bone...Markers of Bone Remodeling...Future DevelopmentsReferences |
|---|
Bone formation can be increased or decreased but is inadequate to compensate for the increased bone resorption. Radiographically, this results in lytic or mixed lytic and sclerotic lesions that can be associated with bone pain, fracture, and neurological symptoms (2). In contrast, prostate cancer and most osteosarcomas typically produce sclerotic lesions characterized at the cellular level by a relative excess of bone formation compared with bone resorption (3). However, based on measurement of bone markers, bone resorption and formation are both markedly increased around prostate bone metastases (4).
Bone resorption around metastatic cancer foci is predominantly mediated by osteoclasts (5). Osteoclast differentiation and activation are regulated at the local level by the relative expression of receptor activator of nuclear factor-
B ligand (RANKL) and osteoprotegerin (OPG) (6). RANKL and OPG are mainly produced by cells of the osteoblastic lineage. RANKL acts directly on osteoclast precursors and mature osteoclasts through its receptor RANK to increase osteoclast differentiation and activation. OPG is a decoy receptor for RANKL (7). The relative expression of RANKL and OPG osteoblasts/stromal cells is regulated by, and mediates, the proresorptive effects of hormones such as PTH and 1,25(OH)2 vitamin D3, inflammatory cytokines, and cancer-produced factors such as PTHrP. RANKL expression is increased around tumors (8) and is induced in stromal cells by coculture with tumor cells (9). Cancer cells could increase RANKL expression in several ways. Tumor cells could themselves express RANKL, as has been reported for prostate cancer cells (10). Many tumor cells produce PTHrP, which has been clearly shown able to increase RANKL and decrease OPG expression in stromal cells (11). Other factors such as inflammatory cytokines and, in myeloma, macrophage inflammatory protein 1 (MIP1
) may also be expressed either by the tumor cells or by host cells in response to tumor and indirectly increase RANKL and/or decrease OPG expression (12).
Prostate cancer cells are thought to facilitate osteosclerosis by expressing the osteoclast-deactivating factor endothelin-1 (13), thus disrupting local bone turnover in favor of a net gain in mineral content and bone matrix production. Furthermore, prostate cancer cells can enhance bone formation by direct osteoblastic stimulation through the N-terminal fragment of urinary plasminogen activator, a growth factor that has been detected in conditioned medium of the human prostate cancer cell line PC-3 (14).
The process of cancer metastasis to bone is multifaceted. The primary tumor must have close association with the vascular system (through promotion of angiogenesis), and the tumor cells need to develop the ability to invade these vessels (intravasation). These steps are likely dependent on motility, chemotaxis, and expression of matrix proteases, particularly matrix metalloproteases. Once the tumor cells invade the vasculature, they can be carried throughout the body. It is likely that multiple factors determine the colonization of tissues and establishment of metastases by circulating tumor cells. Tumor cells must escape the vasculature (extravasation), survive in the host tissue, proliferate to a micrometastasis, and recruit sufficient blood supply for further growth (15).
Tissue perfusion rate determines the probability of a tumor cell passing through a particular tissue and is an important passive determinant of metastasis to different tissues (16). The nature of the vasculature is also likely to be important, with the sinusoidal structures of blood vessels in bone marrow possibly enhancing the ability of tumor cells to escape into the surrounding bone (3). Overall, tissues most frequently the targets of metastasis tend to be the most highly perfused organs, such as lung and liver. However, bone marrow has a slower rate of perfusion than these organs, and tumors of different types metastasize preferentially to different tissues, so other factors must be involved.
Preferential targeting of tumor cells to particular tissues could be an active process with chemotactic signals produced by normal tissue functions directing the movement of tumor cells from the blood vessels to the tissues (17). Alternatively or additionally, tumor cells could move randomly into many tissues but find only in particular tissues an environment conducive to survival and growth.
Continued growth of metastatic foci requires modification of the microenvironment surrounding the tumor cells. Angiogenesis is required to provide the nutritional requirements of the tumor (18). However, removal of physical constraints to growth and enhancement of cytokine and growth factor support provided by the target tissues are also involved. It is likely that bone remodeling activity contributes to many of these processes, and thus inhibition of bone remodeling through the use of antiresorptive therapies has the potential to affect tumor metastasis and growth as well as its established bone protective role.
In young growing nude mice, human breast tumor cells injected by the intracardiac route, target to highly vascularized regions associated with high bone turnover near the growth plates (19, 20). Osteoblasts and osteoclasts may secrete or may lead to the release factors that are chemotactic for tumor cells and promote their active migration from the vasculature into bone. Osteoclasts and osteoblasts each produce a characteristic range of cytokines and growth factors that could support local tumor cell growth, as may be the case of osteoclast generated IL-6 production supporting multiple myeloma (MM) cell survival (21). Bone matrix contains many growth factors that can be released in active form through osteoclastic activity (for example, TGFß, fibroblast growth factor 1 and 2, platelet-derived growth factor, IGF, bone morphogenetic protein) (22). All of these factors have been shown to be chemotactic agents promoting tumor cell invasion, growth, or survival.
Mundy and Guise (21) have developed a concept of a vicious cycle involving osteoclasts and tumor cells in bone (Fig. 1
). Tumor cells act locally to increase osteoclastic bone resorption through secretion of proresorptive factors (most importantly PTHrP) or by cell-to-cell interactions mediated by adhesion molecules [such as bone sialoprotein (BSP)]. PTHrP or cell-to-cell interactions increase osteoblast/stromal expression of RANKL, which promotes the differentiation, activation, and survival of osteoclasts. These osteoclasts resorb bone, releasing active TGFß and other growth factors that may both enhance proliferation of the tumor and increase tumor expression of PTHrP, completing a destructive cycle (23, 24).
|
|
Markers of Bone Remodeling in Metastatic Bone Disease |
|---|
|
TopAbstractMechanisms of Metastatic Bone...Markers of Bone Remodeling...Future DevelopmentsReferences |
|---|
Bone undergoes constant remodeling. Whereas resorption of old bone and formation of new bone are balanced under normal conditions, metastatic bone disease leads to a pronounced imbalance in these processes. Over the past decade, the isolation and characterization of cellular and extracellular components of the skeletal matrix have resulted in the development of biochemical markers that specifically reflect either bone formation or bone resorption. These new biochemical indices have greatly enriched the spectrum of analytes used in the assessment of skeletal pathologies and seem ideal tools to evaluate the changes in bone remodeling associated with the metastatic process. Although the various serum and urinary markers of bone remodeling include cellular derived enzymes, nonenzymatic peptides, and mineral components, they are usually classified according to the metabolic process they are considered to reflect. For clinical purposes, therefore, markers of bone formation are distinguished from indices of bone resorption (Fig. 2 and Tables 1 and 2).
|
|
|
Biochemical markers of bone metabolism: biochemistry
Markers of bone formation. Bone formation markers are direct or indirect products of active osteoblasts. The most commonly used markers of bone formation are alkaline phosphatase (ALP), OC, and the type I collagen propeptides. Importantly, these components are expressed during different phases of osteoblast development and therefore reflect different aspects of bone formation.
ALP is a ubiquitously expressed, cell-membrane-associated enzyme. The total ALP serum pool consists of several dimeric isoforms that originate from various tissues such as liver, bone, intestine, spleen, kidney, and placenta. In addition, certain tumors may express macromolecular forms of ALP (e.g. Nagao ALP). In healthy adults, the liver and bone (BALP) isoforms each account for approximately 50% of the total ALP activity in serum. Thanks to the wide availability of inexpensive detection methods, total serum (s)ALP is still the most often used marker of bone formation. Common causes of elevated sALP levels are metabolic bone disease or cholestatic liver disease. Once liver disease is ruled out, total sALP provides a good impression of osteoblast activity. However, due to its higher specificity, serum bone-specific ALP (sBALP) is increasingly preferred (25, 26).
OC is a 5.8-kDa protein produced by osteoblasts, odontoblasts, and hypertrophic chondrocytes (27, 28). Although OC is present in significant amounts in bone, dentin, and calcified cartilage, it has recently also been found in osteosarcomas, ovarian, lung, brain, and prostate cancers (29). Most of the circulating OC is a product of osteoblast activity and therefore considered an index of bone formation. However, because OC is incorporated into the bone matrix, some fragments may also be released during bone resorption. In serum, OC is readily degraded so that both the intact peptide and fragments of various sizes coexist in the circulation. Assays detecting both intact OC and fragments are therefore considered most appropriate for the measurement of OC in human serum (30). Although there is a good correlation between histomorphometric parameters of bone formation and sOC levels (31), the latter are significantly influenced by age, gender, and kidney function (32).
Type I procollagen propeptides—Collagen type I makes up 90% of the organic bone matrix and is secreted by osteoblasts in the form of procollagen. Extracellular processing of procollagen before fiber assembly includes cleavage of the N- and C-terminal extension propeptides. Because these peptides are generated in a stoichiometric 1:1 ratio with newly formed collagen molecules, their levels in serum are considered an index of collagen synthesis and thus of bone formation (33). Serum levels of the C-terminal propeptide of procollagen type I (PICP) have been demonstrated to correlate with histomorphometric measures of bone formation (34), and hormone replacement or bisphosphonate therapy leads to a reduction in the circulating concentration of this marker (35, 36). Most recent studies, however, suggest that the N-terminal propeptide of procollagen type I (PINP) has a greater diagnostic validity than PICP. In particular, anabolic agents such as PTH seem to greatly affect PINP serum levels. The N-terminal propeptide of type III collagen is considered a marker of collagen type III synthesis, and thus of connective tissue metabolism.
Markers of bone resorption. Most markers of bone resorption are degradation products of skeletal (type I) collagen. However, osteoclast-specific enzymes such as tartrate-resistant acid phosphatase (TRAcP) 5b and noncollagenous proteins such as BSP have also been demonstrated to reflect bone resorptive processes.
OHP is formed intracellular from the posttranslational hydroxylation of proline. Most of the OHP released during the bone resorption is primarily metabolized in the liver and subsequently excreted in the urine (33). Urinary (u)OHP is usually considered an index of bone resorption, although significant amounts of uOHP are derived from the degradation of newly synthesized collagens and, in certain instances, from C1q metabolism (37, 38). Due to this and other factors, uOHP is considered a less specific index of bone resorption and has been largely replaced by other markers.
Hydroxylysine is formed during the posttranslational phase of collagen synthesis and is incorporated into the bone matrix as a component of collagen molecules. Hydroxylysine occurs in two glycosylated forms, namely glycosyl-galactosyl-hydroxylysine and galactosyl-hydroxylysine (GHL). GHL is more specific for bone and is released into the circulation during collagen degradation (33). Both components can be measured in the urine by HPLC after derivatization (39, 40). The advantage of GHL over OHP as a marker of bone resorption is that GHL is not metabolized and not influenced by dietary factors.
The pyridinium cross-links of type I collagen, pyridinoline (PYD) and deoxypyridinoline (DPD), are widely used markers of bone resorption. Whereas PYD is widely distributed throughout the body tissues, DPD is only found in collagen of bone and dentin. Both components are released during collagen breakdown, and histomorphometric analysis shows that their urinary excretion is closely related to bone resorption (41). In contrast to OHP, the measurement of uDPD is not influenced by the degradation of newly synthesized collagens or by dietary collagen intake. In recent years, specific immunoassays have largely replaced the more cumbersome chromatographic techniques (42, 43).
The collagen type I telopeptides form short, nonhelical stretches at the N and C termini of the collagen molecule. During collagen degradation, N- and C-terminal peptide fragments of various sizes and cross-link content are released into the circulation (44, 45). An ELISA for the detection of the N-terminal cross-linked telopeptide (NTX-I) has been established and is available for measurements in urine and serum (46, 47). In addition, there are various assays for the measurement of epitopes related to the C-terminal telopeptide (ICTP, CTX-I, ß-CTX-I) in serum and urine (48, 49, 50). Paget’s disease of bone, primary hyperparathyroidism, rheumatoid arthritis, osteomalacia, and established metastatic bone disease all go along with significantly elevated levels of DPD or telopeptide markers. In postmenopausal osteoporosis, levels of uCTX-I and free uDPD are indicative of high bone remodeling and increased fracture risk and are used to monitor therapeutic efficacy (43, 50, 51, 52, 53).
TRAcP belongs to a family of acid phosphatase isoenzymes found in bone, prostate, platelets, erythrocytes, and spleen. The osteoclast-specific isoform, TRAcP 5b, is a lysosomal enzyme secreted by activated osteoclasts (54, 55). Whereas older kinetic assays measure the entire pool of sTRAcP (56), newer immunoassays are able to detect the TRAcP 5b isoform only (57, 58). Serum TRAcP 5b activity has been found elevated in Paget’s disease of bone, primary hyperparathyroidism, osteomalacia, and breast cancer-induced bone metastases (56, 59, 60).
Urinary calcium levels depend on kidney function; dietary calcium intake; phosphate, sodium, and vitamin-D3 levels; and bone resorption rates (61). Urinary calcium levels are therefore considered an index of overall calcium homeostasis rather than of bone resorption alone. However, bone resorption causes the release of calcium from the mineralized bone matrix. In tumor-induced osteolysis, osteoclast-induced release of calcium into the circulation can be quite significant, and substantial hypercalciuria and hypercalcemia may become apparent, always in conjunction with a marked suppression of serum PTH levels (61).
Markers of bone remodeling in the diagnostic assessment of bone metastases
The early and reliable diagnosis of bone metastases is of clinical relevance because it affects the prognosis and therapeutic intervention strategies. To be useful in a clinical context, markers of bone remodeling would be expected to identify patients with existing bone metastases (sensitivity), provide a good correlation with the extent of the skeletal disease, discern skeletal from soft-tissue metastases (specificity), and reliably predict disease progression or regression (predictive value). Although there is considerable information on the diagnostic use of bone markers in cancer patients (62, 63, 64, 65, 66, 67), the precise value of these indices in regard to the diagnosis, therapeutic response, and prognosis of skeletal metastases has not been studied in controlled trials. Presently, therefore, clinicians are well advised to continue to rely on a synopsis of the patient’s individual history and clinical, imaging, and laboratory data, and to use bone markers only as in an adjunct setting.
In cancer patients with clinically established disease, comparisons are usually made between groups of patients with known bone metastases and those free of such involvement. In postmenopausal women, the predictive value of sBALP alone to discriminate between breast cancer patients with or without bone lesions was found to be poor, but it was improved by additional measurement of a marker of bone resorption (68). Apart from these observations, distinct elevations in sALP have been seen in advanced cases of prostate and breast cancer, particularly when osteoblastic metastases were present (69, 70). Generally, sALP seems most useful when skeletal involvement is suspected in patients with prostate cancer. In these cases, combined measurement of sPSA and sBALP appears to greatly increase the diagnostic sensitivity for bone lesions, compared with healthy subjects or patients with benign prostate hyperplasia (71, 72). In metabolic bone disease with undisturbed osteoid maturation, sBALP levels are usually closely correlated with other markers of bone formation. In contrast, sBALP and sOC levels may dissociate in patients with advanced metastatic bone disease (metabolic uncoupling) (73). It is therefore not surprising that patients with breast cancer metastatic to bone often have normal sOC levels before treatment (74).
In normocalcemic cancer patients with prevalent bone metastases, sOC levels are often found elevated (75) but fall as metastatic hypercalcemia develops (76). In contrast, sOC varies greatly in patients with humoral hypercalcemia of malignancy and seems to remain unaffected by soft tissue metastases (77, 78).
In MM, sOC levels and disease progression or stage seem to be inversely correlated with each other. Low sOC concentrations are considered to indicate a suppression of osteoblast activity and have been associated with poor survival (78, 79, 80, 81). However, this association was not confirmed in other studies (82, 83), and recent studies indicate that sICTP is a better prognostic marker in MM than most other biochemical indices (81, 84).
Although the specificity of sPICP in detecting bone metastases seems to be comparable to some bone resorption markers, its sensitivity appears much lower (85). However, a recent study demonstrated that in patients with breast or prostate cancer, both sPICP and sPINP levels are elevated and that a decreased sPICP/sPINP ratio indicates a more aggressive phenotype with a higher propensity to metastasize to bone (86). Levels of sPINP and sPICP were significantly higher in patients with lung cancer metastatic to bone than in similar patients with no metastases or soft tissue metastases only. Although sPINP and sPICP levels did not predict survival in lung cancer patients (87), a significant correlation between postoperative sPINP levels and poor survival, tumor size, and malignancy grade was observed in a study of 373 patients with node-positive breast cancer (88). Koizumi et al. (89) suggest that sPINP is a useful marker in the evaluation of skeletal spread in patients with prostate cancer, with significantly higher sPINP levels in patients with bone metastasis.
In the past, uOHP and urinary calcium have been widely used to assess bone resorption. However, these parameters lack the necessary specificity to reliably detect bone metastases and are not closely correlated with disease outcome (90, 91, 92). Notably, both serum and urinary calcium concentrations are influenced by changes in tubular reabsorption of calcium, a function often affected by tumor-derived paracrine factors such as PTHrP. In patients with metastatic cancers, uOHP levels were poorly correlated with bone resorption but may reflect breakdown of tissue collagen in extended organ metastases (93). Some reports have evaluated the usefulness of GHL in patients with metastatic bone disease. In patients with breast cancer, this marker seems to be useful in the diagnosis of established bone metastases (94, 95), although so far no follow-up reports have appeared on this topic.
Markers of collagen (type I) degradation currently appear to be the most promising candidates for the biochemical evaluation of malignant bone disease. Approximately 80–95% of patients with breast, prostate, or lung cancers metastatic to bone have increased levels of uDPD (96, 97, 98). However, a significant proportion of patients without evidence for bone lesions on radiographs or bone scans also are found to have elevated urinary cross-link levels (97, 99) (Fig. 3). This observation has been attributed to the presence of undiagnosed bone metastases and/or to a tumor-induced systemic acceleration in bone turnover (100). Whatever the cause is, the clinical validity of uDPD (and probably most other bone resorption markers) to diagnose the presence of metastatic bone disease in patients with solid tumors seems low. This applies in the same way to patients with advanced MM; Pecherstorfer et al. (101) found significantly higher levels of uPYD and uDPD in patients with MM compared with healthy adults, patients with monoclonal gammopathy of undetermined significance (MGUS) or patients with postmenopausal osteoporosis. Although the diagnostic validity of uDPD was high in patients with advanced MM, measurement of uDPD did not help to discriminate patients with MGUS from those with early-stage (stage I) MM (Fig. 4). Increased uDPD levels seemed, however, to identify patients likely to benefit from bisphosphonate therapy (101, 102).
|
|
The clinical relevance of CTX-I vs. ßCTX-I, and particularly their separate measurements in patients with benign or malignant bone diseases remains unclear, and further studies are warranted before a definitive recommendation can be made. It seems, however, that the newer serum telopeptide assays are more sensitive to malignancy-induced changes in bone turnover than the older urine assays (50) (Fig. 5).
|
The value of bone markers to predict future bone metastases in cancer patients without malignant spread has not been studied extensively. The latter often demonstrate abnormally high levels of bone formation and resorption markers, and the significance of this observation remains unclear. The few studies that have looked at these questions were clearly negative. In a recently completed 5-yr prospective study, 3-monthly measurements of an extensive panel of bone remodeling markers did not predict subsequent (i.e. incident) bone metastases (115). The low predictive value of bone markers and the varying results in regard to the detection of bone metastases are likely to be attributable to their high overall and long-term variability in cancer patients (116). Similar observations were made in another, smaller study (Pecherstorfer, M., personal communication).
The question then remains whether bone markers are helpful in the early diagnosis of bone metastases. An older longitudinal study claimed that in 70% of the patients assessed, serial measurements of sBALP levels correctly identified patients with bone metastases, and that the biochemical diagnosis was made on average 7 months earlier than the assessment based upon clinical, radiological, and isotope techniques (117). We and others, however, have found measurements of bone markers useless to detect bone metastases at a preclinical stage (68, 115, 118).
Taken together, most markers of bone remodeling and particularly those of bone resorption are elevated in patients with established bone metastases. Although these observations strongly suggest that bone markers may have a potential as diagnostic tools in cancer patients, the currently available data do not allow final conclusions regarding the accuracy and validity of any of the presently used markers in the (early) diagnosis of bone metastases. The same applies to the prognostic value of abnormal marker results in patients with malignant tumors.
Markers of bone remodeling in the monitoring of antitumor therapy
The widespread use of novel and potent treatments for bone metastases has triggered a momentous demand for clinically useful, simple, and inexpensive tools to monitor therapeutic efficacy. There is little doubt that markers of bone remodeling, especially of bone resorption, are useful to assess the effect of bisphosphonates, which in patients with bone metastases have evolved as first-line therapies to inhibit tumor-induced bone resorption, alleviate pain, and decrease the incidence of pathological fractures (119, 120, 121). Biochemical indices are considered to reflect therapy induced changes earlier than any of the other techniques currently used in clinical settings.
Although sOC did not seem to be of significant diagnostic use in predicting bone metastases, the initiation of bisphosphonate or chemotherapy (CT) leads to a significant fall of sOC levels to, or even below normal values (75, 77, 122, 123). Conversely, in patients with MM or metastatic breast cancer, CT-induced remission leads to a normalization or increase of sOC levels (78, 80). This change in sOC levels is considered to reflect osteoblast recovery and osteogenic repair (124, 125).
Treatment of metastatic prostate cancer with pamidronate results in a reduction of both total sALP and sOC (126), and similar observations have been made in breast cancer patients with metastatic bone disease receiving iv ibandronate (127). However, other studies in patients with metastatic bone disease have reported an increase in total sALP and sBALP 1 month after initiation of treatment with clodronate, and this change was considered to reflect skeletal repair processes (128). Interestingly, patients without detectable bone lesions do not exhibit changes in sBALP or sOC during CT (74). Thus, measurement of sBALP or sOC may be a useful adjunct to other means of monitoring, although no controlled trials have been performed to test this hypothesis.
Markers of bone resorption, such as the pyridinium cross-links (DPD, PYD) or the collagen telopeptides (ICTP, NTX-I, CTX-I) have been found to quickly change in response to bisphosphonate or OPG treatment of cancer or MM patients (129, 130, 131, 132, 133, 134, 135). In patients with bone metastases, iv application of pamidronate or ibandronate results in a significant fall of uDPD and uPYD within 24–48 h after the infusion (127, 136, 137, 138, 139, 140, 141) (Fig. 6). With a lag period of 4–12 wk, markers of bone formation usually follow the change in resorption markers (139, 140). This observation is not unexpected, because osteoclasts are the primary target of bisphosphonates and the reduction in osteoblast activity occurs only as a secondary effect.
|
). Newer studies indicate that pretherapeutic bone resorption rates may predict the response to bisphosphonate treatment. In a double-blinded study on the effects of pamidronate on bone remodeling and clinical outcomes, Vinholes et al. (142) observed that cancer patients with high baseline uNTX-I were less likely to respond to treatment than patients with normal or only slightly elevated bone resorption markers. Also, an increase in uNTX-I excretion was associated with a higher chance of disease progression (143).
|
In the clinical setting, variability of bone markers should be of particular concern when it comes to serial measurements, for example during therapeutic monitoring. Often, a moderate reduction in a bone resorption marker is believed to be the effect of antiresorptive treatment, when it really should be attributed to nonspecific variability or to a regression to the mean. However, a true (significant) response in bone turnover can only be assumed when within a single individual, the change in signal is greater than the imprecision of the measurement. There are several ways to assess and standardize the ratio between true changes (signal) and nonspecific variations in marker levels (background noise). Most of these approaches are based upon the concept of signal to noise ratios, and the calculation of the least significant change (LSC) or minimum significant change (MSC). The LSC can be defined for various levels of confidence (e.g. 80 or 95%) and depends primarily on the short- and long-term within-subject variability (coefficient of variation) of a given marker. The coefficient of variation of bone formation markers is lower than that of most bone resorption markers, and so is their LSC. Thus, for formation markers, a change of more than 30% should under regular circumstances be considered significant, whereas for most bone resorption markers (serum and urine) the LSC is around 60–80%.
The pronounced variability and heterogeneity of bone markers makes is difficult to determine precise thresholds or cut-off values for practical use in individual patients. Using receiver operating characteristics analyses or logistic regression models, attempts have been made to define marker-specific cut-offs at 3–6 months into treatment, predicting the response in bone mineral density after 2 yr of therapy. However, all of these analyses have been retrospective in nature, and so far, none of these cut-off values has been tested in prospective studies (using fracture as an endpoint). In a routine clinical setting, in which exogenous influences may not be controlled easily, a more robust marker that is less susceptible to be biased by these effects might be preferable.
Independent from these considerations, most clinical studies in cancer patients with bone metastases indicate that a reduction in bone turnover (resorption) is associated with a favorable clinical outcome (see above). Therefore, whatever type of marker is being used to monitor treatment, the aim of bisphosphonate therapy should be to normalize increased rates of bone remodeling.
Monitoring urinary calcium excretion has been demonstrated to be useful for the follow-up of patients with advanced stage prostate cancer and established bone metastases (90, 91, 92). Independent of tumor origin, bisphosphonate treatment leads to a dramatic improvement in hypercalciuria and hypercalcemia caused by metastatic bone involvement (145). It is therefore not surprising that both serum and urinary calcium have traditionally been used to monitor response to effective therapy. In one study following cancer patients during bisphosphonate therapy (127), the urinary excretion of calcium and OHP normalized sooner than the excretion of uDPD or uPYD. Because there was no correlation between the urinary excretion of calcium and collagen cross-links in these patients, the authors suggested that bone demineralization and matrix degradation are not necessarily timely coupled processes (127). However, the observed changes may as well reflect the temporary imbalance between bone resorption and bone formation that follows commencement of antiresorptive treatment, because in this situation, a small reduction in bone resorption may induce a substantial reduction in urine calcium due to net influx of calcium into bone.
The effects of antiestrogens such as tamoxifen on markers of bone remodeling are quite different from those of the bisphosphonates. Given as an adjuvant therapy in patients with breast cancer metastatic to bone, tamoxifen induced an increase in pyridinium cross-links (146) and either a suppression (147) or no change (148) in bone formation markers. Changes in urinary calcium were ambiguous in most tamoxifen studies.
The effect of CT on bone markers varies in cancer patients, depending on the type of CT and whether or not glucocorticoids are included in the treatment regimen. It seems that most markers of bone formation change slowly after several cycles of CT as long as no glucocorticoids are involved. In contrast, serum levels of OC suppress profoundly and rapidly once cortisone is introduced (149). In patients with breast cancer, the progression of bone metastases after CT appears to be associated more closely with changes in serum ALP than with carcinoembryonic antigen or CA15.3. In this study, however, measurements of serum ALP were unable to distinguish between responders and nonresponders to CT (69). In patients with breast cancer and osteolytic bone lesions, a rise in serum OC or ALP/BALP after CT has in some studies been associated with focal recalcification and therefore interpreted as a sign of therapeutic success (150). However, the significance of these observations needs to be shown in further and larger studies. In patients with MM, high-dose CT with autografting normalized bone turnover, although these effects were slow to appear (151).
In some animal experiments, bone resorption markers seem to be more useful for the monitoring of hormone or CT than bone formation markers (152). In a study on breast cancer patients, uDPD, urinary calcium, and serum CA27.29 were compared in regard to their ability to detect bone metastasis. Although CA27.29 was the best index for the detection of advanced stage (stage IV) breast cancer metastatic to bone, the primary advantage of uDPD was seen in the monitoring of bone metastases during CT and pamidronate (118).
Taken together, markers of bone remodeling are potentially useful tools in the management and follow-up of patients with malignancies. In particular, markers of bone resorption such as the pyridinium cross-links, ICTP, NTX, and CTX, all react promptly and profoundly to bisphosphonate, hormonal, or chemical treatments and may therefore be used in the monitoring of such interventions. Notably, levels of bone resorption markers before and their changes during and after bisphosphonate treatment seem to be predictive of clinical outcome. Again, most if not all of the data are observational, and there have been no controlled studies on the usefulness of bone markers in the pre- and posttherapeutic monitoring of cancer patients. Although most studies with bisphosphonates indicate that normalization of bone remodeling rates may be advisable, it is presently unknown whether the use of bone markers has any beneficial effects on overall outcome. At the present stage, it certainly is impossible to decide which markers should be used in routine clinical setting. Thus, until appropriate controlled studies have become available, markers of bone remodeling may be used only as an adjunct in the management of cancer patients.
|
Future Developments |
|---|
|
TopAbstractMechanisms of Metastatic Bone...Markers of Bone Remodeling...Future DevelopmentsReferences |
|---|
Recently, BSP has emerged as a new marker of bone resorption in metabolic and metastatic bone disease. The glycoprotein is an osteoblast product of 
70 kDa, undergoes extensive posttranslational modification (glycosylation), and is incorporated into the bone matrix during normal bone formation (153, 154, 155). Although BSP accounts for 10–15% of the noncollagenous organic bone matrix, the protein, or fragments thereof, appear to be released during normal and abnormal osteoclastic bone resorption.
Interestingly, BSP is also synthesized and secreted by a number of other cells, including placental trophoblasts and malignant tumors such as breast, prostate, or thyroid cancers. The expression of BSP in these tumors has been proposed to play a role in tumor microcalcification, the homing of tumor cells to bone, and enhanced survival of tumor cells in the bone microenvironment (156). In vitro, however, some of the cancer cell-derived BSP seems to be degraded before secretion (157).
After developing a specific immunoassay for circulating immunoreactive BSP, Seibel et al. (139) were able to demonstrate that serum BSP levels correlate with markers of bone resorption in both healthy controls and patients with metabolic or malignant bone disease. In addition, iv bisphosphonate treatment resulted in a rapid reduction of serum BSP levels that paralleled and exceeded the changes observed with PYD and DPD. These results suggested that serum BSP levels were associated with bone resorption rather than bone formation.
Further studies indicated that serum BSP levels were often elevated in patients with tumors metastatic to bone. Interestingly, the highest levels seemed to occur in patients with bone metastases from cancers that are known to ectopically express significant amounts of BSP, such as breast, prostate, or thyroid cancers (158, 159, 160). In another study, serum BSP levels were closely related to serum PSA levels (161), although in the same report, serum BSP was also elevated in patients with colon and lung cancer.
Earlier studies by Castronovo’s group (158, 159, 160) found that in patients with metastatic breast cancer, the degree of BSP expression in the primary tumor (as assessed by immunohistochemistry), correlates with the propensity of the cancer to metastasize to bone (159). Consecutive clinical studies by Diel et al. (162) demonstrated that in patients with primary breast cancer, serum BSP concentrations measured at baseline (i.e. at the time of the operation, when no metastases were present) were highly predictive of future bone metastases.
Abnormally high serum BSP values are often found in patients with untreated MM, and measurement of the protein’s serum concentration seems to distinguish between patients with MM and benign osteoporosis (163) (Fig. 8). In general, serum BSP concentrations increase with disease progression, and patients with osteolytic lesion often have higher levels than individuals diagnosed with nonlytic bone disease. Furthermore, serum BSP seems to reflect the response to CT in patients with MM, because the treatment-induced changes in serum BSP values correlate with the changes in the monoclonal protein (163). In prostate cancer, expression of BSP may enable the identification of subgroups of patients that are at different risks of bone metastasis or recurrence (164). However, further studies are needed to precisely determine the role of BSP in metastatic bone disease and are so far hampered by the lack of a suitable, commercially available assay system.
|
As mentioned earlier, RANKL and OPG play essential roles in bone metabolism (Fig. 1
). OPG binds to RANKL and blocks its interaction with RANK, thus inhibiting osteoclast development. OPG mRNA has been found ubiquitously in most body tissues, prostate cancer, and predominantly osteoblastic cells (165).
In bone lesions of metastatic prostate cancer, OPG has been found to be up-regulated and was thought to be involved in the development of osteosclerotic/osteoblastic metastases (8). Furthermore, it has been demonstrated, that patients with advanced disease prostate cancer have significantly higher serum OPG levels than patients with lower disease stages, although there was no significant correlation between OPG and serum CTX-I levels in patients with advanced disease (166). In another study, significant differences in OPG serum levels were found between prostate cancer patients with bone metastases compared with healthy controls, patients with nonmetastasized cancer, and patients with benign hyperplasia. Even when compared with NTX-I serum levels, OPG was found to be of higher sensitivity in detecting bone metastases than NTX-I (167). Comparing OPG serum levels from healthy individuals to levels from patients with solid tumors, hematologic malignancies, or benign disorders, there were no significant differences in OPG levels of patients with solid tumors vs. healthy controls. Analyzed between the tumor groups, elevated serum OPG was only observed in patients with colorectal and pancreatic cancer. There was no OPG elevation in patients with bone metastases compared with individuals with soft organ or liver metastases. Among the group with metastasized tumors, highest levels were seen in patients with soft tissue and liver metastasis (165). Taken together, OPG serum levels seem to reflect the extent of metastatic disease in prostate cancer; whether these can be discriminative of skeletal involvement vs. soft tissue metastases still remains to be clarified. More recently, the promising results of a phase 1 study using a recombinant osteoprotegerin construct in patients with multiple myeloma or breast cancer-related bone metastases were published (168), and the future will show whether OPG has a therapeutic potential in this area.
RANKL
Commercial assays have recently become available for the measurement of RANKL levels in human serum but have not yet been widely applied to metastatic disease. However, in a recent study in newly diagnosed multiple myeloma patients, RANKL levels and more particularly, the RANKL/OPG ratio, were found to correlate closely with the extent of bone involvement and levels of surrogate markers of bone resorption. The RANKL/OPG ratio was also found to be an independent predictor of life expectancy (169). Thus, the measurement of RANKL levels in serum may have potential value in the assessment of metastatic diseases. Measurement of both RANKL and OPG may provide better discrimination of proresorptive stimuli in bone, because the relative expression of these cytokines likely determines osteoclast activity.
Summary
Biochemical markers of bone remodeling are potentially useful tools in the diagnosis and follow-up of patients with malignant bone disease. In patients with established bone metastases, most bone markers are abnormal, indicating that these parameters faithfully reflect changes in bone metabolism associated with the malignant process. However, the currently available evidence does not allow final conclusions in regard to the accuracy and clinical validity of any of these indices in the primary diagnosis of bone metastases. As regards the monitoring of anticancer therapy, markers of bone resorption such as the pyridinium cross-links ICTP, NTX, and CTX all react promptly and profoundly to bisphosphonate, hormonal, or chemical treatments. These indices therefore have the potential to be used in the monitoring of such interventions. Some studies indicate that levels of bone resorption markers before and their changes during and after bisphosphonate treatment seem to be predictive of clinical outcome. Again, most if not all of the available data are observational, and there have been no controlled studies on the usefulness of bone markers in the pre- and posttherapeutic monitoring of cancer patients.
Although it is unlikely that a single marker of bone remodeling has sufficient diagnostic or prognostic value in malignant bone disease, the combination of these markers with other laboratory tests (e.g. tumor markers) and imaging techniques is likely to improve the clinical assessment of patients with bone-seeking cancers.
|
Footnotes |
|---|
, macrophage inflammatory protein 1; MM, multiple myeloma; NTX-I, N-terminal cross-linked telopeptide of collagen type I; OC, osteocalcin; OHP, hydroxyproline; OPG, osteoprotegerin; PICP, C-terminal propeptide of procollagen type I; PINP, N-terminal propeptide of procollagen type I; PSA, prostate specific antigen; PYD, pyridinoline; RANKL, receptor activator of nuclear factor-B ligand; s, serum; sBALP, serum bone-specific ALP; TRAcP, tartrate-resistant acid phosphatase; u, urinary. Received May 23, 2003.
Accepted August 4, 2003.
|
References |
|---|
|
TopAbstractMechanisms of Metastatic Bone...Markers of Bone Remodeling...Future DevelopmentsReferences |
|---|
(4)ß(1)-integrin enhances production of osteoclast-stimulating activity. Blood 96:1953–1960
carboxyglutamic acid-containing protein in primary hyperparathyroidism and in malignant hypercalcemia. Comparison with bone histomorphometry. J Clin Invest 77:985–991
); MM III, patients with MM stage III. Group medians are depicted by a solid line. The dotted line represents the upper limit of the normal reference range. h-PYD, Urinary total PYD measured by HPLC; h-DPD, urinary total DPD measured by HPLC; i-DPD, urinary free DPD determined by immunoassay. All values are corrected for urinary creatinine and are expressed in micromoles of cross-link per mole creatinine. [Reproduced with permission from Pecherstorfer et al., Blood 90:3743–3750, 1997 (
, DPD; 
, PYD; •, CTX-I; 
, OHP; 
