Overview of bisphosphonates

Bisphosphonates (BPs) were first studied over 30 years ago as stable analogues of a naturally occurring inorganic pyrophosphate (PPi), which were shown to be able to inhibit the precipitation of calcium phosphate in vitro and biologic calcification in vivo. A key step occurred when it was shown that they also inhibited bone resorption induced by a wide variety of agents and had profound effects on calcium metabolism in vivo. The first clinical uses of BPs followed soon after in the late 1960s and included their use as agents for bone scanning, based on their ability to adsorb to bone mineral, for which they remain outstandingly useful. Concurrently they were shown to be strikingly effective in clinical disorders associated with increased bone resorption, initially in Paget's disease of bone, then in hypercalcemia of malignancy, myeloma, and bone metastases, and much later in osteoporosis. As a result, several BPs (e.g., etidronate, clodronate, pamidronate, alendronate, and tiludronate) now are licensed as drugs for various indications, and more should follow (risedronate, ibandronate, zoledronate, etc.). Because of the current emphasis on osteoporosis, their value in patients with Paget's disease and in those with malignancy runs the risk of being undervalued. The mode of action of BPs originally was ascribed to their ability to adsorb strongly to hydroxyapatite crystals, and to inhibit their growth and dissolution, but it has gradually become clear that this is insufficient to account for all their effects. The ability to make BPs with minimal inhibitory effects on bone mineralization but increased relative potency on bone resorption was achieved many years ago. Remarkable progress has been made in increasing their potency as inhibitors of bone resorption by factors of 100-10,000 fold. One key feature appears to be the insertion of a nitrogen atom at critical positions in the side chain. These may be in alkyl side chains, (e.g., butyl-NH2) as in alendronate, or in the highly potent -N substituted derivatives of aminohydroxypropylidene bisphosphonate (APD) (pamidronate) (e.g., dimethyl APD, or methylpentenyl APD [ibandronate]). Comparable high potencies are obtained with heterocyclic ring compounds containing nitrogen (e.g., with imidazole, pyridinyl [e.g., as in risedronate] or picolyl groups). These compounds are remarkably potent, and are able to suppress bone resorption in experimental animals at doses of < 1 μg per day. A simple hypothesis for the effects of BPs on bone resorption is that the compounds are selectively concentrated in bone, where they interfere with the action of osteoclasts by as yet poorly defined mechanisms. It is likely that BPs are internalized by osteoclasts and interfere with specific biochemical processes. Thus BPs interfere with osteoclast recruitment, differentiation, and action, and can induce apoptosis. The authors of the current study and others have found several differences between BPs in terms of their biochemical interactions within cells. Recent mechanistic studies show that BPs can be classified into groups with different modes of action. Those that most closely resemble PPi (e.g., clodronate) can be incorporated into toxic ATP analogues, whereas more potent nitrogen-containing BPs interfere with other reactions (e.g., in the mevalonate pathway) and may affect cellular activity such as apoptosis by interfering with protein prenylation, and therefore the intracellular trafficking of key regulatory proteins. Therefore there may be subtle differences between compounds in terms of their clinical effects. The choice of therapeutic regimen itself poses further interesting questions (e.g., the effect of intermittent dosing vs. continuous, intravenous vs. oral therapy, the optimal duration of therapy, and the role of individual variability in response). BPs represent an important class of drugs for the treatment of patients with bone diseases, and their full potential has yet to be realized. The ability of bisphosphonates to inhibit bone resorption is dependent on two features of the bisphosphonate molecule. The first feature is the high affinity for bone mineral,1-3 which allows the rapid and selective targeting of bisphosphonates to bone mineral surfaces in vivo.4 The affinity for hydroxyapatite is a property of the PCP motif because bisphosphonates (like pyrophosphate) can chelate calcium ions by bidentate coordination through the oxygen atoms of the phosphonate groups. The affinity of bisphosphonates for calcium is even greater if the R1 side chain is a hydroxyl (OH) group (as in etidronate)3 because this allows tridentate coordination to calcium ions. Modifications to the phosphonate groups that reduce affinity for calcium (e.g., methylation of one or both phosphonates to give phosphonophosphinates or bisphosphinates, respectively) or substitution of PCP bonds with PNP or PCCP bonds, reduce antiresorptive potency.5, 6 The second important feature of the bisphosphonate molecule is the structure of the side chains, R1 and R2. Because R1 is usually a hydroxyl group to maximize bone affinity, it is the structure of the R2 side chain that is the critical determinant of antiresorptive potency.6 Bisphosphonates that contain a primary amino group (such as pamidronate and alendronate) are up to 100-fold more potent than nonamino-containing bisphosphonates such as etidronate or clodronate.6-8 Methylation of the amino group or inclusion within a heterocyclic ring increases potency still further.9, 10 Hence, bisphosphonates that contain a secondary or tertiary amino group such as ibandronate, risedronate, and zoledronate are among the most potent antiresorptive bisphosphonates.11-13 There have been numerous studies that have demonstrated that minor changes to the structure or three-dimensional conformation of the R2 side chain can dramatically alter antiresorptive potency. For example, elongation of the aminoalkyl side chain by one methylene (CH2) group increases the potency of alendronate 10-fold compared with pamidronate.6, 8, 11 The position of the nitrogen moiety within a heterocyclic ring also can influence antiresorptive potency.6, 14 These observations all suggest that bisphosphonates must interact with one or more molecular targets that contain a binding site complementary in structure to the R2 side chain. Therefore, alterations to the structure of R2 affect the ability of bisphosphonates to bind to or inhibit the activity of such targets. It is likely that the phosphonate groups also play a role in binding to target molecules. In a growth-inhibition assay using Dictyostelium amebae (which accurately reflects the antiresorptive potencies of bisphosphonates but in which there is no bone mineral present, thus ruling out the influence of bone-binding),15, 16 phosphonophosphinates and bisphosphinates are far less potent than the corresponding bisphosphonate.17 Thus, the phosphonate groups not only direct bisphosphonates to bone mineral in vivo but also, in combination with the R2 side chain, interact with (as yet unidentified) molecular targets. The cells that are most likely to be exposed to the highest concentrations of bisphosphonates in vivo are osteoclasts, the highly specialized, multinucleate cells that are formed by the fusion of hematopoietic, unicellular precursors in the bone marrow.18, 19 Due to their high affinity for calcium, bisphosphonates are cleared rapidly from the circulation in vivo and bind to areas of exposed bone mineral (e.g., for example around resorbing osteoclasts).20, 21 Thus, concentrations of bisphosphonate in solution that were very low originally could give rise to very much higher local concentrations after adsorption then release from bone in the acidic environment beneath resorbing osteoclasts. Sato et al. have estimated that pharmacologic doses of alendronate that inhibit bone resorption in vivo could give rise to local concentrations as high as 1 mM in the resorption lacuna beneath resorbing osteoclasts.20 After release from the bone surface, bisphosphonates are most likely internalized into osteoclasts by fluid-phase pinocytosis or by phagocytosis of calcium complexes. Osteoclasts in vivo have been shown to internalize radiolabeled alendronate into numerous intracellular vacuoles, with evidence of the bisphosphonate also being present in other subcellular compartments, including the cytoplasm, mitochondria, and nuclei.20 The site of action of bisphosphonates almost certainly is intracellular, because osteoclasts that are prevented from resorbing bisphosphonate-coated bone by the addition of calcitonin are protected from the deleterious effects of these agents.22-24 Furthermore, microinjection of tiludronate directly into isolated osteoclasts causes disruption of the cytoskeletal arrangement required for bone resorption,25 whereas the potency of bisphosphonates also can be increased after enhanced intracellular delivery of bisphosphonates into phagocytic cells (such as macrophages and Dictyostelium amebae) using liposomes.26, 27 Once internalized, bisphosphonates can affect a multitude of biochemical processes, resulting in a loss of ability to resorb bone or even osteoclast cell death.28-30 It has been well documented that bisphosphonates cause disruption of the osteoclast cytoskeleton, including loss of the actin ring (required for polarization and the formation of a sealing zone at the bone surface)31 and loss of the ruffled border,25, 31-33 a region of highly convoluted membrane at the apical surface adjacent to the bone and across which is secreted hydrolytic enzymes and protons. Two reports recently have demonstrated that clodronate, pamidronate, and risedronate also can cause osteoclast apoptosis in vitro and in vivo,29, 30 a form of cell death characterized by distinct changes in cell and nuclear morphology (including cell contraction and nuclear fragmentation). In retrospect, the observations many years ago by Schenk et al.,34 Rowe et al.,24, 35, 36 and others,22 of degenerative changes in osteoclast morphology, such as nuclear pyknosis and fragmentation, also are indicative of apoptosis.37 In vitro, bisphosphonates cause apoptosis of other cell types, including J774 macrophages38 and human myeloma cells.39 The structure-activity relationships of bisphosphonates for causing J774 macrophage apoptosis in vitro closely match the structure-activity relationships for inhibiting bone resorption, suggesting that bisphosphonates cause apoptosis in macrophages and osteoclasts by similar molecular mechanisms.40 The molecular targets of bisphosphonates, and hence the exact mechanism by which they inhibit bone resorption, have not been identified. However, given the similarity of bisphosphonates to pyrophosphate and the apparent requirement of the PCP moiety for antiresorptive function, it is likely that bisphosphonates inhibit intracellular enzymes that bind naturally occurring phosphate-containing or pyrophosphate-containing substrates. Several bisphosphonates have been found to inhibit protein synthesis in isolated osteoclasts and calvarial cells28, 41 and to inhibit glycolysis and reduce lactate production in isolated cells and calvaria.42-44 More recently, Schmidt et al. have shown that alendronate and other bisphosphonates can inhibit several protein tyrosine phosphatases (PTP) (such as PTPσ and PTPε), without affecting serine or threonine phosphatases.45 Tiludronate also has been shown to increase the level of protein phosphorylation in osteoclasts by inhibiting protein phosphatase activity.46 Hence, inhibition of protein dephosphorylation of critical substrates required for bone resorption in osteoclasts, such as c-src,47 may be a mechanism by which bisphosphonates inhibit the ability of osteoclasts to resorb bone. Proton-pumping ATPases also are vital for acidification of the resorption cavity beneath active osteoclasts.48, 49 Tiludronate recently has been shown to have a direct and potent inhibitory effect on the proton-pumping activity of inside-out vesicles derived from osteoclast plasma membranes.50 By contrast, others have shown that etidronate, pamidronate, and clodronate could inhibit vacuolar acidification in osteoclasts by an indirect effect, possibly by inhibiting cell metabolism or by preventing the insertion of ATPases into the plasma membrane.28, 51 The authors' own work has suggested that bisphosphonates may be divided into two classes that inhibit osteoclastic resorption by different mechanisms. Using amebae of the slime mould Dictyostelium discoideum (the growth of which is inhibited by bisphosphonates)15-17 they and others found that clodronate (dichloromethylenebisphosphonate) and other bisphosphonates that closely resemble pyrophosphate (such as methylenebisphosphonate and difluoromethylenebisphosphonate) can be metabolized by Dictyostelium amebae to nonhydrolysable analogues of ATP that contain a β,γ-methylene group (AppCp nucleotides).52-54 More potent bisphosphonates with larger R2 side chains such as pamidronate or alendronate are not metabolized.53 The bisphosphonates such as clodronate that are metabolized most likely replace pyrophosphate in a back reaction catalyzed by Class 2 aminoacyl-tRNA synthetase enzymes, thus inhibiting the forward reaction (the synthesis of aminoacyl-tRNA).53 Because these enzymes are ubiquitous, it was not surprising to find that the equivalent mammalian enzymes in vitro also appear to incorporate into ATP analogues the same bisphosphonates that are metabolized by Dictyostelium.55 The authors recently confirmed that mammalian cells (murine J774 macrophages and human MG-63 osteosarcoma cells) also are capable of metabolising clodronate,56 confirming the observation many years ago by Felix et al. that clodronate could be "modified" after intracellular uptake by calvarial cells.57 Furthermore, intracellular accumulation of the metabolite of clodronate (AppCCl2p) causes cell death and reduces osteoclast number in long term marrow cultures (Rogers MJ and Frith JC, unpublished data). Thus, it is likely that clodronate inhibits bone resorption owing to intracellular accumulation of a cytotoxic metabolite. Fortuitously, the high affinity of the bisphosphonate for bone mineral probably ensures that cytotoxic concentrations of the metabolite only accumulate within osteoclasts. Because the amino-containing bisphosphonates are not metabolized and do not inhibit the activity of aminoacyl-tRNA synthetases,53, 55 they presumably act by a different mechanism from that of clodronate. Following reports that several of these bisphosphonates can inhibit squalene synthase or other enzymes of the biosynthetic mevalonate pathway in J774 macrophages,58, 59 the authors recently began to accumulate evidence that the aminobisphosphonates affect osteoclasts by inhibition of this pathway. Although required for the synthesis of cholesterol, several intermediates of the mevalonate pathway (farnesylpyrophosphate and geranylgeranylpyrophosphate) also are required for the posttranslational modification (prenylation) of GTP-binding proteins such as the ras, rac, and rho family of proteins that regulate cell proliferation, survival, membrane trafficking, and cytoskeletal organisation.60 The authors have found that aminobisphosphonates such as alendronate and ibandronate inhibit the modification of these proteins in J774 macrophages with farnesyl or geranylgeranyl groups,61 an effect that would prevent membrane association of these proteins and hence affect their function.60 Furthermore, other inhibitors of the mevalonate pathway such as mevastatin also cause macrophage and osteoclast apoptosis in vitro.62, 63 Although yet to be confirmed, it is likely that enzymes of the mevalonate pathway that utilize farnesylpyrophosphate or geranylgeranylpyrophosphate are the molecular targets of the aminobisphosphonates. Once again, it is fortuitous that the selective effect of bisphosphonates on osteoclasts is a consequence of their high affinity for bone mineral. The precise concentration of bisphosphonate that can be achieved in areas of bone other than beneath resorbing osteoclasts (e.g., in the bone marrow) is unknown. Hence, it is possible that bisphosphonates released from the bone surface during resorption could have effects on other bone cells as well as osteoclasts. Because bone resorption in vivo is dependent on the generation of new osteoclasts by the fusion of mononucleate, hematopoietic precursors, bisphosphonates could inhibit bone resorption via effects on osteoclast precursors, in addition to effects on mature osteoclasts. Hughes et al. found that bisphosphonates dose-dependently inhibited the formation of osteoclast-like cells in vitro in human long term bone marrow cultures.64 Furthermore, the order of potency of the five bisphosphonates studied for inhibiting osteoclast formation (risedronate > pamidronate > neridronate > clodronate > etidronate) matched the order of potency of the bisphosphonates for inhibiting bone resorption in vivo. Boonekamp et al. and Löwik et al. also concluded that low concentrations of aminobisphosphonates such as pamidronate could inhibit bone resorption in fetal mouse metacarpals in vitro by inhibiting the transformation, maturation or fusion of osteoclast precursors, possibly by preventing the recognition of some bone matrix factor essential for stimulating mature osteoclast formation.65, 66 Higher concentrations of pamidronate, as well as the less potent bisphosphonates clodronate and etidronate, appeared to inhibit resorption by affecting mature osteoclasts.65 Therefore it is possible that in vivo bisphosphonates affect both osteoclast precursors and mature osteoclasts, depending on the particular bisphosphonate and the concentrations to which the cells are exposed. There have been several reports that bisphosphonates inhibit the proliferation of connective tissue cells (calvarial cells, fibroblasts, and chondrocytes)43, 67 and human osteoblast-like ROS 17/2.8 cells in vitro.68 In the latter case, the order of potency for inhibition of cell proliferation (risedronate > alendronate > pamidronate > clodronate > etidronate) matched the order of antiresorptive potency. Because osteoblasts regulate the activity of osteoclasts, it is possible that bisphosphonates could inhibit osteoclasts or osteoclast formation via effects on osteoblasts, either by preventing the secretion of osteoclast-stimulating factors or by stimulating the production of osteoclast-inhibitory factors.69, 70 Sahni et al. found that treatment of rat osteoblast-like CRP 10/30 cells (which are potent stimulators of osteoclast activity) with ibandronate or clodronate resulted in inhibition of resorption when these cells then were cocultured for 24 hours with osteoclasts.69 The inhibitory effect could be achieved after even a short (5 minute) exposure of the CRP 10/30 cells to bisphosphonates, and also occurred when conditioned medium from the bisphosphonate-treated CRP 10/30 cells was added to osteoclast cultures. This effect appeared to be mediated by a low molecular weight (1-10,000 dalton), osteoclast-inhibitory factor present in the conditioned medium of the bisphosphonate-treated CRP 10/30 cells,71 the identity of which remains to be determined. Macrophages in vitro, like osteoclasts, appear to be particularly susceptible to the deleterious effects of bisphosphonates, probably owing to the highly endocytic nature of the cells and hence the ability to internalize relatively large quantities of bisphosphonate.72 Hence, bisphosphonates inhibit macrophage proliferation, viability, cell migration or function, and can cause apoptosis.26, 38, 73-77 It has been suggested that inhibition of bone resorption in vivo could therefore be accounted for partially by an indirect effect via these cells in the bone marrow, which are a source of resorption-stimulating cytokines.72 However, the amino-containing bisphosphonates such as pamidronate and alendronate also can cause a transient acute phase response, possibly owing to the release of proinflammatory cytokines such as interleukin 6 (IL-6) and tumor necrosis factor-α.78, 79 Pamidronate enhances the production of IL-6 by LPS-stimulated macrophages in vitro.80 However, the antiinflammatory effects of bisphosphonates that have been described in models of arthritis in animals81-83 also could be accounted for by bisphosphonates affecting the monocyte-macrophage system. In particular, liposome-encapsulated clodronate, which is toxic to phagocytic macrophages but not other cells26, 84 and can inhibit the production of cytokines and nitric oxide from activated macrophages in vitro,80, 85 may be of future use in the depletion of macrophages in inflammatory diseases such as rheumatoid arthritis.86-88 Bisphosphonates have become the treatment of choice for hypercalcemia of malignancy and osteolytic bone disease associated with tumor metastases.89 In animal models of metastatic disease, bisphosphonates can slow the development of bone metastases and reduce tumor burden, apparently by an indirect effect through inhibition of osteoclast activity,90 thus preventing the release of tumor-stimulating growth factors. Although there is no evidence as yet that bisphosphonates can affect directly tumor cells in vivo, the authors have found that pamidronate and incadronate (YM175) can cause cell cycle arrest and apoptosis of human myeloma cell lines in vitro, at concentrations (≥ 100 that could be achieved in the of resorbing Because myeloma cells are to be closely associated with of bone resorption, this the that bisphosphonates could have direct effects on myeloma cells in vivo. et al. recently have that bisphosphonates to bone mineral also inhibit the and of cells to bone matrix and therefore could prevent of tumor cells to studies are required to such effects in vivo, because this would the use of bisphosphonates to prevent the development of metastatic bone