Elsevier

Bone

Volume 38, Issue 5, May 2006, Pages 628-636
Bone

Relative binding affinities of bisphosphonates for human bone and relationship to antiresorptive efficacy

https://doi.org/10.1016/j.bone.2005.07.023Get rights and content

Abstract

Potent bisphosphonates (BPs) preferentially bind bone at sites of active osteoclastic bone resorption, where they are taken up by the osteoclast and inhibit resorption. We tested the hypothesis that BP affinity to human bone affects antiresorptive potency. [14C]-Alendronate binding to human bone was saturable and reversible with an apparent Kd of 72 μM by Scatchard analysis. In competition binding assays, unlabeled alendronate (Ki: 61 μM) was slightly more potent than pyrophosphate (Ki = 156 μM) in blocking [14C]-alendronate binding. Likewise, most tested BPs, including etidronate (Ki: 91 μM), ibandronate (116 μM), pamidronate (83 μM), risedronate (85 μM) and zoledronate (81 μM), showed comparable affinities. Interestingly, tiludronate (173 μM; P < 0.05 vs. all other BPs) and especially clodronate (806 μM; P > 0.0001 vs. all other BPs) displayed significantly weaker affinity for bone. The weak affinity of clodronate translated into a requirement for 10-fold higher dosing in in vitro bone resorption assays when bone was pretreated with BP and subsequently washed prior to adding osteoclasts. In stark contrast, neither alendronate nor risedronate lost any efficacy after washing the bone surface. These findings suggest that most clinically tested BPs may have similar affinities for human bone. For those with reduced affinity, this may translate into lower potency that necessitates higher dosing.

Introduction

The bisphosphonates (BPs) are the most effective antiresorptive agents used for the clinical treatment of postmenopausal and other forms of osteoporosis. Several factors contribute to their potency and efficacy as inhibitors of bone resorption, the intracellular target of BPs being one major component. To this extent, the relative IC50s of the nitrogen-containing BPs (N-BPs) such as alendronate (ALN) and risedronate (RIS) for inhibiting their intracellular target, farnesyl diphosphate synthase, fall into the low to mid nanomolar range [4], [7], [40]. For these and other N-BPs, a correlation exists between in vitro potency to inhibit this enzyme and the lowest effective dose in the rat [7]. Clodronate (CL2) falls into a different category regarding intracellular mechanism [13]. For this BP, and probably etidronate (EHDP) and tiludronate (TIL), the BP itself is incorporated into the β and γ positions of ATP through an interaction with aminoacyl-tRNA synthetases. The resulting analog is toxic and causes osteoclast apoptosis. The different mechanisms make difficult the process of comparing intracellular potency between different classes of BP. One can nonetheless make comparisons based upon the relative potency to induce cellular responses such as suppression of bone resorption and induction of osteoclast apoptosis in vitro. In most regards, CL2 is considered to be about one-third as potent as ALN and RIS in these types of in vitro assays.

Osteoclast apoptosis is one cellular mechanism by which osteoclastic bone resorption can be inhibited by a BP. Whereas the non-nitrogen-containing BPs such as CL2 and EHDP employ apoptosis as the dominant mechanism for antiresorptive efficacy, the N-BPs employ this as a secondary mechanism [1], [14]. Nonetheless, since all antiresorptive BPs induce osteoclast apoptosis, this endpoint can serve as an effective in vitro measure of relative potency. In several studies of BP-induced cultured osteoclast apoptosis, ALN, RIS and CL2 all induced cell death within a relatively narrow concentration range of three-fold [3], [8], [14], [17], [19], [30]. Similarly, we previously found that CL2 should be dosed at 1.7-fold higher than ALN or RIS in head-to-head in vitro bone resorption assays in order to achieve a similar antiresorptive effect [14]. In these assays, the BPs were allowed to bathe the bone during the entire assay, and no attempt was made to wash away the unattached BP prior to addition of osteoclast cultures. Based on these in vitro observations, one might predict that in vivo or clinical potency for these three BPs could fall into the same type of narrow range.

Contrary to in vitro predictions, CL2 is substantially less potent than ALN or RIS in vivo than expected. Although ALN and CL2 have never been compared head-to-head in a clinical study, they have been compared in rats in vivo with CL2 dosing set at surprisingly higher values of up to 100-fold [9], [36]. A head to head clinical comparison showed that ALN and RIS increase lumbar spine bone mineral density by 3.7% and 2.6%, respectively, after 1 year [32]. The ability of these BPs to elicit these changes in BMD at relatively low daily (or weekly) doses is striking when compared to the substantially weaker responses seen in the clinic with CL2, which shows no increase in lumbar spine BMD after 3 years of dosing at 65 mg per day [37]. Increases of 3.8% can be seen after 3 years treatment with the substantially higher oral daily dose of 800 mg. This suggests a failure of standard in vitro testing to predict either in vivo animal or clinical potency of CL2 vs. ALN or RIS.

In the present study, we sought to establish how differences in relative binding affinities of clinically tested BPs for human bone could serve as independent contributors to antiresorptive efficacy. Some of the earliest studies on the subject of BP affinity examined inhibition of hydroxyapatite (HAP) mineral formation [10], [12]. Although the earlier studies did not directly compare BPs in modern clinical use, they did establish, for instance, that CL2 has substantially lower affinity for calcium phosphate [35]. Recent studies have employed similar approaches to compare relative binding affinities of BPs such as ALN, EHDP, RIS and zoledronate (ZOL) for inhibition of both hydroxyapatite and octa calcium phosphate crystal formation (or dissolution) as a surrogate for binding to human bone [26]. In an effort to make these studies more relevant to bone, which is comprised of ∼ 35% protein, inhibition of the formation of carbonated apatite crystals (CAP) has also been tested [15]. The data comparing ALN and RIS inhibition of HAP, CAP or octa calcium phosphate suggest a 1.4- to 2.6-fold greater affinity for ALN, depending on the system. The affinity constants for ALN using these chemical in vitro methods fall into the mid nanomolar to low micromolar range. Meanwhile, ALN was found to bind human bone with an apparent Kd of 1 mM [34], and competition binding studies suggest an IC50 for binding to mouse bone in the 70–100 μM range [41]. To date, only ALN and EHDP have been compared for their affinities to human bone, and the data from competition binding analyses suggest that the binding is identical [34]. Likewise, ALN and RIS show similar binding to mouse metatarsals in competition binding analyses [41]. These observations differ from those describing relative effects of ALN and EHDP on in vitro crystal growth, whereby an ALN potency advantage has been reported, depending on the type of crystal [26]. Thus, substantial differences may exist in how BPs affect crystal growth vs. how they bind to human or mouse bone. This suggests the need for a more comprehensive comparison of the various BPs.

In the present study, we measured relative binding affinities for clinically tested BPs for human bone. In addition to measuring the binding affinity and on and off rates for ALN binding, a competition binding approach was used to compare between the clinically relevant BPs. We find that CL2, with a sub mM affinity, bound ∼10-fold more weakly than most tested BPs. All BPs with an OH-group attached to the central carbon bound within a fairly narrow affinity range. This suggests that structural differences in BPs that relate to intracellular potency have little, if any, effect on binding to human bone under these conditions.

Section snippets

Materials

Human bone particles were collected by sifting through a 180 μm mesh screen and then capturing in a 150 μm mesh screen, as previously described [34]. [14C]-ALN and all non-radioactive BPs were from Merck Research Laboratories. Tris–formate, ATP, AMP, a second sample of CL2, pyrophosphate, beta-glycerophosphate and sodium phosphate were all from Sigma-Aldrich.

Binding of [14C]-ALN to human bone particles

Human bone particles were prewashed twice with binding buffer (0.2 M Tris–formate, pH 7.2) at 22°C. The choice of this buffer and

ALN binding to human bone particles

In an effort to elucidate possible differences between BPs that might influence in vivo or clinical potency, we examined the relative binding affinities of BPs towards human bone. In vivo studies of ALN localization on rat bone showed that it preferentially binds to sites of osteoclastic resorption with eight-fold higher BP found beneath the osteoclast than seen on osteoblast-covered surfaces [23]. Resorption sites represent the portions of bone that are the most exposed since osteoclasts are

Discussion

In this study, we examined the equilibrium bone binding affinities of clinically tested BPs as a putative factor influencing antiresorptive potency and efficacy. In contrast to several studies examining effects of BPs on HAP crystal growth or dissolution in vitro published decades ago [6], [10], [12], [35], we assessed binding to human bone with an alternative method. With this approach, we found that the binding affinity of the clinically tested OH-BPs fell into a narrow range. Strikingly, CL2

References (42)

  • A. Alakangas et al.

    Alendronate disturbs vesicular trafficking in osteoclasts

    Calcif. Tissue Int.

    (2002)
  • J. Benedict

    The physical chemistry of the diphosphonates—Its relationship to their medical activity

  • H.L. Benford et al.

    Farnesol and geranylgeraniol prevent activation of caspases by aminobisphosphonates: biochemical evidence for two distinct pharmacological classes of bisphosphonate drugs

    Mol. Pharmacol.

    (1999)
  • A.L. Boskey et al.

    Effect of diphosphonates on hydroxyapatite formation induced by calcium

    Calcif. Tissue Int.

    (1979)
  • J.E. Dunford et al.

    Structure–activity relationships for inhibition of farnesyl diphosphate synthase in vitro and inhibition of bone resorption in vivo by nitrogen-containing bisphosphonates

    J. Pharmacol. Exp. Ther.

    (2001)
  • H. Fleisch

    Bisphosphonates: mechanisms of action

    Endocr. Rev.

    (1998)
  • H. Fleisch et al.

    Diphosphonates inhibit hydroxyapatite dissolution in vitro and bone

    Science

    (1969)
  • I. Fogelman et al.

    Risedronate reverses bone loss in postmenopausal women with low bone mass: results from a multinational, double-blind, placebo-controlled trial

    J. Clin. Endocrinol. Metab.

    (2000)
  • M.D. Francis et al.

    Diphosphonates inhibit formation of calcium phosphate crystals in vitro and

    Science

    (1969)
  • J.C. Frith et al.

    Clodronate and liposome-encapsulated clodronate are metabolized to a toxic ATP analog, adenosine 5′-(beta, gamma-dichloromethylene) triphosphate, by mammalian cells in vitro

    J. Bone Miner. Res.

    (1997)
  • Z.J. Henneman et al.

    Bisphosphonate bone affinity differences predicted by in vitro carbonated apatite model

    J. Bone Miner. Res.

    (2003)
  • Cited by (123)

    • Drugs eluting dental implants

      2020, Dental Implants: Materials, Coatings, Surface Modifications and Interfaces with Oral Tissues
    • Predicting binding affinities of nitrogen-containing bisphosphonates on hydroxyapatite surface by molecular dynamics

      2019, Chemical Physics Letters
      Citation Excerpt :

      Mukherjee and colleagues found the same rank order as Jahnke did [14,15]. Moreover, Leu and colleagues suggested that clinical BPs have similar affinities to human bone [16]. The binding mechanism of NBPs remains an open question.

    View all citing articles on Scopus
    1

    Current address: Department of Biochemistry and Biophysics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA.

    View full text