Visualizing mineral binding and uptake of bisphosphonate by osteoclasts and non-resorbing cells

Abstract

Bisphosphonates (BPs) target bone due to their high affinity for calcium ions. During osteoclastic resorption, these drugs are released from the acidified bone surface and taken up by osteoclasts, where they act by inhibiting the prenylation of small GTPases essential for osteoclast function. However, it remains unclear exactly how osteoclasts internalise BPs from bone and whether other cells in the bone microenvironment can also take up BPs from the bone surface. We have investigated this using a novel fluorescently-labelled alendronate analogue (FL-ALN), and by examining changes in protein prenylation following treatment of cells with risedronate (RIS). Confocal microscopic analysis showed that FL-ALN was efficiently internalised from solution or from the surface of dentine by resorbing osteoclasts into intracellular vesicles. Accordingly, unprenylated Rap1A accumulated to the same extent whether osteoclasts were cultured on RIS-coated dentine or with RIS in solution. By contrast, J774 macrophages internalised FL-ALN and RIS from solution, but took up comparatively little from dentine, due to their inability to resorb the mineral. Calvarial osteoblasts and MCF-7 tumour cells internalised even less FL-ALN and RIS, both from solution and from the surface of dentine. Accordingly, the viability of J774 and MCF-7 cells was drastically reduced when cultured with RIS in solution, but not when cultured on dentine pre-coated with RIS. However, when J774 macrophages were co-cultured with rabbit osteoclasts, J774 cells that were adjacent to resorbing osteoclasts frequently internalised more FL-ALN than J774 cells more distant from osteoclasts. This was possibly a result of increased availability of BP to these J774 cells due to transcytosis through osteoclasts, since FL-ALN partially co-localised with trancytosed, resorbed matrix protein within osteoclasts. In addition, J774 cells occupying resorption pits internalised more FL-ALN than those on unresorbed surfaces. These data demonstrate that osteoclasts are able to take up large amounts of BP, due to their ability to release the BP from the dentine surface during resorption. By contrast, non-resorbing cells take up only small amounts of BP that becomes available due to natural desorption from the dentine surface. However, BP uptake by non-resorbing cells can be increased when cultured in the presence of resorbing osteoclasts.

Introduction

Bisphosphonates (BPs) are non-hydrolysable analogues of pyrophosphate that potently inhibit bone resorption, a property that has led to their widespread application in the treatment of bone diseases characterised by excessive resorption, including post-menopausal osteoporosis, Paget's disease, and tumour-induced osteolysis [1]. This class of drugs have a high affinity for Ca²⁺ ions, due to the presence of the two phosphonate groups, and as a result BPs rapidly localise to bone mineral, particularly at sites of bone resorption [2], [3], [4], following administration in vivo. Although small amounts of BP naturally desorb from the bone surface [5], their release is dramatically increased during bone resorption. The acidic pH of the resorption lacuna (achieved by the transport of H⁺ and Cl⁻ ions across the osteoclast ruffled border membrane [6]) causes protonation of the phosphonate groups of BPs, which decreases their affinity for calcium ions and hence facilitates their release from bone surfaces into solution [7]. It is thought that some of this released BP is ‘recycled’ by reattaching to mineral on the surface of bone [5], while the remainder of this free BP is available for internalisation by osteoclasts into intracellular vesicles [8]. Acidification of these vesicles most likely causes protonation of the BPs, thereby reducing their charge and increasing their membrane permeability, enabling them to enter the cytosol [8]. The potent nitrogen-containing BPs (NBPs) then act by inhibiting farnesyl diphosphate (FPP) synthase, an enzyme in the mevalonate pathway that is required for the synthesis of the isoprenoid lipids, FPP and geranylgeranyl diphosphate (GGPP) [9]. These lipid groups are substrates for the post-translational prenylation of small GTPases of the Ras, Rho and Rab families that play a crucial role in osteoclast function. Inhibition of FPP synthase by NBPs thus depletes cells of FPP and GGPP, resulting in the accumulation of unprenylated forms of small GTPases and disruption of osteoclast function [10].

It is well-established that BPs are taken up by, and inhibit protein prenylation in, all cell types in vitro, and often adversely affect the proliferation and viability of these cells [11], [12], [13], [14]. The susceptibility of cells to the effects of BPs in vitro appears to be largely dependent on their endocytic activity, which determines how much BP these cells will be able to internalise [8], [13]. However, because of their exquisite bone-targeting property, coupled to the ability of osteoclasts to release BPs from the bone surface, it is thought that only these bone-resorbing cells are exposed to relatively high concentrations of BPs in vivo[15], and therefore in vitro susceptibility to BPs does not predict whether a particular cell type will be affected by BPs in vivo[16]. In support of this, recent studies have shown that NBPs have no detectable effect on protein prenylation in non-osteoclast cells in bone when administered in vivo at doses that cause a robust inhibition of protein prenylation in osteoclasts [17], [18]. Despite this, there is some evidence that non-osteoclastic cells in the bone microenvironment, including those on the bone surface such as osteoblasts, or cells in the marrow space, such as cancer cells that have metastasized to bone, may be directly affected by BPs in vivo[13], [19], [20], [21]. Since BPs can directly affect these and other cell types in vitro, it is plausible that such cells will also be affected in the in vivo setting, provided that they are able to internalise sufficient quantities of BP. Although cells such as osteoblasts are likely to be exposed only to the soluble pool of BP that naturally desorbs from the bone surface in vivo, it is possible that resorbing osteoclasts could increase the amount of BP in solution available to non-resorbing cells in their vicinity. Moreover, during bone resorption, degraded matrix proteins are trafficked through the resorbing osteoclast by transcytosis and released at a membrane domain on the basolateral surface known as the functional secretory domain [22], [23]. BPs may also be trafficked through this pathway following their internalisation, thereby further increasing the soluble pool of BP.

We have synthesized a fluorescently-labelled analogue of alendronate to study the mechanism of cellular uptake of BPs, using J774 macrophages as a model, and demonstrated that uptake requires fluid-phase endocytosis and is enhanced by Ca²⁺ ions [8].We have now used a similar fluorescent alendronate analogue (FL-ALN) to visualize both the binding of BP to mineralized surfaces and the uptake of BP from these mineralized surfaces by osteoclasts and other non-resorbing cells that may be present in the bone microenvironment. In addition, we have studied the uptake of unlabelled risedronate (RIS) by these cells, using inhibition of Rap1A prenylation as a surrogate marker for RIS uptake. Finally, we have used FL-ALN to study whether resorbing osteoclasts can increase the availability of BPs to other cells in the local microenvironment, for example by transcytosis and release at the functional secretory domain.