Binding Kinetics of a Fluorescently Labeled Bisphosphonate as a Tool for Dynamic Monitoring of Bone Mineral Deposition In Vivo


  • Robert J Tower,

    1. Section Biomedical Imaging, Department of Diagnostic Radiology, University Hospital Schleswig-Holstein, Campus Kiel, Germany
    Search for more papers by this author
  • Graeme M Campbell,

    1. Section Biomedical Imaging, Department of Diagnostic Radiology, University Hospital Schleswig-Holstein, Campus Kiel, Germany
    Search for more papers by this author
  • Marc Müller,

    1. Section Biomedical Imaging, Department of Diagnostic Radiology, University Hospital Schleswig-Holstein, Campus Kiel, Germany
    Search for more papers by this author
  • Olga Will,

    1. Section Biomedical Imaging, Department of Diagnostic Radiology, University Hospital Schleswig-Holstein, Campus Kiel, Germany
    Search for more papers by this author
  • Claus C Glüer,

    1. Section Biomedical Imaging, Department of Diagnostic Radiology, University Hospital Schleswig-Holstein, Campus Kiel, Germany
    Search for more papers by this author
  • Sanjay Tiwari

    Corresponding author
    1. Section Biomedical Imaging, Department of Diagnostic Radiology, University Hospital Schleswig-Holstein, Campus Kiel, Germany
    • Address correspondence to: Sanjay Tiwari, PhD, Section Biomedical Imaging, Department of Diagnostic Radiology, University Hospital Schleswig-Holstein, Campus Kiel, Am Botanischen Garten 14, 24118 Kiel, Germany. E-mail:

    Search for more papers by this author


Bone mineral deposition during the modeling of new bone and remodeling of old bone can be perturbed by several pathological conditions, including osteoporosis and skeletal metastases. A site-specific marker depicting the dynamics of bone mineral deposition would provide insight into skeletal disease location and severity, and prove useful in evaluating the efficacy of pharmacological interventions. Fluorescent labels may combine advantages of both radioisotope imaging and detailed microscopic analyses. The purpose of this study was to determine if the fluorescent bisphosphonate OsteoSense could detect localized changes in bone mineral deposition in established mouse models of accelerated bone loss (ovariectomy) (OVX) and anabolic bone gain resulting from parathyroid hormone (PTH) treatment. We hypothesized that the early rate of binding, as well as the total amount of bisphosphonate, which binds over long periods of time, could be useful in evaluating changes in bone metabolism. Evaluation of the kinetic uptake of bisphosphonates revealed a significant reduction in both the rate constant and plateau binding after OVX, whereas treatment with PTH resulted in a 36-fold increase in the bisphosphonate binding rate constant compared with untreated OVX controls. Localization of bisphosphonate binding revealed initial binding at sites of ossification adjacent to the growth plate and, to a lesser extent, along more distal trabecular and cortical elements. Micro-computed tomography (CT) was used to confirm that initial bisphosphonate binding is localized to sites of low tissue mineral density, associated with new bone mineral deposition. Our results suggest monitoring binding kinetics based on fluorescently labeled bisphosphonates represents a highly sensitive, site-specific method for monitoring changes in bone mineral deposition with the potential for translation into human applications in osteoporosis and bone metastatic processes and their treatment. © 2014 American Society for Bone and Mineral Research.


Bone is in a continual state of remodeling, which facilitates mineral transport and provides structural integrity. This tightly regulated process is mediated through bone-resorbing osteoclasts and bone-forming osteoblasts. Maintaining the balance of these two dynamic processes is essential to ensure skeletal homeostasis. Many diseases upset this balance and result in either general or localized bone loss or gain. Therefore, the measurement of the rates of bone formation and resorption can be indicative of disease severity. Bone formation and resorption can be measured using immunoassays that measure the serum concentration or urinary excretion of bone turnover biomarkers.[1] However, these markers provide an assessment only of the overall skeletal dynamics and fail to localize aberrant cellular function to a specific location. Current standard radiological modalities, such as radiographs, bone densitometry, and computed tomography (CT), provide a static assessment of bone mineral and structure[2] but do not describe the cellular dynamics of osteoblasts and osteoclasts.[3] Radioisotope labeling provides insights into bone turnover using scintigraphy (bone scan), or tomographic methods like single photon emission computed tomography (SPECT) or positron emission tomography (PET).[4, 5] Optical molecular imaging, where photons are detected from enzymatic reactions (bioluminescence) or fluorescent proteins or dyes (fluorescence), offers the possibility to assess the complex and highly regulated site-specific processes associated with bone remodeling[6] longitudinally, thus circumventing the limitations associated with static assessment of bone mineral and structure.

Because of their high affinity for bone and their ability to inactivate osteoclast activity, bisphosphonates have become a widely utilized treatment option for diseases with high bone turnover, such as osteoporosis, Paget's disease, and cancer-associated bone diseases.[7] Nitrogen-containing bisphosphonates function by inhibiting the intracellular enzyme farnesyl pyrophosphate synthase, depleting isoprenoid lipids used in the prenylation of protein.[8, 9] This results in the accumulation of unprenylated proteins with aberrant functions within the cell.[10] With the conjugation of infrared and near-infrared fluorescent dyes, bisphosphonates are now being assessed for their utility as a tool for monitoring bone dynamics in preclinical models. Recent works have demonstrated the complexity of bisphosphonate binding, showing not only binding to newly mineralized bone surfaces but also uptake by osteoclasts, bone marrow monocytes, and osteocyte lacunae.[11] It has also been shown that although the potency of inhibition of farnesyl pyrophosphate synthase strongly correlates with antiresorptive potency in vivo, bisphosphonate affinity also plays a critical role in skeletal uptake, distribution, and retention.[12]

Along with determining their skeletal distribution, researchers have attempted to use fluorescently conjugated bisphosphonates as a tool for monitoring changes in skeletal metabolism. With the use of molecular imaging, fluorescent bisphosphonates are now being assayed for their ability to label skeletal regions associated with increased bone formation, such as bone fracture healing and skeletal metastases.[13, 14] These studies demonstrate bisphosphonate binding to bone occurs not only to regions of robust osteoblast activity but also to quiescent bone surfaces. Longitudinal studies have also used fluorescent bisphosphonates as a tool to monitor changes in bone metabolism resulting from mechanical loading using fluorescent molecular tomography (FMT).[15] FMT sequentially obtains excitation and fluorescent emission measurements to allow the quantitative, three-dimensional determinations of fluorescence probes in vivo.[16-18] Lambers and colleagues recently showed acceptable reproducibility of FMT imaging and reasonable correlations between loss of fluorescent intensity of OsteoSense and the bone resorption rate determined from micro-computed tomography (micro-CT) (R2 = 0.81, p < 0.01). However, no significant difference in fluorescent intensity loss was observed between treatment groups. Similarly, although significant increases in OsteoSense binding were observed after loading compared with control mice (p < 0.05), large intragroup variation resulted in a lack of significant correlation with dynamic bone formation determined by micro-CT.

These previous studies[14, 15] focused on bisphosphonate localization 24 hours after injection and assayed for binding capacity assuming a correlation to bone formation. Their data document both the strengths and limitations of fluorescent measurements based on assessments of binding plateaus reached 24 hours after injection. Although they were able to detect changes in bisphosphonate uptake, these studies were not able to account for the dynamics of bisphosphonate binding at earlier time points affected by the affinity of the bisphosphonate used in relation to changes in the pattern of mineral deposition and resorption. With the injection of subsaturation concentrations of bisphosphonates in which not all potential binding sites will be occupied, the analysis of bisphosphonate binding levels 24 hours after injection does not necessarily reflect binding capacity. The specific measurement of both the rate of bisphosphonate uptake as well as the binding plateau may better reflect interaction of these compounds within the bone matrix. The aim of this study was to determine whether binding kinetics of bisphosphonates, assessed by repeated FMT measurements during the early phase of bisphosphonate binding, could temporally resolve localized changes in bisphosphonate uptake in bone-loss and bone-gain mouse models in vivo.

Materials and Methods


Twelve-week-old female, CD-1 nude mice were purchased from Charles River (Wilmington, MA, USA). All animals were kept in a temperature- and humidity-controlled environment, with a 12-hour light/dark cycle, with access to food and water ad libitum. Animal experiments and care were in accordance with the guidelines of institutional authorities and approved by the Ethics Committee for Animal Experiments at Christian-Albrechts-Universität-zu-Kiel [V 312-72241.121-33]. Mice were anesthetized with intraperitoneal injections of 80 mg/kg ketamine (Aveco Pharmaceutical, Fort Dodge, IA, USA) and 10 mg/kg xylazine (Rugby Laboratories, Duluth, GA, USA). For long-term anesthetization, additional half-dose administrations of ketamine and xylazine were given upon initial signs of waking. Animals were separated into three groups. Nine nonoperated control animals, 9 ovariectomized (OVX) animals, which were imaged 3 days (short term) and 14 days (long term) after OVX, and 9 parathyroid hormone (PTH)-treated mice, which were OVX for 11 days, then received daily PTH injections for 3 days for a total of 14 days OVX and 3 days PTH treatment (Fig. 1). PTH-treated and long-term OVX mice were further divided into two groups with 7 mice being subjected to kinetics analysis and 2 mice used for nondecalcified sectioning as described later. Animals were ovariectomized via their dorsal side. Human parathyroid hormone fragment 1-34 (Sigma-Aldrich, St. Louis, MO, USA) was given subcutaneously at a dose of 100 µg/kg daily.

Figure 1.

Study design overview. Sample groups are boxed, treatments are bolded, and experimental measurements are italicized.

In vivo micro-CT analysis

Confirmation of expected bone loss or gain resulting from ovariectomy or PTH treatment was obtained from micro-CT assessment of skeletal changes in vivo. Anesthetized mice were placed in full-body holders and the tibias aligned by visual inspection. Scans were made using a vivaCT 40 micro-CT (ScancoMedical, Brüttisellen, Switzerland) at an isotropic voxel size of 19 µm (70 kVp, 114 µA, 250 ms integration time, 1000 projections on 180° 2048 CCD detector array, cone-beam reconstruction). A 60-slice (1.05-mm) volume of interest (VOI), beginning at the most proximal point of the epiphyseal trabecular bone and extending into the metaphyseal region and drawn along the periosteal surface, was defined in the baseline scan of each animal using an automated contouring method.[19] Baseline VOIs were transferred to the follow-up scans using an image registration approach to ensure analysis of consistent VOIs at each time point.[20] Bone mineral density (BMD) was calculated from the grayscale micro-CT images as the mineral content divided by the total volume, encompassing both bone tissue and marrow (Image Processing Language v5.15, ScancoMedical).

Bisphosphonate binding kinetics

Anesthetized mice were injected intravenously with 100 µL phosphate-buffered saline (PBS) containing 2 nmol of dissolved OsteoSense750EX (PerkinElmer, Waltham, MA, USA), a fluorescently conjugated pamidronate derivative,[21] and imaged 2, 4, 6, 8, 10, 15, 20, and 30 minutes after injection using the NightOwl planar imaging system (Berthold Technologies, Bad Wildbad, Germany) to qualitatively determine kinetic distribution. A phantom was placed over the mouse bladder to help position the limbs and prevent obscuring of the limb signal from the urinary pool of bisphosphonate. For all in vivo, quantitative assessments of bisphosphonate binding, anesthetized mice were imaged immediately after injection and every subsequent ∼15-minute interval for 210 minutes using the FMT 2500LX from Visen Medical (PerkinElmer). Images were reconstructed and VOIs around the proximal tibia region, as determined from the photographic image, were quantified using the TrueQuant software as previously described.[14] Kinetics curves were generated using the average fluorescent intensity of both proximal tibia regions over time using one-phase association curves, with Y0 values constrained to zero, generated from Prism (version 5, GraphPad Software, LaJolla, CA, USA). Graphs show only the first 100 minutes of imaging to more clearly show changes in initial binding kinetics. Because of the low resolution of FMT, proximal tibia VOIs may also contain partial fluorescence of the distal femur.

Lab analysis

Blood was collected from the tail vein of control, long-term OVX, and PTH-treated mice before bisphosphonate injection. Levels of skeletal osteoblast and osteoclast activity were assessed using an osteocalcin (DRG Diagnostics, Marburg, Germany) and tartrate-resistant acid phosphatase (TRAP) ELISA assays (Immunodiagnostic Systems, Frankfurt, Germany) on blood serum.

Fluorescent and micro-CT imaging of bone sections

PTH-treated and untreated OVX control mice were injected with OsteoSense750 intravenously, then killed 15 minutes or 100 minutes after injection. Nondecalcified femurs and tibia were fixed in 10% buffered formalin, embedded in methacrylate, then cut in ∼50-µm sections. Bone sections were scanned on the LI-COR Odyssey infrared imaging system with a resolution of 21 µm and an offset of 1 mm, excited at 785 nm and collected at wavelengths greater than 810 nm. Several regions of interest of equal size were placed randomly over growth plate, trabecular, and cortical bone regions of Odyssey bone section scans and subsequent fluorescence signal intensity was quantified using ImageJ. A Leica DM2500 fluorescent microscope equipped with a DFC 360FX camera and Y7 filter cube (Leica Microsystems, Wetzlar, Germany) was used to visualize bisphosphonate binding at the growth plate and along trabecular and cortical bone surfaces. Lacunae and osteocyte labeling was visualized using a fluorescent microscope equipped with an Imager Intense CCD camera from LaVision (Bielefeld, Germany), epilumination source with LP590 filter, and a LP640 emission filter. Micro-CT scans of the slides using the same settings as described above but with a voxel size of 10 µm were made with the slide surface aligned with the axial scan direction. After reconstruction, the images were realigned such that the slide surface was perpendicular to the axial scan direction to allow viewing in 2D. The central slice containing bone was used for analysis. Images from the Odyssey fluorescent scanner and micro-CT scans were overlaid using an affine registration with a normalized mutual information metric (Amira 5.4.3, Visualization Sciences Group, Berlin, Germany). The images were then processed in Matlab (R2010b; Mathworks, Natick, MA, USA), where a threshold (CT value 100, fluorescence value 50) was applied to create binarized masks, which were combined to determine the pixels containing both micro-CT values within the bone tissue region and fluorescence signal. Histograms of the percentage of total fluorescence versus CT value (ie, tissue mineral density [TMD]) were created and fitted to Gaussian distributions.

Statistical analysis

Fluorescent rate constants (K) and plateau values were calculated for curves from the line of best fit for each subject using the formula: Y = Y0 + (Plateau – Y0)(1–exp−Kx). Plateau and rate constant were multiplied to generate a plateau-weighted rate constant for each mouse representing overall curve characteristics. Comparison between groups was made using two-sample t tests using the Welch-Satterthwaite method to avoid the assumption of equal variances. For micro-CT versus fluorescence comparison in coregistered images, Gaussian distributions were fitted to binned data and difference in mean CT values 15 minutes and 100 minutes after injection was assessed. Any p values <0.05 were considered to be statistically significant. T values and effect measures, ie, a z-transform calculated as the difference in mean between groups divided by pooled standard deviations, were determined from two sample t tests to assess the ability of the plateau, rate constant, and plateau-weighted rate constant to detect differences between control, short-, and long-term OVX mice.


Fluorescently labeled bisphosphonate localize to the bone within minutes of injection

To understand the kinetics and distribution of fluorescent bisphosphonates, control mice were subjected to NightOwl planar imaging after injection (Fig. 2). Bisphosphonate fluorescence initially appears diffuse throughout the mouse after injection but is lost in the soft tissue as it accumulates on the bones over 30 minutes.

Figure 2.

Kinetic distribution of fluorescent bisphosphonate. Fluorescently conjugated bisphosphonate was injected and imaged 2, 4, 6, 8, 10, 15, 20, and 30 minutes after injection using the NightOwl fluorescent planar imaging system documenting accumulation in the knee region and clearance from soft tissue. A nontransparent part of the animal holder was used to position the limbs and prevent the obscuring of limb fluorescence the urinary bisphosphonate pool. Signals on this 2D image are affected by attenuation of overlying soft tissue, precluding, eg, visualization of the spine, in this ventral exposure.

Ovariectomized mice show decreased rate constant and binding plateau values compared with control mice

After ovariectomy, serum biomarker analysis of the osteoblast marker osteocalcin showed a significant reduction in osteoblast activity (p = 0.047) in ovariectomized mice (1.70 ± 0.51 ng/mL) compared with control mice (4.18 ± 0.85 ng/mL), whereas the osteoclast activity marker TRAP showed no significant changes (p = 0.711). In vivo micro-CT analysis showed significant losses in bone mineral density after 14 days (long-term) ovariectomy (Fig. 3A). Fluorescent bisphosphonate-injected mice were imaged and the fluorescent intensity of the proximal tibia regions (Fig. 3B) quantified and subjected to nonlinear regression analysis (Fig. 3C). Mice ovariectomized for 3 days (short-term OVX) showed significantly reduced rate constants and binding plateaus compared with control mice (Fig. 3D, E). Long-term ovariectomized mice showed a further significant reduction in rate constant values from short-term ovariectomized mice but no significant change in binding plateau values. To quantify overall changes in bone mineral deposition in a single parameter, plateau-weighted rate constants were calculated for each individual mouse (Fig. 3F). Long-term ovariectomized mice showed significantly lower plateau-weighted rate constants compared with short-term ovariectomized mice, which in turn showed significantly lower values than control mice.

Figure 3.

Ovariectomy results in decreased rate constant and binding plateau values of fluorescently labeled bisphosphonate. Control, short-term (day 3), and long-term (day 14) OVX mice were imaged by micro-CT to confirm loss of bone mineral density (BMD) (A). Mice were injected with fluorescently conjugated bisphosphonate and imaged by FMT. (B) Isosurface rendering of 3D FMT reconstruction. Expanded windows show representative, control tibia region fluorescence over time after injection. Tibia regions were quantified and subjected to nonlinear regression analysis. (C) Overall best-fit nonlinear regression curves are shown for each group. Rate constants and plateau values were then calculated for each individual mouse. Ovariectomized mice showed significantly reduced rate constants (D) and binding plateau values (E) compared with control mice. Plateau values and rate constants were multiplied to generate a single numerical parameter for each mouse (F). Long-term OVX mice show a significant reduction in plateau-weighted rate constants from short-term OVX mice, which in turn show a significant decrease from control mice, with no overlap between groups. Dotted lines represent 95% confidence interval of fitted curve. Graphs represent mean values ± SD. Whiskers represent extreme maximum and minimum values. (*p < 0.05, **p < 0.01, ***p < 0.001) (n = 9)

Bisphosphonate binding kinetics as a tool to monitor pharmacological intervention

In our assessment whether bisphosphonate binding kinetics could be used as a tool to monitor increases in new bone formation resulting from pharmacological intervention, binding curves were created for ovariectomized mice with or without PTH treatment (Fig. 4A). Mice treated with PTH showed a significantly increased rate constant (Fig. 4B) and plateau-weighted rate constant values (Fig. 4D), as well as significantly increased osteocalcin (6.44 ± 1.49 ng/mL) (p = 0.017) and TRAP levels (6.58 ± 0.28 U/L) (p = 0.044) compared with untreated controls (1.70 ± 0.51 ng/mL and 4.69 ± 0.69 U/L, respectively). No significant changes in binding plateau values were observed (Fig. 4C). BMD, assessed by micro-CT, showed no significant increase 3 days after PTH treatment (p = 0.233) (Fig. 4E); however, to confirm the expected bone anabolic effect, additional micro-CT assessment was carried out after 14 days of PTH treatment, which showed a significant BMD increase of 32.7 mg HA/cm3 compared with untreated controls (p = 0.007).

Figure 4.

Ovariectomized mice treated with PTH showed increased rate constants and plateau-weighted rate constants. OVX mice were treated with PTH for 3 days, then assayed for bisphosphonate binding kinetics using FMT (A). Bisphosphonate binding curves were used to calculate rate constants (B), binding plateaus (C), and plateau-weighted rate constants (D) for each group. (E) Micro-CT showed a nonsignificant 2% increase in BMD (p = 0.233). Dotted lines represent 95% confidence interval. Graphs represent mean values ± SD. Whiskers represent extreme maximum and minimum values. (***p < 0.001) (n = 9)

Bisphosphonates bind preferentially to low TMD regions associated with bone ossification and modeling

To determine the spatial distribution of bisphosphonate binding at early and late time points, nondecalcified limb sections of OVX and PTH-treated mice were imaged ex vivo 15 and 100 minutes after bisphosphonate injection using a near infrared imaging scanner (Fig. 5A). Relative fluorescent intensity in the growth plate, trabecular, and cortical bone regions were quantified (Fig. 5B). Mice treated with PTH showed significantly increased bisphosphonate binding near the growth plate, as well as a trend toward increased binding in the trabecular region 15 minutes after injection (p = 0.053), compared with untreated controls. Bisphosphonate localization in OVX and PTH-treated mice was confirmed by fluorescent microscopy (Fig. 5C). Analysis of these images shows bisphosphonates highly localize to regions adjacent to, but not within, the growth Plate 15 minutes after injection and, to a lesser extent, along cortical and trabecular bone surfaces in OVX mice 100 minutes after injection. Treatment with PTH resulted in increased labeling of all bone surfaces, including osteocyte lacunae near the cortical bone surfaces 15 minutes after injection. To spatially correlate bisphosphonate binding and local TMD, micro-CT and fluorescent Odyssey scans from PTH-treated mouse sections were coregistered, and pixels containing both bone mineral and fluorescence were compared (Fig. 6A). Analysis shows binding of fluorescent bisphosphonates occurs preferentially at regions of low TMD 15 minutes after injection and later to regions of significantly greater TMD (p = 0.0175) 100 minutes after injection.

Figure 5.

Bisphosphonates preferentially bind to bone regions associated with new bone formation. Ovariectomized, untreated mice and OVX mice treated with PTH were injected with fluorescent bisphosphonate and euthanized 15 or 100 minutes after injection. Nondecalcified sections of the tibia and femur were prepared and imaged using the Odyssey near infrared fluorescent scanner (A). Regions of interest encompassing the growth plate, trabecular bone, or cortical bone were quantified in relative fluorescent units (RFU) (B). Fluorescence localization was confirmed by fluorescent microscopy (C). Top panels show bright-field images to depict microstructures of bone sections and middle panels show overall fluorescent localization of bisphosphonates. Labeling initially occurs adjacent to the growth plate (OVX 15 minutes) and later along trabecular and cortical surfaces (OVX 100 minutes). Treatment with PTH results in increased labeling of all bone surfaces 15 minutes after injection. Arrows indicate the growth plate region; arrowheads indicate the labeled trabecular bone surfaces. Scale bar = 200 µm. Lower panels show high-magnification fluorescent images of the cortical bone, showing labeling of bone surfaces, as well as osteocyte lacunae. Scale bar = 50 µm. Graphs represent mean values ± SD. (*p < 0.05, **p < 0.01, ***p < 0.001)

Figure 6.

Bisphosphonates preferentially bind to regions of low bone density. Nondecalcified sections from PTH-treated mice euthanized 15 and 100 minutes after bisphosphonate injection were imaged using the Odyssey near infrared fluorescent scanner as well as by micro-CT (A). Micro-CT image was binarized and bone tissue segmented from bone marrow signal. Images were registered and subjected to pixel-by-pixel analysis within the bone tissue compartment using grayscale micro-CT images. (B) Gaussian distributions show bisphosphonates bind preferentially to significantly lower tissue mineral density regions 15 minutes after injection compared with 100 minutes (p = 0.0175).


In this study, we demonstrate fluorescently labeled bisphosphonates to be a valuable tool for quantifying localized changes in bone metabolism in vivo. Although previous longitudinal studies have focused on plateau values of fluorescence days after injection, our results suggest expanded utility may be obtained by monitoring bisphosphonate binding immediately after injection. By analyzing the early binding kinetics of bisphosphonates, we were able to monitor dynamic properties of the bone environment. Furthermore, the ease and simplicity of this approach allows for expanded application in the field of preclinical drug testing and a greater understanding of the complex processes associated with bone remodeling.

Mice ovariectomized for 3 days showed significant reductions in binding rate constants and plateau binding values with further reductions in both parameters after 14 days of ovariectomy. These results are consistent with serum data, which show a significant reduction in osteoblast activity leading to a reduction in new mineral deposition. Although no significant changes in serum TRAP levels were observed in our model, we cannot discount the important role of osteoclasts, and resorbing bone surfaces, on bisphosphonate binding. Previous works have suggested extensive bisphosphonate binding in resorption pits, especially in the case of lower-affinity bisphosphonates.[12] A lack of detectable change in osteoclast activity may be attibutable in part to the absence of T cells in nude mice, which have previously been implicated in stimulating osteoclastogenesis and shown to play a role in estrogen-deficient bone loss.[22-25] However, the lack of significant serum TRAP changes suggests that the changes in binding kinetics observed in this study are primarily reflecting changes in osteoblast activity. With no detectable difference in serum osteoclast activity and a significant decrease in osteoblast activity, a reduction in total bone can be expected, consistent with micro-CT data showing a significant reduction in bone mineral density as the result of ovariectomy. By combining the rate constant with the plateau binding values, we have generated a single numerical parameter for each subject.[26] This plateau-weighted rate constant summarizes changes in the rate constant and plateau values, thereby facilitating interpretation of the kinetic data. Furthermore, plateau-weighted rate constants showed greater average t values and effect measures (t = 5.0786, z-transform = 2.625) compared with either plateau (t = 4.385, z-transform = 2.261) or rate constant values (t = 4.446, z-transform = 2.293), suggesting plateau-weighted rate constant values show the greatest ability to distinguish control, short-term, and long-term OVX sample groups. This may suggest greater utility in using plateau-weighted rate constants, perhaps not only for fluorescent markers in animals but also in a clinical setting using radioactive tracers to easily distinguish healthy patients from those with altered bone metabolism.

Next, we sought to investigate whether binding kinetics could be used to assess pharmacological intervention leading to increased bone formation. After 3 days of intermittent PTH treatment, no significant differences were observed by micro-CT or plateau binding values. However, rate constants showed a highly significant increase after PTH treatment compared with the untreated control. Serum analysis showed significantly elevated osteoblast and osteoclast activity, suggesting that, although micro-CT showed no significant net bone mineral gained within 3 days of PTH treatment, composition of the bone surface may have been altered. Other researchers have reported that administration of PTH for 3 days in rats failed to show significant changes in the uptake of 99mTc-pyrophosphate 2 hours or 6 hours after radiotracer injection.[27] Using our binding kinetics approach, we were able to detect early changes in bone metabolism in mice after 3 days of PTH treatment, before conventional micro-CT or single time-point FMT measurements 24 hours after injection, further supporting increased utility of monitoring binding kinetics. It would be of interest to investigate whether different types of bone-anabolic treatments may show dissimilar patterns of rate constants and plateau values potentially reflecting variable degrees of bone activation.

Localization studies of bisphosphonate binding 15 and 100 minutes after injection show preferential binding of the bisphosphonate probe OsteoSense to the region adjacent to the growth plate, associated with endochondral ossification[28] and, to a lesser extent, along trabecular and cortical regions associated with bone remodeling, or quiescent bone surfaces. Quantification of Odyssey bone scans showed increased bisphosphonate binding near the growth plate in PTH-treated mice 15 minutes after injection compared with the untreated controls. Further analysis of early binding shows bisphosphonates localize preferentially to low-TMD bone, associated with new bone in a state of primary mineralization, shortly after injection and, to a lesser extent, to highly mineralized bone associated with fully mineralized, quiescent bone.[29] These data are consistent with increased binding of bisphosphonate at regions associated with high osteoblast activity and newly forming bone.

From these observations, we propose that the dynamics of bisphosphonate binding observed in this study reflect the binding to two bone types, high-uptake capacity and low-uptake capacity, comprising the bone surface. High-uptake-capacity bone would be composed of newly deposited mineral at the site of osteoblast activity and contains a high volume of exposed surface minerals. The amount of high-uptake-capacity bone reflects both the density of active osteoblasts along the bone surface, as well as the level of bone-forming activity of these osteoblasts.[30] Additionally, newly deposited bone may transition to fully mineralized, low-uptake-capacity bone when not undergoing bone remodeling. This transition may also affect the form of mineral present within the bone. Calcium phosphate in an amorphous state, present in newly forming bone, has previously been shown to have higher uptake of diphosphonates compared with calcium phosphate in its crystalline form, present in fully mineralized, quiescent bone, in vitro.[31] We propose changes in the amount of high- and low-uptake-capacity bone, along with injected dose, would both be reflected in the rate constant and binding plateau of bisphosphonate binding kinetics.

This ability to distinguish sites of high bone turnover from regions of low or no bone activity may also have important implications in tumor site identification and treatment. Osteosclerotic bone lesions, such as in the case of prostate cancer, result in marked increases in bone formation. According to the findings presented here, the expected increase in bone turnover at the tumor site would accumulate bisphosphonate probes more rapidly than in the surrounding tissue, enabling the detection and preferential targeting of the lesions. Binding kinetics may also prove useful for detecting metastases that result in mixed osteolytic/osteosclerotic lesions.[32, 33] These lesions may result in mild or no overall bone loss or gain, making them difficult to detect by CT, but the marked increase in bone turnover within the tumor environment could still be detected from the corresponding kinetic uptake parameters.

One aspect that requires consideration in future methodological refinements is the differentiation of localized changes (as induced by bone metastases) versus systemic changes (eg, owing to medications) and the effects of surrounding soft tissue on fluorescent measurements. In this study, we present data on the proximal tibia region because of the relatively small soft tissue present to attenuate and scatter fluorescence, as well as its importance as a high-frequency site of bone metastases in preclinical models. However, overall changes in skeletal bone uptake outside our region of interest will affect the blood pool levels and, correspondingly, the ability of the tibia to bind free bisphosphonate. In the instance of PTH treatment, mice showed rapid uptake of bisphosphonate at the proximal tibia region but failed to show any significant changes in binding plateau values. This may be because of rapid uptake of available bisphosphonate by other skeletal sites, reducing the blood pool of bisphosphonate available for binding and, thus, preventing further binding at the proximal tibia at later time points.[34] Clinical evaluation of bisphosphonate binding using radioactive labels supplement bone accumulation measurements with blood serum levels, urinary pool, or soft-tissue retention of unbound bisphosphonates.[35-38] Body mass of the mouse will also play a role in bisphosphonate uptake, clearance, and dosing. And, although body mass was not significantly altered in this study, consideration must be given for situations in which significant weight loss may be observed. A measure of blood pool bisphosphonate may reveal that tibia fluorescence was limited more by the rapid clearance of bisphosphonate from the blood and not by the uptake of bisphosphonate by the tibia itself. FMT is tomographic, but with limited spatial resolution and scan region obtained, we were unable to calculate arterial input functions, renal clearance, or blood perfusion within the proximal tibia region for each mouse. As a result, we cannot discount either decreased delivery of bisphosphonate to the bone or increased renal clearance of unbound bisphosphonate in the case of ovariectomy. No significant changes in health or weight, common symptoms of impaired renal function, were detected; however, recent bone perfusion experiments using laser Doppler techniques suggest blood perfusion may be affected in ovariectomized mice depending on their genetic background.[39] These uncontrolled parameters limit the exact quantification of bone kinetic parameters as done in this study, but with improvements in imaging technology and implementation of more complex imaging protocols may become feasible in the future.

It is also of interest to note that binding curves in this study were obtained using a pamidronate derivative, reflecting the binding kinetics of relatively high-affinity bisphosphonates. Previous works have shown that different bisphosphonates vary in their affinity for bone surfaces and their level of bone penetration.[12, 40] The use of medium- or low-affinity bisphosphonates as the targeting molecule may better reflect changes in osteoclast activity because of their relatively greater uptake in regions of bone resorption compared with quiescent surfaces.[12] Additional histological analyses will be needed to more precisely define subregions within the bone with more rapid uptake of high- versus low-affinity bisphosphonates. In particular, methodological refinement should include histological analyses of bone surfaces revealing quiescent, resorbing, and bone-forming regions and the corresponding levels of bisphosphonate binding.

The use of fluorescent probes suggests this method has the possibility to be applied to monitoring multiple wavelengths and multiple probes detecting multiple aspects of bone dynamics in vivo at macroscopic resolution simultaneously, while also allowing the use of more high-resolution modalities such as fluorescent confocal or two-photon microscopy for analysis at the microscopic level. The challenge remains to validate this method in patients using radioactive tracers conjugated to bisphosphonates and translate this simpler approach into wider clinical use. Because 99mTc-MDP display relatively slow blood and soft-tissue clearance, the measurement of bisphosphonate binding at early time points is confounded by the high percentage of tracer retained in the soft tissue in the moments immediately after injection of tracer. The application of new radioactive bisphosphonates with variable affinities for the bone matrix can have an impact in nuclear medicine in the refined assessment of metabolic bone disorders and warrants clinical investigation.

In conclusion, we have developed and tested a new molecular imaging method for in vivo assessment of bone metabolism, specifically bone mineral deposition. Using OsteoSense, a fluorescently labeled bisphosphonate, we were able to noninvasively visualize localized changes in bone turnover, including bone loss and bone anabolic treatment models, at very early time points. The analysis of three parameters—the rate constant, the binding plateau, and the plateau-weighted rate constant—provides noninvasive insights into the presence of high- and low-uptake-capacity bone mineral, important aspects in osteoporosis and for the detection and differentiation of lytic versus osteosclerotic bone metastases. Our method has potential for further refinement with the goal of better quantification and for translation to human application using radionuclide tracers instead of fluorescent markers.


All authors state that they have no conflicts of interest.


The authors are grateful to Gabriele Trompke and Gaby Nessenius from the UKSH clinic of oral and maxillofacial surgery for their assistance preparing nondecalcified bone sections. We thank Dr Vladimir Ermolayev and Sarah Glasl (Technische Universität München), Dr Twan Lammers (University of Aachen), and Dr Arndt Rohwedder (UKSH) for assistance with fluorescent microscopy images, and Dr Jürgen Baudewig for his assistance with image registrations using Amira.

Financial support was provided by the Deutsche Forschungsgemeinschaft (DFG) through the forschergruppe 1586 SKELMET and by the research grant from the state of Schleswig-Holstein and the European Union ERDF-European Regional Development Fund (MOIN CC, Zukunftsprogramm Wirtschaft). Research support was provided by PerkinElmer in the form of discounted reagents.

Authors' roles: Study design: RJT, GMC, OW, CCG, and ST. Study conduct: RJT, GMC, MM, OW, and ST. Data collection: RJT and MM. Data analysis: RJT, GMC, and MM. Data interpretation: RJT, GMC, and CCG. Drafting manuscript: RJT. Revising manuscript content: GMC, MM, CCG, and ST. Approval of final version of manuscript: RJT, GMC, CCG, and ST. ST takes responsibility for the integrity of the data analysis.