Purpose: The aim of this study was to assess the antiangiogenic treatment effects of zoledronic acid (ZA) and sunitinib malate (SM) noninvasively in experimental breast cancer bone metastases by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and vessel size imaging.
Experimental Design: Nude rats bearing bone metastases after inoculation of MDA-MB-231 breast cancer cells were treated with ZA (40 μg/kg weekly; n = 8 rats), SM (20 mg/kg daily; n = 8 rats), or their combination (ZA and SM; n = 8 rats) and compared with sham-treated controls (n = 10 rats). Vascular changes in bone metastases were longitudinally imaged in vivo using DCE-MRI [amplitude (A) and exchange rate coefficient (kep)] and vessel size imaging [blood volume (BV) and vessel size index (VI)]. In addition, antiresorptive and antitumor changes were assessed in these lesions by flat-panel volumetric computed tomography as well as morphologic MRI and diffusion-weighted imaging.
Results: In bone metastases, significant changes in A, kep, BV, and VI in accordance with decreased blood volume and vessel permeability as well as with increased mean vessel diameters were observed after application of ZA and SM as compared with controls. In this longitudinal study, antiangiogenic changes preceded the inhibition of osteolysis and antitumor effects after treatment.
Conclusions: These results indicate vessel remodeling in breast cancer bone metastases on ZA and SM treatment and implicate substantial effects on imaging and treatment of malignant bone lesions. Clin Cancer Res; 16(12); 3215–25. ©2010 AACR.
This article is featured in Highlights of This Issue, p. 3089
The noninvasive imaging methods used in our study for assessment of antiangiogenic treatment response in experimental bone metastases, including dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and vessel size imaging, can be applied to humans with such skeletal lesions accordingly. Translation of the presented results into the clinical situation is in particular facilitated as a regular MRI scanner for humans at 1.5 T was used, which is widely available.
The observed antiangiogenic effects of treatment with zoledronic acid and sunitinib malate in experimental bone metastases were associated with antiresorptive as well as antitumor effects and require further validation in clinical studies. When applying zoledronic acid and sunitinib malate in patients, DCE-MRI and vessel size imaging could be used for early and quantitative assessment of treatment response in malignant bone lesions.
Bone is among the most frequent locations of metastasis in patients with breast cancer. In these patients, bone marrow represents a fertile soil for cancer cells by homing factors that stimulate various processes in the development and progression of bone metastases such as tumor cell proliferation, bone resorption, and angiogenesis. Angiogenic factors in the bone marrow are essential for the cross-talk of osteoclasts, osteoblasts, and bone marrow endothelial cells. In patients with skeletal metastases, the expression of angiogenic factors was associated with osteoclast-mediated osteolysis and, vice versa, antiangiogenic therapy resulted in antiresorptive treatment response in these lesions (1–4). Recently, a close anatomic relation between capillaries and bone remodeling units consisting of osteoclasts and osteoblasts was observed in osteolytic lesions, indicating a possible interaction between angiogenesis and bone resorption (5). For these reasons, angiogenesis plays a crucial role in the pathogenesis of breast cancer bone metastasis, but the effect of the inhibition of angiogenesis is still unclear in these lesions.
Both antiangiogenic and antiresorptive effects were reported for bisphosphonates as current standard treatment for breast cancer bone metastases as well as for the tyrosine kinase inhibitor sunitinib malate (SM; ref. 6). On treatment with bisphosphonates including zoledronic acid (ZA), vascular endothelial growth factor (VEGF) was decreased in the blood of breast cancer patients (7). ZA was furthermore shown to inhibit proliferation of endothelial cells, to modulate endothelial cell adhesion and migration, as well as to reduce vessel sprouting (8, 9). The multitargeted receptor tyrosine kinase inhibitor SM inhibits important vascular targets such as VEGF receptor and platelet-derived growth factor receptor. SM has already been approved for the treatment of gastrointestinal stroma tumors and renal cell carcinomas, but was shown to exhibit strong antiangiogenic activities in many more primary tumors including breast cancer (10). Interestingly, SM has been reported to inhibit osteolysis in experimental breast cancer bone metastases (11, 12). However, the antiangiogenic effects of ZA and SM have not been studied in bone metastases.
For noninvasive assessment of angiogenesis, magnetic resonance imaging (MRI) techniques such as dynamic contrast-enhanced MRI (DCE-MRI) and vessel size imaging (VSI) can be applied to assess vascular changes such as vessel remodeling by the calculation of quantitative biomarkers (13, 14). DCE-MRI is widely used in preclinical cancer research as well as in oncologic patients, including applications in bone metastases, whereas VSI is a comparatively new technique (15, 16). The current standard to assess treatment response in breast cancer patients with bone metastases is the determination of tumor size in bone by MRI and computed tomography (CT; ref. 17). Furthermore, by detecting changes in cellularity, the apparent diffusion coefficient (ADC) from diffusion-weighted imaging (DWI) was found to be an imaging biomarker of early treatment response in bone metastases (18).
The aim of this study was to assess the antiangiogenic treatment effects of the bisphosphonate ZA and the receptor tyrosine kinase inhibitor SM in experimental breast cancer bone metastases using DCE-MRI and VSI, as well as the resulting antitumor and antiresorptive effects of these therapies by morphologic MRI, DWI, and flat-panel volumetric CT (VCT). To our knowledge, this is the first study to noninvasively investigate the antiangiogenic effects of ZA and SM in bone metastases in vivo.
Materials and Methods
Cell lines and culture conditions
The human estrogen-independent breast cancer cell line MDA-MB-231 was obtained from the American Type Culture Collection. Tumor cells were cultured routinely in RPMI 1640 (Invitrogen) supplemented with 10% FCS (Sigma). All cultures were kept under standard conditions (37°C, humidified atmosphere, 5% CO2) and passaged two to three times a week to keep them in logarithmic growth.
Animal model and treatment
Experiments were approved by the governmental animal ethics committee (Regierungspräsidium Karlsruhe, Germany). Nude rats (RNU strain) were obtained from Harlan Winkelmann at the age of 6 weeks. They were housed at specific pathogen-free conditions in a mini-barrier system of the central animal facility under controlled conditions. For all procedures including tumor cell inoculation and in vivo imaging, rats were anesthetized with a mixture of nitrous oxide (1 l/min), oxygen (0.5 l/min), and isoflurane (1-1.5 vol%). For tumor cell inoculation, 105 MDA-MB-231 tumor cells were injected in the right superficial epigastric artery of nude rats as described previously (19). Resulting bone metastases were observed exclusively in the femur, tibia, and fibula of the right hind leg of these animals.
Nude rats (n = 34) bearing bone metastases were randomly distributed into four groups. After the initial imaging on day 0 (30 days after tumor cell injection; for details, see below), SM (SU-11248, Pfizer; n = 8, oral application of 20 mg/kg daily), ZA (Roche; n = 8, s.c. application of 40 μg/kg weekly), or their combination (n = 8; 20 mg/kg SM, oral application daily, and 40 μg/kg ZA, s.c. application weekly) was given to nude rats until the end of the observation time at day 25 after initiation of therapy and compared with untreated controls (n = 10, s.c. application of saline instead of SM or ZA).
In vivo imaging
In vivo imaging of experimental breast cancer bone metastases was done at day 0 (before initiation of therapies) as well as 5, 15, and 25 days after initiation of therapies using VCT and MRI as described previously (20, 21).
VCT imaging was done on a prototype flat-panel equipped volumetric computer tomograph (Volume CT, Siemens) with the following parameters: tube voltage, 80 kV; tube current, 50 mA; scan time, 51 seconds; rotation speed, 10 seconds; frames per second, 120; matrix, 512 × 512; and slice thickness, 0.2 mm.
Magnetic resonance imaging
MRI was done on a 1.5-T MR-scanner (Symphony, Siemens) using an appropriate home-built coil for radiofrequency excitation and detection (20, 21). The animals were imaged with the following sequences. (a) T2 weighted images: turbo spin-echo sequence [orientation, axial; TR, 3,240 ms; TE, 81 ms; matrix, 152 × 256; field of view (FOV), 53.4 × 90 mm2; slice thickness, 1.5 mm; averages, 3; images, 15; scan time, 3:40 min]; (b) diffusion-weighted imaging: half-Fourier acquisition single-shot turbo spin-echo sequence (b = 0, 50, 100, 200, and 600 s/mm2 through the largest diameter of the tumor; orientation, axial; TR, 4,000 ms; TE, 155 ms; matrix, 72 × 128; FOV, 45 × 80 mm2; slice thickness, 2 mm; averages, 10; images, 15; scan time, 5 × 2:00 min); (c) dynamic contrast-enhanced imaging: saturation recovery turbo flash sequence (through the largest diameter of the tumor; orientation, axial; TR, 373 ms; TE, 1.86 ms; matrix, 192 × 144; FOV, 130 × 97.5 mm2; slice thickness, 5 mm; measurements, 512; averages, 1; images, 512; scan time, 6:55 min) while infusing 0.1 mmol/kg Gd-DTPA i.v. (Magnevist, Schering) over a time period of 10 seconds; and (d) vessel size imaging: T2 weighted spin-echo sequences (TR, 6,000 ms; TE, 80 ms; matrix, 128 × 104; resolution, 0.5 × 0.5 × 1.5 mm3; averages, 1; images, 40; scan time, 13:00 min) were acquired before and 3 minutes after i.v. administration of ultrasmall supraparamagnetic iron oxide particles (200 μmol Fe/kg; Sinerem, Laboratoires Guerbet) as well as T2* weighted multigradient echo sequence (FLASH; TR, 320 ms; TE, 4.76-47.6 ms; matrix, 128 × 104; resolution, 0.5 × 0.5 × 1.5 mm3; averages, 3; images, 108; scan time, 7:00 min) pre- and post-injection of the ultrasmall supraparamagnetic iron oxide particles.
The volumes (in microliters) of the osteolytic lesions and the soft-tissue components were determined on unenhanced VCT and magnetic resonance images by volumetry (The Medical Imaging Interaction Toolkit, Germany). For DCE-MRI, DWI, and VSI, regions of interest were drawn manually around the tumor lesion to determine the respective values for amplitude (A; arbitrary units), exchange rate constant (kep; 1/min), vessel size index (VI; μm), blood volume (BV; arbitrary units), and ADC (mm2/s × 10−5) in bone metastases. DCE-MRI data were postprocessed according to the pharmacokinetic two-compartment model of Brix et al. (22) and maps for the ADC were calculated as reported previously (20, 21). Pixel-based data analysis for VSI to obtain the parameters VI and BV was done on an analysis software built with the rapid prototyping software Rad-Builder (Siemens Corporate Research). Due to technical failure during measurement, the number of VSI measurements on days 0 and 5 was reduced for animals that received ZA (day 0, n = 6 and day 5, n = 6) and for control animals (day 0, n = 5 and day 5, n = 6). Motion correction of the images was not necessary as the tumor-bearing hind leg of animals was fixed on a plastic bedding during image acquisition.
Bone storage and histology
The lower limbs from each animal were excised at sacrifice (day 25) and processed as described previously (20). For immunohistochemical staining, the following antibodies were used: primary antibodies—rabbit polyclonal against rat collagen IV (Progen Biotechnik GmbH; dilution 1:50) and rat α-smooth muscle actin (mouse monoclonal, Sigma Aldrich; dilution 1:400); secondary antibodies—donkey anti-rabbit Texas red sulfonyl chloride (Jackson ImmunoResearch; dilution 1:100) and goat anti-mouse Cy2 (Jackson ImmunoResearch; dilution 1:50). The preparations were rinsed again with PBS, counterstained using 4′,6-diamidino-2-phenylindole (Serva), dried, and coverslipped.
After immunohistochemical staining, five FOVs from sections of five representative animals of each group were examined using a Leica microscope (DMRE) with an adapted digital camera (F-view XS, Soft Imaging System). On these sections, quantitative analysis of marker densities was done by calculating positive area fractions; mean vessel diameters were also measured using Analysis Software (cellF, Olympus Soft Imaging Solutions). The microvessel density (MVD) was determined by calculating positive area fractions of collagen IV. When comparing immunostainings for CD31 and collagen IV to determine the MVD, a positive correlation was observed in soft-tissue components of bone metastases (data not shown).
For each animal, absolute values for osteolytic lesion and soft-tissue component volumes, A, kep, ADC, VI, and BV were plotted versus time after initiation of therapy. After quantitative histologic analysis, values of the positive area fractions as well as the vessel diameters were compared. For statistical comparisons, the respective values were compared between the control and treatment groups as well as between the treatment groups using the two-sided Wilcoxon test; P values <0.05 were considered significant.
Flat-panel VCT and MRI
After ZA monotherapy, decreased values for the volumes of osteolytic lesions and soft-tissue components were observed at days 5, 15, and 25 after initiation of therapy as compared with controls, but differences did not reach statistical significance (Fig. 1A and B; Table 1A and B). SM monotherapy and combination treatment resulted in a significant reduction of osteolytic lesion and soft-tissue component volumes at day 25 and additionally at day 15 for the combination treatment group compared with controls (Fig. 1A and B; Table 1A and B). Representative figures of osteolytic lesions and soft-tissue components of the control and treatment groups are shown in Fig. 2.
Although ADC values in animals treated with ZA were higher than those of controls at each time point after initiation of therapy, no statistically significant differences between the groups were observed (Figs. 1C and 2A-B; Table 1A and B). However, significantly increased ADC values were found in rats after SM monotherapy at days 15 and 25 as compared with controls, but not after combination therapy (Figs. 1C and 2C-D; Table 1A and B).
Dynamic contrast-enhanced MRI
ZA monotherapy resulted in significantly decreased A values at all observed time points after initiation of therapy as compared with controls, whereas an even greater reduction was observed after SM monotherapy and combination treatment at each time point of the observation period (Fig. 3A; Table 1A and B). Values for kep were significantly decreased in animals after ZA treatment than in untreated rats at day 25 (Fig. 3B; Table 1A and B). Furthermore, significantly decreased kep values were observed at days 5, 15, and 25 after SM monotherapy as well as at days 5 and 25 after combination treatment in comparison with controls (Fig. 3B; Table 1A and B). Representative color maps from bone metastases of the control and treatment groups are shown in Fig. 4.
Vessel size imaging
Compared with controls, significantly decreased values for the BV were observed in animals treated with ZA, SM, and the combination at day 25 as well as with SM and combination treatment additionally at day 5 (Fig. 3C; Table 1A and B). Values of the VI after ZA monotherapy were higher than in untreated controls, but were not significantly different between the groups (Fig. 3D; Table 1A and B). During SM mono- and combination therapies, significantly increased values of the VI could be calculated at days 5 and 25 as compared with controls (Fig. 3D; Table 1A and B). Representative color maps from bone metastases of the control and treatment groups are shown in Fig. 4.
Comparison of treatment groups
When comparing rats receiving ZA monotherapy with animals receiving SM monotherapy or combination treatment, significant differences between the groups for all vascular parameters (A, kep, BV, and VI; Fig. 3A-D; Table 1B) and for soft-tissue component volumes (Fig. 1B; Table 1B) were observed, but not for osteolytic lesion volumes (Fig. 1A; Table 1B) or ADC values (Fig. 1C; Table 1B). Comparison of rats receiving SM monotherapy and combination treatment resulted in significantly lower ADC values in animals of the combination treatment group at day 5 after initiation of therapy (Fig. 1C; Table 1B). However, no additional significant differences were observed for any parameter acquired by noninvasive imaging between these groups (Table 1B).
Immunofluorescent analysis in control animals revealed multiple irregular vessels with small diameters indicated by collagen IV staining, which were not colocalized with smooth muscle actin (SMA), along with few larger vessels showing collagen IV/SMA colocalization (collagen IV/SMA ratio, 3.5:1.0; Fig. 5A; Table 1C). After ZA monotherapy, the MVD as derived from collagen IV staining was significantly decreased when compared with controls (Fig. 5B; Table 1C). In these animals, particularly larger vessels showed an increased collagen IV/SMA colocalization (collagen IV/SMA ratio, 2.6:1.1; Fig. 5B; Table 1C) after ZA monotherapy in comparison with controls. In animals treated with SM mono- and combination therapies, decreased MVD (P < 0.01) and collagen IV/SMA ratios (0.9:0.9 for SM monotherapy and 0.9:1.0 for combination therapy) and increased mean vessel diameters (P < 0.01) were observed as compared with controls (Fig. 5C and D; Table 1C).
The aim of this longitudinal study was to assess the antiangiogenic, antiresorptive, and antitumor effects of ZA and SM in experimental breast cancer bone metastases using noninvasive imaging methods.
According to the two-compartment model of Brix, changes in the vasculature of bone metastases could be assessed noninvasively and in vivo by the DCE-MRI parameters A and kep (20–22). In these osseous lesions, we characterized the antiangiogenic properties of ZA for the first time and observed a decrease of A as early as 5 days after the initial application of ZA in comparison with control rats. This finding indicated an early reduction of regional blood volume after ZA treatment corresponding with a decreased MVD observed by immunofluorescent analysis, which was most probably a consequence of the previously described inhibition of endothelial cell proliferation and vessel sprouting on ZA treatment (8, 9, 23). Furthermore, a significant decrease of kep under ZA monotherapy was observed at the end of the observation time in bone metastases, indicating reduced vessel permeability as compared with controls. This change in kep was in accordance to a more pronounced pericyte coverage of the remaining vasculature due to vessel maturation.
In breast cancer bone metastases of animals treated with SM monotherapy and combination treatment, significantly decreased values for A and kep were assessed at almost each time point of the observation period. Compared with the above-mentioned results on ZA treatment, these findings indicated an even stronger decrease in blood volume and vessel permeability in line with the findings from immunohistology revealing vessel regression and maturation on SM monotherapy and combination treatment. These vascular changes after inhibiting the VEGF-VEGF receptor axis have been found in various preclinical and clinical studies in solid tumors, but such changes have not been observed and analyzed in bone metastases (24–27).
Differences in mean vessel size and regional blood volume of vasculature can be assessed by alterations in T2 and T2* relaxation times induced by superparamagnetic iron oxide nanoparticles using VSI (16). Applying this technique to bone metastases, the significant decrease in blood volume after treatment with ZA, SM, and the combination therapy corresponded to a significant decrease of DCE-MRI parameter A as an alternative measurement for regional blood volume. After SM monotherapy and combination treatment, significantly increased vessel size indices were observed, which was in line with a recent study showing increased mean vessel size diameters and vessel regression in subcutaneous tumors on SM treatment (14). When comparing vessel diameters determined by VSI and histology, an overestimation of mean vessel calibers by noninvasive imaging was observed, which has been described before (16, 28). In our study, small immature vessels most probably regressed due to VEGF inhibition, whereas bigger mature vessels persisted leading to an overall larger mean vessel diameter (29). Taken together, the results of significant changes in A, kep, BV, and VI values as assessed noninvasively in experimental breast cancer bone metastases indicate vessel remodeling including small vessel regression and vessel maturation after SM treatment. For ZA, antiangiogenic properties such as the inhibition of new vessel formation leading to a decreased regional blood volume have not been shown in bone metastases. However, these findings have to be confirmed in patients suffering from malignant bone lesions (e.g., skeletal metastases or multiple myeloma; ref. 30).
Besides the antiangiogenic effects of ZA and SM in our model, we observed an inhibition of bone resorption with these treatments. Also, previous reports described antiresorptive effects after antiangiogenic therapy, indicating a link between angiogenesis and osteolytic activity in bone metastases (4, 31, 32). A probable explanation for this phenomenon is the recently discovered fact that small capillaries are in close vicinity of bone remodeling units, where capillaries route pre-osteoclasts to the site of bone resorption (5). As a consequence, in particular, the disruption of small tumor vessels including capillaries, as observed in our study after SM treatment, might have impaired the migration and maturation of osteoclasts resulting in antiresorptive effects. Another possible mechanism for the inhibition of osteoclastic bone resorption by these drugs might be the direct inhibition of VEGF- and platelet-derived growth factor–associated effects. As an explanation, VEGF, platelet-derived growth factor, as well as the respective receptors were reported to be expressed by tumor and stroma cells in bone metastases and their interaction has been associated with osteoclast-mediated osteolysis (20, 33–36).
SM and ZA treatment resulted not only in antiangiogenic and antiresorptive effects but also had an effect on the soft-tissue component of bone metastases consisting of tumor cells and stroma as assessed by morphologic MRI and DWI. The ADC from DWI was associated with the cellularity of tissues and introduced as an imaging biomarker for early detection of antitumor effects in prostate cancer bone metastases (18, 37, 38). In our study, treatment with SM alone or in combination with ZA significantly decreased the volumes of the soft-tissue component and increased ADC values, indicating a reduction in cellularity. Also, ZA monotherapy decreased the tumor volume and increased ADC values in comparison with controls, but changes did not reach significance. These findings suggest a strong antitumor efficacy of SM and, to a lesser extent, of ZA monotherapy as assessed with DWI and morphologic MRI.
When comparing treatment groups in our study, antiangiogenic and antitumor effects were more pronounced in rats receiving SM monotherapy compared with rats receiving ZA monotherapy. However, the combined therapy with ZA and SM was not beneficial over SM monotherapy because no statistically significant advantage of the combination treatment was found for any parameter assessed by noninvasive imaging in our study. During treatment, even significantly lower ADC values indicating increased cellularity were found for the combination treatment as compared with SM monotherapy. Further studies with different dosages of SM and ZA should be conducted to analyze dose-dependent effects resulting from mono- and combination therapies with these drugs.
In patients with bone metastases, assessment of treatment response is pivotal in clinical practice (15). According to the most recent version of the Response Evaluation Criteria in Solid Tumors, measurements of the osteolytic lesion size by CT and the soft-tissue component by MRI are proposed methods to determine changes in the morphology of osseous lesions (17). However, a change in lesion size of bone metastases is usually assessed weeks or months after initiation of treatment because a change of morphology is slow, particularly in bone (39). According to our results, it seems worthwhile validating the noninvasive imaging parameters from DCE-MRI and VSI as potential early markers for assessment of treatment response after bisphosphonate and antiangiogenic treatments in patients with bone metastases by capturing vascular changes before a reduction in lesion size becomes apparent.
Here, we report on vessel remodeling in response to ZA, which is even more pronounced after SM treatment, in experimental breast cancer bone metastases as assessed noninvasively by quantitative imaging parameters from DCE-MRI and VSI. In our longitudinal study, these vascular changes preceded the antiresorptive and antitumor effects of ZA and SM in bone lesions, highlighting the imaging of angiogenesis in bone metastases as a valuable method for early and quantitative assessment of response after bisphosphonate and antiangiogenic treatments. Further clinical studies are required to confirm the proposed link between antiangiogenic effects and the inhibition of osteolysis of these drugs in patients with malignant bone lesions.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
We thank Karin Leotta and Renate Bangert for excellent technical assistance as well as Bram Stieltjes, Frederik Laun, and Dirk Simon for their support and valuable discussions about DWI techniques. We also thank Pfizer for providing SU-11248, Guerbet for supplying Sinerem, and the Deutsche Forschungsgemeinschaft (SFB-TR23) for their financial support.
Grant Support: Tobias Bäuerle and Dorde Komljenovic (Deutsche Forschungsgemeinschaft, SFB-TR23).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received November 3, 2009.
- Revision received March 2, 2010.
- Accepted April 22, 2010.