In order to accurately stage and treat malignancies, precise knowledge of tumor location, size, and lymphatic or distant spread is essential. While imaging with computed tomography (CT), magnetic resonance imaging, and ultrasound imaging provides valuable anatomical information, they lack the high sensitivity and specificity offered by a functional imaging modality such as positron emission tomography (PET). However, despite its ability to detect functional processes in the body, PET imaging has several drawbacks including the production, transportation, and cost of the radiopharmaceuticals which limit a wider spread of this technology. Also, PET images do not provide anatomical information, and are therefore inadequate for radiation or surgical treatment planning. Recent development of hybrid PET/CT scanners and sophisticated image registration algorithms allows for combined image sets from CT and PET to be used in the diagnosis and staging of malignant diseases. However, despite the benefits of combined PET/CT, the full potential of CT imaging cannot be realized because the superb spatial resolution provided by CT scans is not shared by the PET images. Current PET technology has reported limitations in detecting tumors of fewer than 109 cells (∼1 cm in diameter) [1]. This has significance for the early diagnosis of cancer, where small malignant lesions can be missed by PET scans.

In the last decade, studies have demonstrated that CT imaging can, when combined with an X-ray contrast agent, offer both anatomical and functional data [2]. However, this technique, termed functional CT, has not gained widespread clinical use because of the limitations of current contrast agents. To date, the most commonly used X-ray contrast agents are iodine-based compounds. Despite their clinical use, iodine-based contrast agents have several drawbacks including a high osmolality and a short blood half-life (<10 min) that requires imaging immediately after administration. Also, iodine has a moderate atomic number (Z) that limits the level of achievable CT contrast, decreasing its usefulness in radiation therapy planning, which relies almost exclusively on these values. More importantly, commercially available iodine-based X-ray contrast agents lack tumor-specific targeting ability. Conjugates with targeting moieties, such as antibodies, have thus far failed to deliver iodine to disease sites at a concentration detectable by current CT scanners. In addition to iodine-based agents, several other experimental X-ray contrast materials have been tested with varying degrees of success [3–6]. Nonetheless, the development of intravascular X-ray contrast agents based on other mid-Z to high-Z material, especially those agents with tumor-specific targeting capability, has not been successful due to performance, cost, and toxicity issues [7–9].

Recent advancements in nanotechnology have allowed for the development of novel contrast agents, including gold nanoparticles (AuNP), to improve functional CT imaging. Several investigators have demonstrated the feasibility of using various materials at the nanometer scale as X-ray contrast agents [10–15]. Due to its size, atomic number, and surface area to volume ratio, AuNP offers advantages over iodine-based compounds. Gold, because of its higher atomic number, attenuates X-rays more effectively than iodine and therefore produces superior contrast. Furthermore, the large molecular weight of the AuNPs, ranging from tens to hundreds kilodaltons, provides longer biological half-life than iodine-based compounds. Lastly, a unique property of the AuNPs that remains to be explored is their large surface-to-volume ratio. This large ratio differs distinctively from that of molecular or bulk materials and allows more freedom for surface modification.

In order to enhance their uptake by tumor cells, and thus increase the contrast in comparison to the surrounding normal tissue, we conjugated AuNPs with the glucose analog 2-deoxy-d-glucose (2-DG). We term the 2-DG-labeled gold nanoparticle AuNP-DG. We hypothesize that AuNP-DG will preserve the essential biochemical features of glucose, thus resulting in higher accumulation in tumor cells, and therefore may be used to trace glucose uptake via CT imaging. While the exact mechanism by which the AuNP-DG is internalized and retained in the tumor cells is the substance of future experiments, the purpose of this study is to evaluate the feasibility of using the AuNP-DG as a tumor-targeting CT contrast agent through in vitro experiments. To this end, we synthesized colloidal AuNP using a citrate acid reduction method similar to that reported previously [16, 17]. The AuNP suspension was repeatedly centrifuged until we obtained a concentration of 60 mg Au/mL. The solvent of the final product was changed from deionized water to phosphate buffered saline (PBS, pH 7.4). The prepared AuNPs were observed to be a dark-red-colored aqueous suspension with a mean particle size of 4 nm in diameter as determined using transmission electron microscopy (TEM, FEI Tecnai™ F30, FEI Co., USA). Fig. 1 shows a TEM picture of the unlabeled gold nanoparticles. The conjugation of the 2-DG with the AuNP core was accomplished using mercapto group in the 2-carbon position by condensation reaction of 2-amino-deoxyglucose with mercaptosuccinic acid [18]. The overall size of the AuNP-DG is estimated to be slightly larger than that of the unlabeled AuNP. However, because of the lack of an effective staining technique for the visualization of the 2-DG molecules, TEM images of the AuNP-DG were not acquired.

Fig. 1
figure 1

Transmission electron microscopy (TEM) image of the gold nanoparticles.

Full size image

As an example of the contrast-producing property of the gold nanoparticles, an AuNP sample at a concentration of 30 mg Au/mL was imaged using a microCT scanner (TRIUMPH™ X-O™ CT System, Gamma Medica-Ideas Inc., USA). A vial of water was also included in the CT scan serving as the reference. The images were acquired at 75 kVp, 135 μA with a 1,148 × 1,120 matrix size and 360 views, averaging five frames per view. The dataset was reconstructed into a 512 × 512 × 512 image volume. Image reconstruction was performed using a general-purpose filtered back-projection algorithm, implemented in the reconstruction software supplied with the imaging system. The reconstructed image data were transferred to a remote computer for further analysis. The reconstructed image data were analyzed using a commercial image processing software package (AMIRA™ 5.2, Visage Imaging Inc., USA). To assess the contrast enhancement, the CT signal intensity was expressed in Hounsfield units (HU) [19]. To make this conversion, a two-point calibration method was used. In this method, the CT signal intensity in the water volume was set to 0 HU, and the CT signal intensity in the air volume was set to −1,000 HU. The HU values of other materials were then obtained by linear extrapolation. Fig. 2a shows one coronal image from these microCT scans.

Fig. 2
figure 2

a A coronal slice from the microCT scans of AuNP (30 mg Au/ml) and water. The contrast between the AuNP and the water is clearly seen; b Histogram of voxel intensity expressed in Hounsfield units (HU) for the water and AuNP samples in (a).

Full size image

The contrast enhancement in the AuNP sample is clearly seen in this picture. The measured CT intensity of the AuNP sample is 900 ± 50 HU, consistent with the result shown by Kim et al. [14]. The voxel intensity histograms for both the water and the AuNP sample are shown in Fig. 2b.

The human alveolar epithelial cancer cell line, A-549, was chosen for the in vitro cellular uptake assay because of its positive response to fluorodeoxyglucose (FDG) [20]. Prior to the imaging experiments, trypan blue staining was performed on the A-549 cells after incubation with varying concentrations of AuNP-DG suspended in PBS and Dulbecco's Modified Eagle Medium (DMEM) media in order to assess cell membrane integrity as a marker for cell viability. These data were used to find non-toxic levels of AuNP-DG concentration in the growth media that would not disrupt the cell membrane, but still be high enough to achieve a suitable contrast level.

Approximately 1 × 105 A-549 cells, maintained in DMEM media supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin at 37°C in 7% CO2, were plated in 24-well cell culture plates 24 h prior to the experiment. Two groups of cell samples were prepared, each containing four wells. Twenty-four hours after plating, the A-549 cells were treated with AuNP and AuNP-DG, respectively, for 30 min (37°C, 7% CO2). Following the incubation, the cells were washed with sterile PBS six times to remove the excess gold nanoparticles. The cells were then transferred to plastic vials containing fresh growth media and spun to cell pellets using a centrifuge. The cell pellets were approximately 2 mm in diameter and adhere to the bottom of the vials in a half-moon shape. Fig. 3 sketches the cell preparation process.

Fig. 3
figure 3

Cell preparation flow chart.

Full size image

The cell samples were imaged using the microCT scanner immediately after the centrifugation. For image acquisition, the vials containing the cell pellets were stacked horizontally with their long axes parallel to the bed of the microCT scanner. The scanner settings used were identical to those used to image the pure AuNP suspension.

Fig. 4 shows the axial CT slices of the cell samples in both group-A (incubated with AuNP) and group-B (incubated with AuNP-DG). The intensity is expressed in HU as previously described. Because the cell pellets span more than one slice, only one representative slice is shown for each cell sample. In these images, the contrast enhancement in group-B over that in group-A is apparent.

Fig. 4
figure 4

Axial CT slices of two groups of cell samples. a Cells incubated with AuNP and b with AuNP-DG. The intensity is expressed in Hounsfield units (HU) with color-coding shown in the legend on the right. Because the cell pellets span more than one slice, only one representative slice from each cell sample is shown. The contrast enhancement in group-B is apparent. The dark blue annuli are the walls of the plastic vials. Trypan blue staining (not shown) was performed on cell samples after imaging in order to assess cell membrane integrity as a marker for cell viability.

Full size image

The CT contrast values (HU), measured in the center of the cell pellets as presented in Fig. 4, are shown in Fig. 5. As shown in Fig. 5, the CT contrast in the cells incubated with the AuNP-DG is, on average, more than three times higher than that of the cells incubated with the unlabeled AuNP. The CT contrast is directly proportional to the amount of the contrast agent taken up by the cells. Therefore, the higher contrast enhancement in the group-B samples strongly suggests enhanced uptake of the 2-DG-labeled gold nanoparticles over the unlabeled gold nanoparticles by the highly glycolytic A-549 cells. Both groups exhibit much higher CT contrast with respect to typical soft tissue (typical CT contrast for soft tissue ranges from 0 to 50 HU). We speculate that the enhanced contrast in the AuNP-incubated cell samples is due to insufficient wash-out or permeation of the AuNP through the cell membrane. However, the exact mechanism by which the unlabeled AuNP enhances the contrast in these in vitro assays remains to be investigated in future experiments.

Fig. 5
figure 5

CT contrast values (HU) of A-549 cancer cell samples incubated with AuNP (A1 to A4) and those incubated with AuNP-DG (B1 to B4).

Full size image

The A-549 cells used in these experiments are derived from human lung cancers where FDG/PET is commonly employed. Since the A-549 cells contain wild-type p53 and undergo cell cycle arrest following serum withdrawal, we chose to use complete cell growth media in our experiments that may have led to some competition with the glucose receptors. However, translation of these results to in vivo environment awaits direct experimental verification in the future. While the exact mechanism by which the unlabeled AuNP enhances the contrast in these in vitro assays remains to be investigated in future experiments, the cell membrane appeared intact in a majority of the cells after the AuNP-DG incubation in our cell viability assays.

In this study, we demonstrated, through in vitro experiments, that the addition of 2-DG to the AuNP enhances tumor cell uptake. This enhanced cellular uptake suggests that AuNP-DG could serve as a CT contrast agent with tumor-specific targeting capability. Validation of whether AuNP-DG is similar to 2-DG in its metabolism is currently being investigated in our laboratory. Uptake of 2-DG, similar to glucose, appears to depend on both the glucose transport proteins and the hexokinase enzyme. While the former allows intracellular uptake of glucose and its analogs, the latter ensures the trapping of these substances inside the cell by phosphorylating their 6th carbon. Both events lead to the intracellular accumulation of 2-DG-6-phosphate which is limited in its further metabolism resulting in accumulation of this metabolite. The upregulation of either glucose transporters or hexokinases in cancer cells may independently play a role in the greater accumulation of 2-DG [21]. Thus, it is important to investigate whether addition of AuNP to 2-DG will interfere with either of these enzymes. A priori, we hypothesize that AuNP-DG should be an active substrate for glucose transporters since previously it was demonstrated that replacement of the hydrogen group on the second carbon of 2-DG with larger moieties results in minimal steric hindrance [22]. On the other hand, the glucose binding pocket of hexokinase appears to be more conservative to the size of the molecule on the second carbon of 2-DG [23]. Therefore, the nanoparticle may interfere with the binding of AuNP-DG to hexokinase. However, glucose analogs that cannot be phosphorylated by hexokinase, for instance, 3-O-methyl-glucose and 6-fluoro-deoxy-d-glucose, have been shown to be a tracer of greater glucose uptake in cells that over-express glucose transporters [24]. Thus, even if the AuNP-DG is a substrate of only the glucose transporter but not hexokinase, it may still be sufficient to employ this agent for functionally targeted CT imaging since it may exploit high glucose uptake of tumor cells. Overall, our results in which the cellular AuNP-DG signal is found to be superior to AuNP alone suggest that the 2-DG moiety on the nanoparticle preserves some of its activity in facilitating the uptake of this molecule into tumor cells.

Functional CT technique offers a new set of capabilities in cancer imaging by providing unmatched high-resolution anatomical and functional images in a single CT scan. Higher-resolution functional and molecular image information can be expected to be one of the most important contributors in improving cancer therapy, since early detection of cancer still offers the best prognosis [25]. Since X-ray CT and AuNP-DG technologies could be widespread and readily available, successful clinical implementation of the AuNP-DG will offer an effective functional contrast agent for high-resolution X-ray and CT imaging that could have a large impact on cancer research, diagnosis, and treatment planning. Results from these experiments strongly suggest enhanced uptake of the AuNP-DG over the unlabeled AuNP by the highly glycolytic cancer cells in vitro and indicate that AuNP-DG could serve as a CT contrast agent with tumor targeting capability. The exact mechanism by which the AuNP-DG is internalized and retained in the tumor cells remains to be studied in future experiments.