Synthesis and characterization of intrinsically radiolabeled quantum dots for bimodal detection

Abstract

A novel approach was developed to synthesize radioactive quantum dots (r-QDs) thereby enabling both optical and radionuclide signals to be detected from the same intrinsic bimodal probe. This proof-of-concept is exemplified by the incorporation of the radionuclide ¹⁰⁹Cadmium into the core/shell of the nanoparticle. Green and near infrared (NIR) emission intrinsic r-QDs were synthesized and characterized. Zwitterionic and Poly-polyethlene glycol (PEGylated) ligands were synthesized and used to coat r-QDs. Zwitterionic NIR r-QDs (quantum yield = 11%) and PEGylated NIR r-QDs (quantum yield = 14%) with an average size of 13.8 nm and 16.8 nm were obtained respectively. The biodistribution of NIR zwitterionic and PEGylated r-QDs in nude mice was investigated and zwitterionic r-QDs showed longer blood circulation (t_1/2 = 21.4±1.1 hrs) than their PEGylated counterparts (t_1/2 = 6.4±0.5 min). Both zwitterionic and PEGylated r-QDs exhibited progressive accumulation in the liver and spleen, but the magnitude of the accumulation (%ID/g) was about 3-6 fold higher with the PEGylated r-QDs at all the time points. The results demonstrated the feasibility of r-QDs synthesis in quantitative yield and retention of fluorescence following incorporation of radioactivity into the core/shell of the nanoparticle. The gamma signal from the same fluorescent elemental material enabled quantitative and robust pharmacokinetic measurements and how these changed depended on the type of coating ligands used. This strategy for intrinsically radio-labeling the QDs is currently being implemented in our laboratory for the incorporation of other radiometals.

Introduction

Multi-modality molecular imaging technologies continue to be developed since the successful impact made by clinical modalities such as PET/CT and SPECT/CT, where the strength of each modality has been exploited in the hybrid system. The combination of high resolution anatomical CT and the high sensitivity, coupled with the functional and molecular capability of PET and SPECT has significantly enhanced diagnostic sensitivity and specificity compared to the individual modality [1-4]. Such advances have stimulated further research in other combinations such as hybrid PET/MR [1,3]. These have culminated in the recent FDA approval of a truly PET/MR clinical system. These advancements have stimulated a number of studies on the development of hybrid molecular or nano-scale probes combining two or more emission properties such as magnetization, optical, or gamma radiation [5-9]. Hybrid probes combining fluorescence and gamma signals would also enhance the quantitative capability of fluorescence-guided surgery and ultimately result in more efficient resection of small residual tumor lesions [10]. Herein, we describe a method for synthesizing hybrid quantum dot nanoparticles that combine fluorescence and gamma signals.

Quantum dots (QDs) have attracted attention as fluorescent nano-probes in biomedical imaging because of their unique properties of broad absorption, narrow emission, tunable emissive wavelength and excellent resistance to light-bleaching [11-16]. When used by itself, fluorescent imaging however has low spatial resolution and limited depth, but provides high sensitivity in the nano- to pico-molar range. Instead of using two different agents, (one for each modality that could be distributed differently in the body), a more robust way is to fuse the signaling moieties for the two imaging techniques into one molecule (monomolecular multimodality imaging agents or MOMIAs) [17-21]. This unique structural feature ensures that both signals emanate from the same source allowing for the fusion of contrast data with high precision [1].

Recent studies have exploited an extrinsic chelator-label approach to make radio-labeled QDs (Figure 1A) for combined PET/optical or SPECT/optical imaging in vivo [22-25]. Introduction of a chelator on the surface of nanoparticles, however, could potentially influence one or more pharmacokinetic component, such as absorption, distribution, metabolism and excretion. Few investigations were reported for making intrinsically radioactive QDs (r-QDs) (Figure 1B and 1C) and evaluated their potential applications for multi-modality imaging [26,27]. Compared with extrinsic radio-chelate labeled QDs, the intrinsic r-QDs are more conveniently synthesized, more stable in vivo and minimize change in the pharmacokinetics of nanoparticles, which could be induced by surface modification necessary for chelation of radio-chelate QDs. Furthermore, the intrinsic approach enables, a variety of radionuclides which can be incorporated into the quantum dots including those of copper, indium, zinc, selenium, tellurium amongst others. This intrinsic r-QDs technique provides a potential for hybrid nanoprobes for molecular imaging.

Figure 1.

Schematics of approaches to radioactive quantum dots: extrinsic radio-chelate labeled QDs (A); intrinsic radio-core/shell r-QDs (B); and core/radio-shell r-QDs (C).

Cadmium-based QDs are most developed and used for fluorescent biological imaging, despite their potential in vivo toxicity. In this work, we tested and validated the intrinsic radiolabeling proof-of-concept by synthesizing and evaluating r-QDs using the radionuclide cadmium-109 (E_γ = 88 keV, K-X-rays = 22 keV). A general and highly reproducible strategy was developed to synthesize intrinsic r-QDs in a high radiochemical yield (Figure 1C). By designing and adjusting the surface coating of quantum dots, compact zwitterionic or PEGylated r-QDs were obtained. Biodistribution studies showed distinct pharmacokinetic profile of the two coated r-QDs. Bimodal detection of fluorescent and gamma signals was demonstrated using the r-QDs. Cadmium-free r-QDs, with short and medium half-life radioisotopes, are currently under development in our laboratory for in vivo multi-modality imaging.

Materials and methods

Chemicals

All chemicals were used as received without further purification. Cadmium oxide (99.99%), Selenium (99.99%), Trioctylphosphine oxide (TOPO, 99%), Trioctylphosphine (TOP, 90%), Octadecene (ODE, 90%), Hexadecylamine (HDA, 99%), Oleylamine (OA, tech grade), Trioctylamine (TOA, 98%), Zinc oxide (99%), Zinc stearate, Hexamethyldisilthiane (synthetic grade), Lipoic acid (95%), Carbonyldiimidazole (CDI, reagent grade), N,N-dimethylethylenediamine (98%), 1,3-propane sultone (95%), Sodium borohydride (96%), Rhodamine 6G, 1,1’,3,3,3’,3’-hexamethylindotricarbocyanine iodide (HITCI, 97%), Poly(acrylic acid) (PAA, M_w:1,800) were purchased from Sigma-Aldrich. Oleic acid (90%), and Tellurium (99.99) were obtained from Alfa Aesar. Tetradecylphosphonic acid (TDPA, 99%) was purchased from PCI Synthesis. Cadmium oleate in oleic acid (0.5 M) and Zinc oleate in oleic acid (0.5M) were prepared according to published methods [28]. ¹⁰⁹CdCl₂ aqueous solution (specific activity: 95,534 GBq/g) was purchased from Oak Ridge National Laboratory.

Synthesis of radioactive quantum dots

Synthesis of CdSe/¹⁰⁹CdZnS r-QDs: CdTeSe and CdSe were synthesized according to literature [28,29] with minor modification. Synthesis of CdSe/¹⁰⁹CdZnS was followed by surface ion layer adsorption and reaction (SILAR) procedure with introduction of Cd-109 nuclide. Briefly, ¹⁰⁹CdCl₂ (200 μCi) aqueous solution was loaded into a 25 mL three-neck flask. Water was removed under heating and gentle argon flow. TOPO (7.76 mmole), HDA (6.20 mmole) and the solution of CdSe (0.3 μmole) in chloroform were added. The chloroform was removed under vacuum. Zinc stearate (0.5 mmole) and Hexamethyldisilthiane (0.5 mmole) were dissolved in TOP (3 mL) separately. The zinc solution (0.5 mL) was added dropwise into the reaction mixture and then kept the reaction for 10 min at 160°C. Hexamethyldisilthiane solution (0.5 mL) was added dropwise subsequently and reacted for 10 min at the same temperature. The addition of two reactants was repeated once with 1.0 mL and 1.5 mL respectively. Finally the reaction mixture was kept for half an hour at 160°C and then was cooled to room temperature (r.t). Chloroform was added to form a clear solution at r.t. Methanol was added to precipitate the CdSe/¹⁰⁹CdZnS. After centrifugation at 6,000 rpm for 5 min, 5 mL chloroform was used to dissolve the pellet and the solution was filtered by 0.2 μm nylon filter (radioactive reaction yield: 98%, specific activity: 653 mCi/mmole).

Synthesis of CdTeSe/¹⁰⁹CdZnSe r-QDs: Synthesis of CdTeSe/¹⁰⁹CdZnSe was followed by thermal cycling procedure with the introduction of Cd-109 radionuclide. Briefly, ¹⁰⁹CdCl₂ (200 μCi) aqueous solution was loaded in a 25 mL three-neck flask. Water was removed under heating and gentle argon flow. TOA (5 mL) and CdTeSe (0.12 μmole) solution in chloroform (5 mL) were loaded into the flask. Chloroform was removed under vacuum. TOP (0.5 mL), solution of Cd (oleate)₂ in oleic acid (0.5M, 0.3 mL) and Zn (oleate)₂ in oleic acid (0.5M, 0.7 mL) were added into the reaction mixture. The reaction mixture was heated to 80°C for half an hour under vacuum and then heated up to 230°C. The selenium in TOP (0.5M, 0.3 mL) was added dropwise into the reaction within a few minutes. Temperature of reaction was raised up to 280° C very quickly and then cooled down to 230°C again. Selenium solution in TOP (0.5 M, 0.7 mL) was once again added dropwise to the reaction mixture and the same thermo-cycling procedure was followed. CdTeSe/¹⁰⁹CdZnSe was obtained following similar purification procedure described above (radioactive reaction yield: 96%, specific activity: 1.6 Ci/mmole).

Synthesis of zwitterionic ligand (3) and PEGylated ligand (8)

To a lipoic acid solution (2 g, 9.7 mmole) in chloroform (10 mL) was added CDI (1.88 g, 11.6 mmole) and stirred for 30 min under argon. The final solution was added dropwise to the solution of N, N-dimethylethylenediamine (1.28 mL, 11.6 mmole) in chloroform (20 mL) at 4°C under argon. The reaction mixture was stirred overnight at r.t. The product was washed with brine (15 mL×3) followed by aqueous NaOH solution (15 mL×3, 10 mM). The organic phase was dried over anhydrous Na₂SO₄. Removal of organic solvent yields 1 as a yellow oil (2.4 g, 90%). ¹H NMR (300 MHz, [D₁]CHCl₃, 25°C, TMS): δ = 5.98 (s, 1H), 3.57 (m, 1H), 3.34-3.28 (t, 2H), 3.21-3.06 (m, 2H), 2.50-2.37 (m, 3H), 2.22 (s, 6H), 2.18 (t, 2H), 1.96-1.85 (m, 1H), 1.74-1.62 (m, 4H), 1.52-1.43 (m, 2H) ppm. ¹³C NMR (75 MHz): δ = 58.06, 56.63, 45.30, 40.42, 38.65, 36.84, 36.55, 34.82, 29.09, 25.60 ppm.

To a solution of 1 (0.51 g, 1.85 mmole) in chloroform (10 mL) was added 1,3-propane sultone (0.236 g, 1.93 mmole). The reaction was kept at 60°C under argon and stirred overnight. The yellow precipitate was filtered and washed by chloroform (15 mL×3). The product was dried under vacuum to give 2 as a yellow solid (0.56 g, 76%). ¹H NMR (300 MHz, [D₂]H₂O, 25°C): δ = 3.70 (t, 2H), 3.62-3.49 (m, 5H), 3.23 (m, 2H), 3.19 (s, 6H), 3.00 (t, 2H), 2.50 (m, 1H), 2.34-2.24 (m, 4H), 2.03 (m, 1H), 1.79-1.60 (m, 4H), 1.49-1.41 (t, 2H), 1.20 (t, 1H) ppm. ¹³C NMR (75 MHz): δ = 177.08, 63.00, 61.85, 56.74, 51.15, 49.00, 47.38, 40.53, 38.45, 35.53, 34.06, 33.21, 28.29, 25.04, 18.40 ppm.

To a solution of 2 (0.56 g, 1.44 mmole) in a mixture of water (12 mL) and ethanol (3 mL) was added NaBH₄ (62 mg, 1.68 mmole) aliquot at 4°C. The reaction was gradually warmed to r.t and kept for 4 hrs. After removal of the solvent, the product was obtained as a white solid 3 without further purification. 20 mL water was added to dissolve 3 and the solution was used for ligand exchange reaction directly. ¹H NMR (300 MHz, [D₂]H₂O, 25°C): δ = 3.72 (t, 2H), 3.59-3.50 (m, 4H), 3.20 (s, 6H), 3.02 (m, 3H), 2.67 (m, 2H), 2.40-2.25 (m, 5H), 1.90-1.47 (m, 10H) ppm.

A solution of Poly(acrylic acid) (PAA) (0.5 g, 0.27 mmole) in THF (10 mL) was added dropwise to the solution of CDI (1.24 g, 7.65 mmole) in THF (10 mL) under argon at r.t. The reaction was stirred for 12 hrs at r.t to form the activated polymer 4. The above solution of 4 was added dropwise to a mixture of 5 (0.86 g, 3.5 mmol) and 6 (1.9 g, 3.5 mmol) in chloroform (20 mL) under argon at 4°C. The reaction solution was stirred overnight at r.t. The product was washed by brine (15 mL×3), and the aqueous phase was extracted by chloroform (15 mL×3). The combined organic phases were dried over anhydrous Na₂SO₄. The crude product was purified by size exclusion chromatography (Bio-Beads SX1 gel, 200-400 mesh, BIO-RAD) with THF as the mobile phase to give 7 as a yellow oil (1.44 g, 44%). ¹H NMR (300 MHz, [D₁]CHCl₃, 25°C, TMS): δ = 5.48 (broad), 4.33-4.28 (broad), 3.79-3.49 (broad), 3.33 (broad), 2.48-2.43 (broad), 2.25-2.17 (broad), 1.91-1.78 (broad), 1.61-1.18 (broad) ppm.

To a solution of 7 (0.94 g, 0.083 mmole) in ethanol (15 mL), was added sodium borohydride (57.5 mg, 1.55 mmole) aliquot at 4°C. Reaction was stirred at r.t for 4 hrs. The solvent was removed under vacuum. 15 mL water was added to dissolve the residue 8 and the solution was filtered. The solution of 8 in H₂O was used directly for ligand exchange reaction. ¹H NMR (300 MHz, [D₂]H₂O, 25°C): δ = 3.69-3.55 (broad), 3.44-3.37 (broad), 2.95 (broad), 2.65 (broad), 2.24-2.19 (broad), 1.82-1.28 (broad) ppm.

Ligand exchange reactions

The solution of r-QDs (6 nmole) in chloroform (3 mL) was mixed with a solution of zwitterionic ligand (3) (140 mmole) in water (2 mL). The reaction mixture was stirred vigorously under 60°C and the r-QDs were transferred from the organic phase to the aqueous phase. After quiescence and cooling to r.t, the r-QDs aqueous solution was filtered through 0.2 μm nylon filter. The filtered solution was centrifuged to separate excess ligands and concentrated over Amicon filters (30,000 MWCO) at 10,000 rpm for 30 min.

The solution of r-QDs (6 nmole) in chloroform (2 mL) was mixed with excess of PEGylated ligand (8) (5 mmole). Reaction mixture was stirred vigorously under 45°C for 30 min. Reaction solution became turbid when ligand exchange completed. PEGylated coated r-QDs were precipitated by n-hexane. After centrifugation at 3,000 rpm for 5 min, the pellet was collected and dried, then dissolved in water (2 mL). The r-QDs aqueous solution was filtered through 0.2 μm nylon filter. The filtered solution was centrifuged to separate excess ligands and concentrated over Amicon filters (30,000 MWCO) at 10,000 rpm for 30 min.

Characterization

TEM images were recorded on JEOL JEM-1230 transmission electron microscope operating at an accelerating voltage of 120 KV. The samples were prepared by dropping a diluted solution of r-QDs in toluene on the carbon film supported copper grids (Formvar/Carbon 300 Mesh Cu). Zeta Sizer Nano Series ZEN3600 was used to measure the hydrodynamic size of r-QDs in water. UV-vis absorption and fluorescent emission spectra of r-QDs in solution were recorded by Hewlett Packard 8453 and Varian Cary eclipse spectrophotometer respectively. Quantum yield (QY) of green r-QDs was measured relative to Rhodamine 6G (QY is 95% in ethanol) under excitation of 490 nm. QY of NIR r-QDs was measured relative to HITCI (QY is 26% in methanol) under excitation of 630 nm. Solutions of r-QDs in water and dye in relative solvents were optically matched at the excitation wavelength. The UV-vis absorption density was kept below 0.1 between 300-700 nm. The integration of emission spectra, corrected for differences in index of refraction and concentration, were used to calculate the QYs using following expression: QY_QD = QY_Dye × (Absorbance_dye/Absorbance_QD) × (Peak Area_QD/Peak Area_dye) × (n_QD-solvent)²/(n_dye-solvent)². The radioactive reaction yield was calculated from gamma counting. Gamma counter (LKB Wallac 1282 compugamma CS universal gamma counter/Perkin Elmer) was calibrated for ¹⁰⁹Cd using an energy window centered on 88 keV. ¹H and ¹³C NMR spectra were recorded on varian Mercury 300 spectrometer with solvent proton resonance as reference.

Bimodal detection of optical and radioactive emissions

To demonstrate bimodal detection of both fluorescence and radioactive emission from the r-QDs, capillary tubes were first loaded with 10 μL of 0.7 μM of QDs (CdSe/ZnS) and r-QDs (CdSe/¹⁰⁹CdZnS) in water suspension and the ends of the tubes were sealed to prevent evaporation. Tubes were then imaged using Maestro-2 multispectral imaging system (Cambridge Research and Instrumentation, Inc., Woburn, MA) to detect the fluorescence signal (emission peak = 570 nm). Following this, the tubes were positioned over a photostimulable phosphor imaging screen (Fuji Medical Systems, USA) and were exposed overnight in a light proof cassette to capture the radioactive emission of the ¹⁰⁹Cd (Gamma and X-rays; 88 keV). The imaging screen was then scanned using a BAS-5000 phosphorimager (Fuji Medical Systems, USA) and the radioactive emission was imaged.

Cell viability

Mouse mammary carcinoma (MMC) cells established from a spontaneous tumor harvested from FVBN202 mice, as previously described [30], were used for the initial evaluations in a cell line of mouse origin before moving onto the in vivo bio-distribution studies in nude mice. These MMC cells, cultured in RPMI 1640 supplemented with 10% fetal bovine serum, were plated in quadruplets in 96-well plates (30000/well) and were allowed to adhere to the plates overnight. They were then incubated for 24 hrs in the presence of various concentrations of either zwitterionic r-QDs or PEGylated r-QDs. At the end of the exposure period, the cell viability was enumerated using a luminescent assay (CellTiter-Glo, Promega) for the quantitation of ATP content, which signals the presence of metabolically active cells. In this homogeneous method, addition of assay reagents results in cell lysis and generation of a luminescent signal directly proportional to the amount of ATP present. The amount of ATP is directly proportional to the number of metabolically active cells present in culture. Similar assay was also done in human cell lines such as Colo-205 and Jurkat cells (ATCC) incubated with r-QDs. These were grown in media and conditions recommended by ATCC.

Biodistribution

Animal experiments were approved and performed according to the policies and guidelines of the Animal Care and Use Committee (IACUC) at Virginia Commonwealth University. Adult (8-week-old) female nude mice (Harlan) were injected with the solution of zwitterionic or PEGylated ligands coated r-QDs in saline (0.5-1 μCi; 25-50 pmole in 200 μL) through tail vein. Mice were euthanized, blood samples and other major organs were collected at different times after injection (5 min, 30 min, 1 hr, 6 hrs, 24 hrs, 48 hrs, 96 hrs and 168 hrs; n=3 per time point). We measured radioactivity of each sample by gamma counter. The percentage of the injected dose per gram (%ID/g) of tissues was calculated according the results of gamma counting. The blood activity data was used to calculate the blood half-life and uncertainty. Mice bearing Colo-205 (ATTC) xenograft (5x10⁶ cells injected about 14 days earlier) were used for the biodistribution study of zwitterionic coated r-QDs. In order to demonstrate the avid uptake of the PEGylated r-QDs by the liver and spleen, non-tumor bearing normal mice were used.

In vivo and ex vivo fluorescence imaging

Nude mice were anaesthetized with isoflurane (2% in oxygen) and injected intravenously via the tail vein with zwitterionic CdTeSe/¹⁰⁹CdZnSe or PEGylated CdTeSe/¹⁰⁹CdZnSe at a dosage of 600 pmole or 300 pmole, respectively. One hour post-injection of the r-QDs, NIR fluorescent multispectral images of these mice were obtained using the Maestro-2 in vivo Imaging System (Cambridge Research and Instrumentation, Inc., Woburn, MA). A combination of excitation (684 to 729 nm band-pass) and emission (745 nm long-pass) filters with the acquisition settings of 740-950 in 10 nm increments was used to capture the NIR emission from the r-QDs. The resultant multispectral images were unmixed into their component spectra (r-QDs, autofluorescence, and background) using the in-built software. Following in vivo imaging, mice were sacrificed and dissected samples of the liver, spleen, heart, lungs, intestine, kidney, stomach, muscle, bone, skin, brain and femur were washed in normal saline, drained and ex vivo imaging was done with the same acquisition and unmixing settings.

Results

Radioactive quantum dots synthesis

Two methods were used to synthesize intrinsically radioactive core/shell QDs. The schematics of these two approaches are shown in Figure 1B and 1C, in which the orange color represents the incorporation of the radionuclide. In order to prevent leakage of radioactivity from the r-QDs, the radionuclide needs to be incorporated in the core or on the interface layer between core and shell. In this manner, the radionuclide is well protected by the non-radioactive shell. In the first method (Figure 1B), the radionuclide is introduced during the synthesis of the quantum dots core. ¹⁰⁹CdSe or ¹⁰⁹CdTeSe r-QDs were synthesized by doping a tracer amount of ¹⁰⁹Cd in the reaction of Cd(TDPA)₂/Se/Te following traditional organometallic synthetic strategy. Results from gamma counting gave a product radiochemical yield of 40-50%, which was limited by the reaction efficiency of formation of quantum dots. The second method (Figure 1C) was designed to synthesize the CdSe/¹⁰⁹CdZnS in order to improve the radiochemical yield and specific activity of r-QDs. The ¹⁰⁹Cd was introduced in the second step (shell synthesis step) by SILAR method. This method afforded a quantitative radiochemical yield approaching 100% incorporation in many preparations and has also proven effective for depositing different radionuclides on the surface of quantum dots. In the synthesis of CdTeSe/¹⁰⁹CdZnSe r-QDs, thermal cycling and SILAR methods were combined and used to produce ¹⁰⁹CdZnSe dual-layer shell. The morphology and fluorescent quantum yield of synthesized NIR r-QDs were obviously improved. Spherical r-QDs instead of peanut-shaped r-QDs (data not shown) were obtained after thermal cycling process. And the fluorescent quantum yield of NIR r-QDs in chloroform enhanced from 24% to 40%. Dual-layer cold shell also provides good protection to the radioactive layer. The size and distribution of synthesized r-QDs were characterized by TEM. The morphology and uniform size distribution of r-QDs are similar to those of non-radioactive QDs. The size of green r-QDs, as measured by TEM is around 4 nm (Figure 3A) and that of NIR r-QD is 6 nm (Figure 3B).

Figure 3.

Transmission electron microscopy (TEM) images and size distribution by dynamic light scattering (DLS): Green emission r-QDs (A) and NIR emission r-QDs (B). DLS of zwitterionic green emission r-QDs (C); zwitterionic NIR emission r-QDs (D); PEGylated green emission r-QDs (E) and PEGylated NIR emission r-QDs (F) in H₂O.

Ligands design and synthesis

Recently, small zwitterionic ligands such as cysteine or DHLA-based anion/cation counterpart were proposed to coat QDs and the final hydrophilic QDs showed compact hydrodynamic size, good stability in aqueous solution and reduced nonspecific binding to cell membranes [31,32]. Here a zwitterionic ligand dihydrolipoic acid-sulfobetaine (DHLA-SB) (3) was designed and synthesized within three steps conveniently (Figure 2). First, lipoic acid was coupled with N,N-dimethylethylenediamine under the activation of CDI to give compound 1. Secondly, tertiary amine of 1 was sulfonated to form zwitterionic ligand directly through the ring-opening reaction of 1,3-propane to give precursor 2. Due to the high reactivity of 1.3-propane, the reaction can be performed under mild condition (60°C) with high reaction yield. Finally, the 1,2-dithiolane group of lipoic acid was reduced to dihydrolipoic acid (DHLA) by sodium borohydride to achieve zwitterionic ligand 3. The whole synthetic procedure is economic and simple and the reaction yield is reasonably high. An additional advantage of this method is that it allows for flexibility in the synthesis of other related zwitterionic ligands with different counter-anions.

Figure 2.

Schematics of ligand synthesis used to coat the r-QDs. Zwitterionic ligand DHLA-SB (3), and multidentate PEGylated ligand PAA-DHLA-PEG (8). Reaction conditions: (i) CDI, N, N-dimethylethylene-diamine, chloroform, 4°C to r.t, overnight; (ii) 1,3-propane sultone, chloroform, 60° C, overnight; (iii) NaBH₄, ethanol/H₂O, 4°C to r.t in 4 hrs; (iv) CDI, chloroform, r.t, overnight; (v) 5, 6, THF/chloroform, r.t, overnight; (vi) NaBH₄, ethanol, 4°C to r.t in 4 hrs.

PEGylation has been demonstrated as a general strategy to make water soluble nanoparticles and improve their biocompatibility as well as reduce nonspecific binding with serum proteins [33-35]. In comparison to small DHLA-PEG ligands coated QDs, our PEGylated based r-QDs provided better stability in aqueous solution over a period of three months (data not shown). The synthesis of poly-PEG ligand PAA-DHLA-PEG (8) (Figure 2) is based on introducing lipoic acid and PEG (M_w: 550) into the PAA main chain followed by reduction of lipoic acid to DHLA, a multi-dentate PEGylated polymer ligand was achieved. The short length of the PEG chain afforded a compact hydrodynamic size of PAA-DHLA-PEG coated r-QDs that is comparable to that of the zwitterionic ligand coated r-QDs. The ligand exchange reaction of r-QDs with PAA-DHLA-PEG was performed in chloroform based on the methods described in the literature [29].

Ligand exchange, hydrodynamic sizes, fluorescent quantum yields and autoradiography of water soluble r-QDs

Using the DHLA-SB (3) and PAA-DHLA-PEG (8), it was feasible to achieve ligand exchange under mild conditions due to strong interaction of DHLA with surface of QDs. The r-QDs were transferred from the organic phase to aqueous phase within 30 min, resulting in tight ligation water soluble r-QDs. Following ligand exchange, there was a slight bathochromic shift relative to the emission of hydrophobic r-QDs, which is consistent with previous reports on ligand exchange with thioalkyl acids [36]. The hydrodynamic sizes of water-soluble green (Figure 3C) and NIR (Figure 3D) emission zwitterionic r-QDs are around 10 nm and 13.8 nm respectively, which are relatively compact due to the small size of zwitterionic ligand. The hydrodynamic sizes of PAA-DHLA-PEG coated r-QDs were in the range of 15-17 nm for the green (Figure 3E) and NIR (Figure 3F) r-QDs. The stability of zwitterionic ligand coated r-QDs was investigated as a function of pH ranging from 3 to 11. The r-QDs show excellent stability in neutral and basic conditions from pH 7 to 11 over 3 months (data not shown). However a precipitate was observed at pH 5 after one week and even earlier at pH 3. Multi-dentate PEGylated r-QDs also showed high stability over 3 months when stored under ambient conditions (room temperature and pH 7-9).

Fluorescence spectra of green and NIR r-QDs are shown in Figure 4A, emitting at about 570 and 780 nm respectively. The quantum yields of green and NIR r-QDs (90% and 40% respectively in chloroform) are comparable with the standard Rhodamine 6G and the NIR standard HITCI. After the zwitterionic ligand exchange, the quantum yields of water soluble green and NIR r-QDs were reduced to 25% and 11% respectively. These values are comparable to the quantum yield values of non-radioactive QDs and similar results were reported previously [28,36]. PAA-DHLA-PEG ligand coated green and NIR r-QDs also gave similar fluorescence emission spectra and quantum yields to the zwitterionic r-QDs (23% and 14%). Bimodal detection of the fluorescence from these r-QDs along with the autoradiographic radionuclide signal were demonstrated in zwitterionic coated green r-QDs samples as shown in Figure 4B, therefore providing proof-of-concept for dual modality.

Figure 4.

Fluorescence emission spectra of (A): □ CdSe in CHCl₃ (1), ◊ CdSe/¹⁰⁹CdZnS in CHCl₃ (2), ▲ zwitterionic CdSe/¹⁰⁹CdZnS in H₂O (3) and ▼ CdTeSe in CHCl₃ (4), ◄ CdTeSe/¹⁰⁹CdZnSe in CHCl₃ (5), ► zwitterionic CdTeSe/¹⁰⁹CdZnSe in H₂O (6). Images of r-QDs in capillary tubes (B): bright field image (top row), fluorescent signal (middle row) and the radioactive gamma signal, imaged by autoradiography (bottom row). Left capillary tube is for CdSe/ZnS and the right capillary tube is for CdSe/¹⁰⁹CdZnS.

Effect of r-QDs on cell viability

As shown in Figure 5, no significant decrease in viability of MMC cells was observed for concentrations up to 20 nM for both zwitterionic and PEGylated r-QDs. Significant drop in cell viability was seen for higher concentrations with 20% and 40% non-viable cells observed at 100 nM of zwitterionic and PEGylated r-QDs respectively. Similar results were obtained with Colo-205 and Jurkat cell lines (data not shown).

Figure 5.

In vitro viability assay of MMC cell lines, obtained by using luminescent based ATP assay. No toxicity was observed for either r-QDs, up to 20 nM concentration (*= p<0.05) in comparison to 0 nM.

Biodistribution and fluorescence imaging of r-QDs

The %ID/g values of the NIR r-QDs in blood and other major organs collected from nude mice at different time points after tail vein injection are given in Table 1 and Figure 6A for the Zwitterionic NIR r-QDs and in Table 2 and Figure 6B for the PEGylated NIR r-QDs. The zwitterionic r-QDs showed relatively much longer dwell time in the blood (t_1/2 = 21.4±1.1 hrs) than the PEGylated r-QDs (t_1/2 = 6.4±0.5 min). Similar blood half-life of PEGylated QDs has also been reported by others [34,37]. Both zwitterionic and PEGylated r-QDs exhibited progressive accumulation in the liver and spleen, but the magnitude of the uptake was higher with the PEGylated r-QDs at all the time points. The accumulation of zwitterionic r-QDs was found in liver (16.49±1.7 %ID/g) and spleen (5.49±1.2 %ID/g) at 24 hrs post-injection, then slowly increased to 19.63±2.5 % ID/g and 7.17±0.7 %ID/g respectively at 48 hrs post-injection. In comparison, the uptake of PEGylated r-QDs in the liver and spleen was rapid and about 3-6 fold higher at all time-points. For instance, the maximum accumulation of PEGylated r-QDs in the liver (74.88±1.6 %ID/g) and spleen (27.96±0.1 %ID/g) was reached within 6 hrs post-injection. Also, marked uptake of zwitterionic r-QDs in skin (up to 14.21±1.4 %ID/g at 96 hrs) was observed whilst the PEGylated r-QDs showed only marginal accumulation (0.32±0.1 %ID/g) in skin at 24 hrs post-injection. The uptake of zwitterionic r-QDs in femur (2.28±0.3 %ID/g at 24 hrs post-injection) was found less than PEGylated r-QDs (8.58±1.7 %ID/g at 24 hrs post-injection).

Table 1.

Biodistribution of i.v. injected zwitterionic NIR r-QDs in nude mice (n=3 per time point). Data are presented as %ID/g (mean ± sem) values determined by gamma counting

Organs	5 min.	30 min.	1 hr	6 hrs	24 hrs	48 hrs	96 hrs	168 hrs
Blood	28.7±2.0	26.0±3.2	14.2±3.0	19.6±0.2	8.8±1.4	9.1±1.2	2.7±0.3	0.1±0.0
Heart	5.3±0.4	4.3±1.3	4.1±1.5	4.1±0.6	3.4±0.5	3.4±0.4	1.4±0.3	0.6±0.1
Lungs	7.5±1.2	7.1±1.1	4.9±0.7	6.4±0.5	3.0±0.6	3.7±0.4	5.3±0.4	0.5±0.0
Liver	10.5±1.5	6.2±0.2	20.1±1.6	13.8±0.5	16.5±1.7	19.6±2.5	19.4±2.5	31.9±3.1
Spleen	4.2±1.1	3.0±0.3	6.0±1.6	4.4±0.1	5.5±1.2	7.2±0.7	6.6±0.7	7.1±0.8
Stomach	0.8±0.1	0.5±0.1	0.4±0.1	0.7±0.2	0.8±0.3	1.0±0.3	0.9±0.3	0.4±0.1
Intestine	1.2±0.3	1.2±0.2	1.0±0.3	1.3±0.3	1.3±0.3	1.6±0.6	0.9±0.6	0.3±0.1
Kidneys	4.0±0.6	3.9±0.4	2.9±0.3	4.1±0.1	1.7±0.1	2.9±0.4	3.7±0.4	3.2±0.2
Skin	4.6±0.3	5.7±0.9	6.5±0.4	8.3±0.4	6.7±0.5	11.5±1.4	14.2±1.4	9.0±0.7
Muscle	0.3±0.2	2.6±0.9	2.1±0.7	1.8±0.2	1.3±0.2	1.8±0.4	1.8±0.4	1.6±0.4
Skull	3.1±0.5	3.4±0.5	2.8±0.9	3.8±0.0	4.0±1.6	2.7±0.2	1.3±0.2	3.6±2.7
Brain	0.8±0.2	0.6±0.1	0.9±0.5	0.7±0.2	0.4±0.0	0.3±0.1	0.1±0.1	0.01±0.0
Femur	1.9±0.6	1.4±0.2	2.6±1.5	2.3±0.5	2.3±0.3	2.3±0.5	2.7±0.5	2.6±0.6
Tumor	0.7±0.1	1.1±0.1	0.6±0.2	2.0±0.3	3.1±0.6	5.1±0.8	4.9±0.8	1.2±0.2

Figure 6.

Biodistribution (%ID/g) at various time points post i.v. administration (0.5-1 μCi; 200 μL; 25-50 pmole) of zwitterionic CdTeSe/¹⁰⁹CdZnSe r-QDs (A) and PEGylated CdTeSe/¹⁰⁹CdZnSe r-QDs (B). The calculated half-lives (t½) of r-QDs in blood were 21.4±1.1 hrs for zwitterionic and 6.4±0.5 min for PEGylated r-QDs.

Table 2.

Biodistribution of i.v. injected PEGylated NIR r-QDs in nude mice (n=3 per time point). Data are presented as %ID/g (mean ± sem) values determined by gamma counting

Organs	5 min	15 min	30 min	1 hr	6 hrs	24 hrs
Blood	19.2±1.9	3.7±1.4	0.8±0.3	0.8±0.1	0.3±0.0	0.1±0.0
Heart	2.4±0.5	1.0±0.3	0.5±0.1	0.3±0.0	0.2±0.1	0.2±0.1
Lungs	4.1±0.1	1.4±0.3	0.6±0.1	0.5±0.1	0.3±0.1	0.4±0.1
Liver	50.5±4.1	55.5±6.2	71.3±0.5	82.0±1.3	74.9±1.6	81.2±9.0
Spleen	19.6±1.1	21.0±2.9	22.7±1.3	29.2±0.7	28.0±0.1	26.2±2.3
Stomach	0.5±0.2	0.2±0.1	0.2±0.0	0.9±0.7	0.3±0.1	1.5±1.0
Intestine	0.8±0.2	0.3±0.0	0.2±0.0	0.2±0.0	0.2±0.0	0.1±0.1
Kidneys	2.7±0.2	0.8±0.1	0.5±0.0	1.0±0.4	0.3±0.1	0.4±0.2
Skin	1.0±0.2	0.4±0.1	0.2±0.0	0.4±0.1	0.3±0.1	0.3±0.1
Muscle	0.3±0.2	0.4±0.1	0.5±0.3	0.1±0.1	0.3±0.1	0.01±0.0
Skull	3.3±1.3	3.7±0.6	3.0±0.4	3.9±1.6	4.1±0.2	4.1±1.1
Brain	0.4±0.0	0.2±0.1	0.1±0.0	0.03±0.0	0.1±0.1	0.01±0.0
Femur	5.9±0.7	10.9±1.7	8.3±1.9	7.2±2.2	12.9±3.0	8.6±1.7

The in vivo and ex vivo fluorescence imaging of both zwitterionic and PEGylated r-QDs in nude mice obtained 1 hr post tail vein injection is shown in Figure 7. Whilst zwitterionic r-QDs show diffuse signal (Figure 7A), the PEGylated r-QDs clearly show higher liver uptake (Figure 7B) consistent with the ex vivo fluorescent image and the radiotracer biodistribution data (Figure 6B).

Figure 7.

In vivo and ex vivo fluorescence images of nude mice injected with 600 pmole zwitterionic CdTeSe/¹⁰⁹CdZnSe (A), and 300 pmole PEGylated CdTeSe/¹⁰⁹CdZnSe (B). Images were acquired 1 hr post tail vein injection. These images are unmixed fluorescent images from which the background and autofluorescence have been subtracted. Therefore the red color implies the presence of the r-QDs, whereas black color indicates absence of fluorescent signal from the r-QDs. HE: heart, LU: lungs, LI: liver, SP: spleen, IN: intestine, KI: kidney, ST: stomach, MU: muscle, BO: skull, SK: skin.

Discussion

The main objective of this study was to develop intrinsically radioactive quantum dots that combines both radioactivity and fluorescence signals. A highly efficient radio-labeling method was developed to incorporate the radionuclide between cold core and shell of quantum dots by SILAR method. The radioactive layer was formed between the cold core and shell to take advantage of the higher reactivity of Cd. Multi-layer cold shell provides better protection to the radio -label layer and improves fluorescent quantum yield of QDs. Because of high efficiency of adsorption and reaction of radio-tracer on the surface of core, almost quantitative radiochemical reaction yield was achieved. To our best knowledge, this is the first report that utilizes SILAR method to synthesize intrinsically radioactive QDs. This method can also be adapted for the synthesis of intrinsically radioactive nanoparticles other than QDs, for instance, other metal-and metal oxide-based nanoparticles. Furthermore, the method requires one less step of radioactive synthesis, making it more compatible with short and medium half-life radionuclides. The shorter synthesis could also be adapted to an automated micro-fluidic synthesizer, to minimize personnel exposure to radiation, particularly during scale up, and to provide a reproducible production method for much needed quantitative in vivo biological studies of nano-particles in general. Although we used a small amount (200 μCi) of ¹⁰⁹Cd for the radiosynthesis, to enable ex vivo quantitative studies, more radioactivity can certainly be incorporated to allow in vivo SPECT imaging. However, since ¹⁰⁹Cd was only used to demonstrate proof of concept and not intended for future imaging applications due to its long half-life, low gamma energy abundance and high cost, the use of a large radioactivity amount was not considered in the present work. Bimodal imaging with fluorescence and SPECT, using more clinically relevant radionuclides such as ¹¹¹In is currently being carried out in our laboratory. This will enable not only the use of high radioactivity synthesis, but also high specific activity r-QDs.

For biological application, hydrophobic radioactive quantum dots have to be coated by either amphiphilic molecules through non-covalent interaction or hydrophilic ligands through ligand exchange to make them water-soluble and biocompatible. The surface functionalization of nanoparticles could profoundly influence their in vivo pharmacokinetics and interaction with biological compartments [38]. Therefore design and selection of ligands to coat the nanoparticles are crucial for achieving specific properties for biological applications. It is well known that PEG, polysaccharides, peptidomimetic or zwitterionic surface modification tend to reduce undesirable nonspecific protein binding of nanoparticles [39]. In this paper, we designed and synthesized two types of DHLA-based ligands, which contain either zwitterionic or PEGylated moiety. Both ligands coated r-QDs showed high stability in aqueous solution and good biocompatibility. The zwitterionic and PEGylated ligands coated r-QDs also possess compact hydrodynamic size, which is due to the small size of zwitterionic ligand and short PEG-chain of PEGylated ligands. No radioactivity was lost from the r-QDs after ligand exchange reaction, which further verify that the shell provides excellent protection to the radioactivity of r-QDs. Fluorescence and gamma radioactive signals were detected from r-QDs in the same samples, demonstrating a bimodal detection. The incorporation of imageable and more clinically relevant radionuclides into the r-QDs is under development in our laboratory.

Biodistribution study of zwitterionic and PEGylated r-QDs based on gamma counting indicated obvious difference of their in vivo pharmacokinetics according to their surface coating. Much longer blood circulation and slower accumulation of zwitterionic r-QDs in liver and spleen was observed. The longer blood half-life of zwitterionic r-QDs could be due to the reduced nonspecific binding with serum proteins and highly hydrophilic property of zwitterionic ligands [23] in addition to the decreased accumulation in liver and spleen. In our hands the zwitterionic r-QDs showed significantly less accumulation in bone than the PEGylated r-QDs. Relatively high bone uptake of PEGylated QDs has also been reported [37]. An interesting observation is also seen in the significantly higher skin uptake with the zwitterionic r-QDs than the PEGylated ones. Previous reports have shown that the uptake of untargeted QDs in NHEK skin cells is dependant on surface coating. These studies demonstrated that the order of nanoparticles uptake in these skin cells increases as follows: QD-COOH>QD-NH₂>QD-PEG [40,41]. This helps to understand the reason behind the higher skin uptake we observed with zwitterionic r-QDs than that of the PEGylated r-QDs. The outer most layer of zwitterionic r-QDs is sulfonate groups which are negatively charged (similar to carboxylic acid) and therefore have greater interaction with lipid layers on the skin cell membrane, as these negatively charged groups are the first part of the coated quantum dots to be exposed to the lipid layers. This outer most group coating effect could also play a part to affect the interactions of quantum dots with other cell lipid layers. The Intrinsic radio-labeling technology described in this paper, combined with quantitative gamma counting, and autoradiography, as well as high power resolution microscopy will allow better understanding of r-QDs interactions with different cellular compartments from initial cell membrane interactions to intracellular internalization and distribution. It is important to point out that high sensitivity of the radiotracer technique, enabled by the intrinsic radioactivity in our r-QDs, allowed us to measure very small changes in the tissue accumulation such as that observed in skin and bone. Capturing these small changes in tissue uptake would not have been possible by optical methods alone due to the non-quantitative nature and attenuation of the fluorescent signal.

Ex vivo and in vivo fluorescence imaging of r-QDs in nude mice were also acquired. The ex vivo and in vivo fluorescence imaging reflected the similar distribution of r-QDs in different tissues and consistent with the results of biodistribution based gamma counting. The expected prolonged blood circulation time with the zwitterionic coated r-QDs prompted us to test the passive tumor targeting due to the enhanced permeability retention (EPR) effect in Colo-205 (colon adenocarcinoma) xenografts. The tumor uptake reached a maximum of 5.14 %ID/g 48 hrs after injection. Due to the very short blood half-life of the PEGylated r-QDs, it is not expected that any tumor uptake would have been observed, though this was not tested in the present study. Work is underway to target zwitterionic coated r-QDs (with various radionuclides) to receptor expressing tumors. This should yield enhanced tumor uptake beyond what we have observed with non-targeted r-QDs.

Conclusions

In summary, a novel method to synthesize intrinsically radio-labeled quantum dots, with radioactivity incorporated either in the core (radio-core) or in the shell (radio-shell) was developed. The latter method yielded r-QDs at almost quantitative radiochemical reaction yield. The design and synthesis of two different coating ligands affected the biodistribution profile of the r-QDs. Overall the zwitterionic coated r-QDs exhibited better pharmacokinetic profile, which could provide targeted imaging with enhanced signal contrast. Efforts are underway to synthesize r-QDs using other diagnostic and therapeutic radionuclides.