Environ. Sci. Technol. 2008, 42, 6264–6270
Computational and Ultrastructural Toxicology of a Nanoparticle, Quantum Dot 705, in Mice PINPIN LIN,† JEIN-WEN CHEN,† LOUIS W. CHANG,† JUI-PIN WU,† LAUREL REDDING,† HAN CHANG,‡ TENG-KUANG YEH,§ CHUNG SHI YANG,| MING-HSIEN TSAI,† HSIU-JEN WANG,† YU-CHUN KUO,† AND R A Y M O N D S . H . Y A N G * ,†,⊥ Division of Environmental Health and Occupational Medicine, National Health Research Institutes, Zhunan, Taiwan, Department of Pathology, Chung Shan Medical University, Taichung, Taiwan, Division of Biotechnology and Pharmaceutical Research, National Health Research Institutes, Zhunan, Taiwan, Center for Nanomedicine Research, National Health Research Institutes, Zhunan, Taiwan, and Quantitative and Computational Toxicology Group, Department of Environmental and Radiological Health Sciences, Colorado State University, Fort Collins, Colorado 80523-1680
Received January 25, 2008. Revised manuscript received June 6, 2008. Accepted June 6, 2008.
We conducted pharmacokinetic and toxicology studies on Quantum Dot 705 (QD705) in male ICR mice for up to 6 months after a single intravenous dose. Time-course sacrifices were carried out at 1, 4, and 24 h; 3, 7, 14, and 28 days; and 6 months on groups of six mice per time point. Mass balance studies were also carried out at 24 h, 28 days, and 6 months. Using inductively coupled plasma mass spectrometry, various tissues, urine, and feces were analyzed for cadmium (Cd111), which is a major (46%) component of QD705. On the basis of these experimental studies, a physiologically based pharmacokinetic computer simulation model was developed with excellent predictive capability for the time-dependent kinetic and distributional changes of QD705 in tissues. QD705 persisted and accumulated in the spleen, liver, and kidneys for at least 28 days with little or no disposition but was gradually and partially eliminated by 6 months. Although histological alterations of the spleen, liver, and kidney by light microscopy are unremarkable, investigation using electron microscopy on numerous renal samples revealed definitive mitochondrial alterations in renal tubular epithelial cells at 28 days and 6 months postdosing. Health implications and potential beneficial applications of QD705 are suggested.
* Corresponding author phone: 970-491-5652; fax: 970-491-7569; e-mail:
[email protected]. † Division of Environmental Health and Occupational Medicine, National Health Research Institutes. ‡ Chung Shan Medical University Hospital. § Division of Biotechnology and Pharmaceutical Research, National Health Research Institutes. | Center for Nanomedicine Research, National Health Research Institutes. ⊥ Colorado State University. 6264
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 16, 2008
Introduction Quantum Dots (QDs) are semiconductor nanocrystals with unique fluorescence properties, such as size- and composition-tunable emission from visible to infrared wavelengths and high levels of photostability (1). The most recently synthesized and commercially available QDs comprise metallic components of cadmium-tellurium (CdTe) or cadmium-selenium (CdSe) cores with zinc sulfide (ZnS) shells (2). These nanocrystals may be surface-modified with methoxy-polyethylene glycol (PEG-5000) to enhance biological compatibility. QDs have great potential to become one of the most useful tools for biomedical applications. One of the potential applications of QDs is to be used as optimal fluorophores for in vivo biomedical imaging (1, 3, 4). Other applications of QDs have also been proposed for biological labeling or cell targeting (e.g., labeling neoplastic cells) following conjugation with specific bioactive moieties (2, 5–8). Despite the immense potential for QDs’ clinical and research applications (1–8), their biological effects and safety are still unclear (9, 10). We conducted quantitative pharmacokinetic and toxicology studies on QD705 in male ICR mice for up to 6 months after a single intravenous dose. Experimental data were used to construct a physiologically based pharmacokinetic (PBPK) model which consisted of five compartments (blood, spleen, liver, kidney, and body). The overall objectives of the present study were (1) to develop a PBPK computer simulation model to predict the timedependnet kinetic and distributional changes of QD705 in tissues of the mouse with an aim for future extrapolation to a human PBPK model and (2) to evaluate toxicity at the microscopic and ultrastructural levels in those tissues where QD705 accumulated.
Experimental Section Materials. The nanoparticles, QD705, used in our experiments were commercially available from Quantum Dot Corporation (Hayward, CA) as Qtracker 705 nontargeted Quantum Dots. Each particle had a CdSeTe core (with approximately 46% Cd, 10% Se, and 1% Te) and ZnS shell with a methoxy-PEG5000 coating. The size was approximately 18.5 nm, and the molecular weight was 1.5 × 106 g/mol. According to a Quantum Dot Corporation Certificate of Analysis, Novermber 30, 2005, and Giepmans et al. (11), fluorescence of QD705 was in the range of 650-750 nm with an emission max around 700-715 nm. Animals and Animal Treatment. Six-week old male ICR mice were purchased from BioLASCO (Taiwan) and acclimated for two weeks in the animal facilities at the National Health Research Institutes (NHRI). All mice were under a 12 h light/dark cycle, a temperature of 23 ( 1 °C, and 39-43% relative humidity; water and food were provided ad libitum. At the start of the experiments, the mice were approximately eight weeks old and weighed between 32.9 and 38.7 g each. Six mice per time point were randomly selected for timecourse studies. Each mouse was injected via tail vein with 40 pmol of QD705 (20 µL of a 2 µM stock solution) in saline; the injection volume was 100 µL/mouse. The IV route was chosen to (1) mimic potential human medical imaging applications and (2) provide cleaner pharmacokinetic profiles without the complication of the absorption phase. Serial sacrifices (under pentobarbital anesthesia) were carried out at 1, 4, and 24 h; 3, 7, 14, and 28 days; and 6 months following dosing of QD705. Body weights as well as organ weights of each animal were measured and recorded at each of these sacrifices. Tissue samples collected at each time point 10.1021/es800254a CCC: $40.75
2008 American Chemical Society
Published on Web 07/12/2008
FIGURE 1. A conceptual PBPK model for QD705 in mice. CVM, CVK, CVL, and CVS represent QD 705 concentrations in venous blood, kidneys, the liver, and the spleen, respectively. CA is the QD 705 concentration in arterial blood. QMC, QKC, QLC, and QSC represent blood flow to the body, kidneys, liver, and spleen.
modeled this phase of the study separately, in addition to the overall modeling of the entire 24-week duration of the experiment. Similarly, as explained below in the section on time-dependent concentrations in the tissues, the same PBPK model was used in the simulation of the pharmacokinetics of QD705 in mice during the later phase, where significant elimination of QD705 between 28 days and 6 months had to be accounted for. PBPK modeling was carried out using Berkeley Madonna software (University of California, Berkeley). Derivation of Tissue Distribution Coefficients. In this paper, a tissue distribution coefficient (DC) was considered as the ratio of the affinity of QD705 to a given tissue over that to blood; it would vary with time depending on the respective instantaneous QD705 concentrations in the blood and the individual tissues, as well as the microenvironment at the tissue site. The latter was governed by many factors including gaps of the capillary vessels, the topography of the tissue site, possible binding with proteins and/or receptors, and a variety of cellular processes. We further defined DC, for a time period, t1 to t2, as the area under the curve (AUC) of QD705 concentration in tissues divided by the AUC of QD705 concentration in the blood. When we plotted the calculated DCs for each organ/compartment versus time, we noticed that each curve was sigmoidal in shape. In deriving the timedependent DCs for our PBPK modeling, the Hill equation was able to describe the sigmoid shape. To avoid the zero DC values at the origin (0,0) that could not be utilized in the simulation system, we modified the Hill equation and incorporated a new parameter, initial value, for our Hill function. The original and modified Hill equations were shown as eqs 1 and eq 2, respectively DC )
DCm )
FIGURE 2. The mechanistic processes involved in the transport of QD705 from blood to various tissues. included plasma, red blood cells, liver, lungs, kidneys, spleen, muscle, thymus, fat, brain, skin (ear), and bones. These tissues were analyzed for quantification of QD705 on the basis of Cd111 contents. Some of these tissues were also collected for light and electron microscopy examinations. Analytical. QD705 analyses were based on the quantification of Cd111 concentrations in various tissues utilizing inductively coupled plasma mass spectrometry (ICP-MS; PerkinElmer, Elan6100, USA). Briefly, tissues were dried at 95 ( 2 °C for 16 h, and the dried tissues were then weighed and liquefied with 5 mL of 15N HNO3 and 1 mL of 30% H2O2 before being subjected to microwave digestion (Multiwave 3000, Anton Paar GmbH, Graz, Austria). QD705 was quantified by measuring Cd (mass ) 111 m/z) with ICP-MS. PBPK Modeling. The spleen, liver, and kidneys were major organs for QD705 accumulation, and remarkable timedependent redistribution of QD705 occurred between the rest of the body and these three organs (12). We therefore constructed a PBPK model consisting of five compartments (blood, spleen, liver, kidneys, and body) as shown in Figure 1. Throughout this study, we used only one PBPK model for all of our computer simulations. However, because in the earlier phase (0 to 28 days) there was no evidence for metabolic transformation or fecal and renal excretion, we
DCMAX · T n
(1)
n DCT50 + Tn
Inn + DCMAX · T n
(2)
n DCT50 + Tn
where DCm was the modified DC equation, DCMAX was the maximum DC, DCT50 was the time giving half-maximum DC (hr), T was the experimental time (hr), In was the initial value for DC, and n was the Hill coefficient, which determines the overall shape of the curve. The AUC ratios between tissues and blood, as described above, were used for regression analyses (Crystal Ball software, version 2000.2, Decisioneering, Inc., Co, USA) to obtain the optimal values for DCMAX and DCT50 for our modified Hill equation for each organ or tissue. These equations were then incorporated into the mass balance differential equations for the respective tissue or body compartments in the PBPK model (see PBPK model in the Supporting Information). Time-Dependent Concentrations in the Tissues. During the experimental period of 6 months, the factors influencing tissue concentrations of QD705 in mice changed in a timedependent manner in two different phases. In the early phase, which was between 0 and 28 days where only one factor, the distribution coefficient (see above section), exerted all of the influence on the tissue concentration of QD705 because there was neither metabolism nor urinary and fecal excretion (12). As reported in our earlier publication (12), QD705 levels in the spleen, liver, and kidneys remained high, and we were able to recover quantitatively the administered dose even 28 days following a single IV dose in our mass balance studies. During the later phase (28 days to 6 months), however, we suspected that, in addition to DC, metabolic activity and/or the possible involvement of a transporter contributed to the tissue kinetics of QD705. Mass balance study results at the end of 6 months indicated that the mouse now only retained VOL. 42, NO. 16, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6265
FIGURE 3. Comparison of PBPK model simulations with experimental tissue concentrations of QD705 28 days following an intravenous dosing. 40-60% of the initially administered dose. Presently, we do not have data on the extent of metabolism of QD705 in mice. Neither do we have any detailed quantitative information on fecal or renal excretion of QD705 in mice. To keep things simple, therefore, we assumed that all QD705 disposition in this phase was associated with the first-order rate of metabolism, kf, in the liver. Similar to the DC, the timedependent nature of this rate constant was described by a Hill function as shown below: kf )
k fn kfINI + kfMAX · T kfn k fn kfHalf + Tkfn
(3)
where kfINI is the initial value of kf, which is a very small nonzero value; kfMAX is the maximum kf; kfn is the Hill coefficient, which determines the overall shape of the curve; and kfHalf is the time giving half-maximum kfMAX. During the earlier phase (first 28 days), kfn ) 1 and kfMAX and kfHalf were all zero. Thus, kf which equaled kfINIkfn was a small nonzero number; the tissue concentrations were largely the results of DCs. During the later phase between 28 days and 6 months, kfMAX and kfHalf became progressively more significant. Together with DCs, they affected the time-course kinetics of the tissue QD705. Specimens Preparation for Transmission Electron Microscopy (TEM). Tissues were fixed with 2% glutaraldehyde-2% formaldehyde in phosphate-buffered saline at 4 °C overnight and then gradually dehydrated with ethanol and propylene oxide. The tissues were infiltrated in an embedding medium containing 50% (v/v) propylene oxide (Electron Microscopy Sciences, Ft. Washington, PA) for one hour, and then 33.3% propylene oxide overnight followed by 100% embedding resin EMbed 812 (Electron Microscopy Sciences, Ft. Washington, PA) for 2 h at room temperature. Tissue blocks polymerized in an oven at 60 °C for 24 h. Ultrathin sections were cut with an ultramicrotone. Sections were examined and photographed with TEM (H-7650, Hitachi, Japan).
Results and Discussion The Integrity and Fate of QD705 under the Experimental Conditions. Under the analytical preparation conditions of 15N HNO3 and 30% H2O2, we believe that the degradation of QD705 would have been complete. The subsequent microwave digestion would further ensure the complete degradation of QD705. At this stage, all metal components are in ionic form. Therefore, when we did ICP-MS analyses, we were indeed quantifying Cd111. Regarding the fate of 6266
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 16, 2008
QD705, our belief is as follows: during the initial 28 days following the IV treatment, QD705 was neither metabolized nor excreted to any significant extent. We reach this conclusion on the basis of the quantitative recovery of the administered dose in the body of the mouse in our mass balance studies at 24 h and 28 days. The fact that 100% recovery of the administered dose was obtained in the mouse body also provided evidence that QD705 per se could not possibly be excreted intact. However, between 28 days and 6 months, there was a significant level of degradation and/or metabolism of QD705, as the recovery in the mouse body was down to 40-60% in our mass balance studies conducted at 6 months. Thus, somewhere between 28 days and 6 months, the integrity of QD705 had been compromised presumably by either chemical degradation and/or metabolic disposition. PBPK Modeling for the Early Phase (0 to 28 days). In our pharmacokinetic studies, there was no significant excretion or metabolism of QD705 in the 28 days following dosing, and considerable tissue redistribution occurred to concentrate QD705 in the spleen, liver, and kidneys (12). Therefore, we believed that the affinity of QD705 to blood versus that to tissues was likely the most important factor in blood and tissue kinetics of this nanoparticle during the early phase. Since QD705 could not be dissolved in the blood or tissues, there was no such thing as blood/tissue partition coefficients, which, collectively, was an important parameter in PBPK modeling (13). Instead, we proposed that time-dependent “tissue distribution coefficients” were operational between blood and various tissues. These distribution coefficients represented hybrid constants which described the physicochemical and physiological processes of deposition, protein binding, transport through openings in discontinuous or fenestrated capillaries, phagocytosis, transcytosis, and endocytosis of QD705 in vivo at any given time. As shown in Figure 2, because gaps of capillaries in the liver and spleen (discontinuous capillaries) were large enough (
