Anal. Chem. 2007, 79, 8863-8869
Accelerated Articles
Multiparameter Magnetic Relaxation Switch Assays Sonia Taktak,†,‡ David Sosnovik,†,§ Michael J. Cima,| Ralph Weissleder,† and Lee Josephson*,†
Center for Molecular Imaging Research and Department of Cardiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02129, Department of Material Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, and T2 Biosystems, Inc., 286 Cardinal Medeiros Avenue, Cambridge, Massachusetts 02141
Magnetic nanoparticles (NPs) can serve as magnetic relaxation switches (MRSw’s), switching from a dispersed to a clustered state, or the reverse, due to the presence of molecular targets, with changes in the spin-spin relaxation time of water (T2). Biotinylated NP probes reacted with an avidin molecular target to form stable NP clusters, which permitted several NMR parameters to be measured as a function of cluster size. Associated with avidin-induced NP cluster formation was an increase in the spin-spin relation rate (1/T2), while the spin-lattice relaxation rate (1/T1)was unaffected. On the basis of the selective effects of NP cluster formation on T2, we developed a T1/T2 interrogation method where NP probe concentration and avidin analyte were unknown and both were determined. A third NMR parameter examined was the replication of T2 measurements, which were used to rapidly determine whether the ratio of avidin to biotinylated NP was optimal or whether additional biotinylated NP was needed. The T1/T2 and T2 replication interrogation methods illustrate how MRSw assays can employ multiple parameters, instead of relying only on T2, to obtain information about the reaction of NPs with molecular targets. The need to determine the levels of molecular targets has led to the development of probes whose physical properties change * Corresponding author. Tel: 617-726-6478. Fax: 617-726-5708. E-mail:
[email protected]. † Center for Molecular Imaging Research, Massachusetts General Hospital and Harvard Medical School. ‡ T2 Biosystems, Inc. § Department of Cardiology, Massachusetts General Hospital and Harvard Medical School. | Massachusetts Institute of Technology. 10.1021/ac701976p CCC: $37.00 Published on Web 11/06/2007
© 2007 American Chemical Society
in a manner that reflects the presence of molecules in the probe environment. Light-based environmentally sensitive (smart) probes have been used to measure the levels of materials in living systems, including hydrolytic enzymes like proteases1 and galactosidases,2 ions like calcium,3 and sequences of nucleic acids.4 Magnetic nanoparticles (NPs) can act as magnetic relaxation switches (MRSw’s), undergoing target-mediated clustering, which results in changes in the T2 relaxation times of water protons detected by NMR.5 Since MRSw methods are radiofrequency based, they are indifferent to light-based interferences (scattering, absorption, fluorescence) in tissues or fluids. Use of NP probes as MRSw’s can detect chemically diverse analytes in vitro5,6 or in vivo.7 Many rapid and routine assay methodologies are limited to measuring a single parameter (fluorescence intensity, absorbance, number of photons, densitometric peak area, intensity of light scattered, rotation of plane polarized light, radioactivity) to obtain the concentration (or amount) of an unknown analyte present in a sample. Using biotinylated magnetic NPs with an avidin analyte as a model for the MRSw methodology, we show how different types of NMR-based water relaxation measurements can be used to obtain different types of information from a sample. First, we show that measurements of T1 (spin-lattice relaxation time) and (1) Weissleder, R.; Tung, C. H.; Mahmood, U.; Bogdanov, A., Jr. Nat. Biotechnol. 1999, 17, 375-378. (2) Louie, A. Y.; Huber, M. M.; Ahrens, E. T.; Rothbacher, U.; Moats, R.; Jacobs, R. E.; Fraser, S. E.; Meade, T. J. Nat. Biotechnol. 2000, 18, 321-325. (3) Meldolesi, J. Trends Pharmacol. Sci. 2004, 25, 172-174. (4) Tyagi, S.; Kramer, F. R. Nat. Biotechnol. 1996, 14, 303-308. (5) Perez, J. M.; Josephson, L.; O’Loughlin, T.; Hogemann, D.; Weissleder, R. Nat. Biotechnol. 2002, 20, 816-820. (6) Perez, J. M.; Josephson, L.; Weissleder, R. Chembiochem 2004, 5, 261264. (7) Atanasijevic, T.; Shusteff, M.; Fam, P.; Jasanoff, A. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 14707-14712.
Analytical Chemistry, Vol. 79, No. 23, December 1, 2007 8863
T2 (spin-spin relaxation time) can be used to determine concentrations of a target where the NP probe concentration is not known a priori. The “T1/T2 interrogation method” is based on the fact that T1 reflects NP probe concentration, while T2 reflects both the avidin-induced formation of NP clusters and the concentration of biotinylated NP probes. We show that the physical basis of the T1/T2 interrogation method is consistent with a model where NP clusters behaved as porous fractal aggregates. A third parameter useful with magnetic NP cluster-based assays is the reproducibility of T2 measurements. Using the “T2 replication method” (rapid determination of T2 variability in a relaxometer), we show that one can ascertain whether the target avidin concentration in an unknown sample is within the accurate dynamic range of assay for an initial concentration of biotinylated NP. When out of range, more biotinylated NP can be added and the final concentration of NP can be unknown. However, by use of the T1/T2 interrogation method, the target avidin concentration can be determined. The use of three parameters (T1, T2, T2 replication) to interrogate the reaction between biotinylated NPs and an avidin molecular target can be used to improve the performance MRSwbased assays5,6 or sensors based on them.8 EXPERIMENTAL METHODS General Information. The N-hydroxysuccinimide ester of biotin, D-biotin-SE, and avidin were from Invitrogen (Carlsbad, CA). The diameter of NP or NP clusters was determined by photon correlation light scattering using a Zetasizer 1000HS (Malvern Instruments, Marlboro, MA). Nanoparticle Probe Synthesis. The amino-cross-linked dextran-coated iron oxide (amino-CLIO) nanoparticle was prepared as described elsewhere.9 Amino-CLIO (1.05 mg of Fe, 18.8 µmol of Fe) in 400 µL of PBS was brought to pH 8.0-8.3 with 40 µL of sodium bicarbonate (1 M). D-Biotin-SE (0.64 mg, 1.87 µmol) in 100 µL of DMSO was added and allowed to react (1 h, room temperature). Unreacted biotin-SE was removed with the Sephadex G-25 PD10 columns (GE Healthcare, Uppsala, Sweden) equilibrated with PBS. The amount of biotin attached was quantified using the EZ Biotin Quantitation Kit (Pierce Chemical, Rockford, IL). There were between 62 and 72 biotins/NP based on 8000 Fe atoms/NP.10 The biotinylated nanoparticle or NP had a diameter of 32 nm. Reaction of Avidin with NP. Avidin in PBS with 1 mg/mL BSA was mixed with NP solution in PBS. Reaction was at room temperature for 1-2 h after which T1 and T2 were determined. Samples were warmed for 10 min at 40 °C prior to measurements. Phantom Design. Phantoms consisted of six samples of three different clusters of NP and avidin (30 nm, no avidin, dispersed nanoparticles; 80 nm, 0.5 avidin/nanoparticle; and 120 nm, 1 avidin/nanoparticle) at two iron concentrations (0.2 and 0.4 mM Fe for 9.4-T experiments; 0.4 and 0.6 mM Fe for 1.5- and 4.7-T experiments). For studies at 9.4 and 4.7 T, nanoparticle solutions were placed in NMR tubes (diameter, 3 mm) and immersed in a 1% v/v Magnevist (Berlex, Montville, NJ), to decrease the signal (8) Sun, E. Y.; Weissleder, R.; Josephson, L. Small 2006, 2, 1144-1147. (9) Josephson, L.; Tung, C. H.; Moore, A.; Weissleder, R. Bioconjugate Chem. 1999, 10, 186-191. (10) Reynolds, F.; O’Loughlin, T.; Weissleder, R.; Josephson, L. Anal. Chem. 2005, 77, 814-817.
8864
Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
from the water surrounding the phantoms. For studies at 1.5 T, NMR tubes (diameter, 10 mm) were immersed in a water bath. Minispec Relaxation Time Measurements. Relaxation times were measured at 0.47 T and 40 °C using a Bruker Minispec mq20 (Bruker Optics Inc., Woodlands, TX). For T1 relaxation times, an inversion-recovery sequence was used, which consisted of 12 data points with pulse separations ranging from 5 to 1000 ms and 4 scans each. For T2 relaxation times, a spin-echo pulse sequence was used, which consisted of 200 data points with a pulse separation of 0.5 ms and 8 scans. Relaxation times of the pure solvent (PBS) were determined with pulse separations for T1 ranging from 5 to 2000 ms and by taking 800 data points for T2 determination. The T1 and T2 of the solvent were 3.4 ( 0.4 and 2.73 ( 0.04 s, respectively. Solvent relaxation rates were subtracted from relaxation rates in the presence of nanoparticles. Relaxivity was the slope of plots of 1/T versus nanoparticle concentrations. Relaxivity values were obtained by a linear least-squares fit of the data, with coefficients of correlation greater than 0.95 obtained in all cases. In a typical experiment, three measurements of T2 and two of T1 were employed. Values are the mean and standard deviation. T2 replication was a function of avidin concentration as discussed. A single T2 measurement took of ∼30 s, while T1 took ∼80 s. MRI. For the 1.5-T images (Avanto, Siemens Medical, Malvern, PA), phantoms were placed in a commercial head coil. For T1 weighted (T1W) images, a segmented inversion recovery gradient echo sequence was used. The optimal inversion time of 170 ms was selected from an inversion recovery SSFP scout sequence, where single-shot images were obtained at set increments after the inversion prepulse. Other parameters in the segmented gradient echo sequence included the following: FOV 21 × 21 cm, slice 8 mm; matrix 256 × 256; TR 1000 ms; TE 2.8 ms; and flip angle 25°. For T2 weighted (T2W) images, a spinecho sequence at the same resolution was employed, but using a TR of 2000 ms and a TE of 27 ms. T2* weighted (T2*W) images were acquired with a gradient echo sequence using a TR of 2000 ms, a TR of 19 ms, and a flip angle of 60°. For 4.7- and 9.4-T images (Biospec, Bruker, Billerica, MA), pulse sequences and parameters were chosen to maximize SNR and CNR and were guided in part by the gradients and rf coils available on the two scanners. T1W images at 4.7 T were acquired using a gradient echo sequence with TR ) 20 ms, TE ) 2 ms, and flip angle of 60°. T2W images were acquired with a spinecho sequence using the following parameters: FOV 30 × 30 mm, slice 2 mm, matrix 150 × 150, TR ) 2000 ms, and TE ) 28 ms. T2*W images were acquired with a gradient echo sequence with TR ) 1000 ms, flip angle 30°, and TE of 16 ms. Sequence parameters at 9.4 T were identical other than the following: T1W images were acquired with an inversion recovery RARE sequence using a TR of 4000 ms and an echo train of 2. T2W spin echo imaging was performed with TR ) 3000 ms and TE ) 36 ms. T2*W gradient echo images were acquired with TR ) 500 ms and TE ) 22 ms. RESULTS AND DISCUSSION In order to study the magnetic relaxation properties of avidininduced NP clusters, it was essential to establish that the clusters were stable with respect to time. As shown in Figure 1a, the
Figure 1. Avidin effects on biotinylated NPs. (a) Time independence of T2 changes for clusters of 32 nm, no avidin (9), 80 nm, 0.5 equiv of avidin/NP (4), and 120 nm,1 equiv of avidin/NP (1). (b) Cluster size determined by light scattering as a function of avidin added per nanoparticle. Avidin/NP reaction was at room temperature. (c) Detection of aggregates by relaxometry using the CV of three T2 measurements at 40 °C. CVs below 2% indicate formation of stable clusters. CVs above 2% indicate precipitation of aggregates. (d) Fate of NPs reacting with different amounts of avidin.
addition of avidin to biotinylated NPs resulted in decreased T2’s that were stable for up to 48 h. The size of avidin-NP clusters obtained was then determined by light scattering as shown in Figure 1b. Time-independent T2 measurements (Figure 1a) were obtained when the maximum cluster size was below 150 nm and the ratio of avidin to NP was below 1.2 (Figure 1b). The addition of larger amounts of avidin resulted in the formation of NP aggregates that separated from solution within hours. A feature of stable clusters (avidin to NP ratios less than 1.2; size, 30-150 nm) was that the coefficient of variation (standard deviation/mean) of three T2 measurements performed on the relaxometer (0.47 T) over 1-2 min was below 1.5% (Figure 1c). We therefore employed a coefficient of variation (CV) of below 2% as a rapid NMR-based criterion, indicating an avidin concentration was present yielding stable clusters. The T2 replication method (T2’s with CV of
