Optical drug monitoring: photoacoustic imaging of nanosensors to monitor therapeutic lithium in vivo.

Poly(vinyl chloride), high molecular weight (PVC), bis(2-ethylhexyl) sebacate (DOS), sodium tetrakis[3,5-bis(trifluoromethyl)phenyl] borate (NaTFPB), 9-Dimethylamino-5-[4-(15-butyl-1,13-dioxo-2,14-dioxanonadecyl)phenylimino]benzo[a]phenoxazine (Chromoionophore VII;CHVII), 6,6-Dibenzyl-1,4,8-11-tetraoxacyclotetradecane (Lithium Ionophore VI; LiI VI), trioctylphosphine oxide (TOPO), tetrahydrofuran (THF), 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), dichloromethane (DCM), 9-Dimethylamino-5-[4-(16-butyl-2,14-dioxo-3,15-dioxaeicosyl)phenylimino]benzo[a]phenoxazine (chromoionophore II; CHII), N,N-Dicyclohexyl-N′,N′-diisobutyl-cis-cyclohexane-1,2-dicarboxamide (Lithium Ionophore III; LiI III), 2-nitrophenyl octyl ether (NPOE), 2-nitrophenyl phenyl ether (NPPE), and lithium chloride were purchased from Sigma Aldrich (St. Louis, MO). 1,2-disteroyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-550] ammonium salt in chloroform (DSPE-PEG) was purchased from Avanti Polar Lipids (Alabaster, AL). 1,1'-Dioctadecyl-3,3,3',3'-Tetramethylindotricarbocyanine Iodide(DiR) was purchased from Life Technologies (Grand Island, NY) 2-Amino-2-hydroxymethyl-propane-1,3-diol, 2M solution, (TRIS, 2M) was purchased from Fisher Scientific (Waltham, MA). Phosphate buffered saline (PBS with Ca++ and Mg++, pH=7.4) was purchased from Boston Bioproducts(Ashland, MA). Nanosensor Fabrication Protocols used in this report are based on those previously described 37 - 38 . In brief, the process of fabricating optodes and nanosensors starts with formulation of an optode cocktail comprising 500 µL THF containing PVC, DOS, and the sensing components. The ratio of these components is tuned to control the response of the nanosensors. The formulation of the nanosensors used in the manuscript is 60 mg PVC, 120 µL DOS, 20 mg TOPO, 20 mg LiI VI, 22 mg NaTFPB, and 2 mg CH VII dissolved in 1 mL of THF. The formulation for fluorescence nanosensors also includes 0.5 mg of DiR in each batch. To fabricate nanosensors from this optode cocktail, 2 mg of DSPE-PEG (80 µL of a 25 mg/mL in chloroform) was dried in a 4 dram scintillation vial and then resuspended in 5 mL PBS with a probe tip sonicator for 30 seconds at 20% intensity (Branson, Danbury CT). 50 µL of the optode cocktail was combined with 50 µL of dichloromethane, and added to the PBS/PEG-lipid solution under probe tip sonication (3 minutes, 20% intensity). The nanosensor solution was filtered with a 0.8 µm syringe filter to remove excess polymer (Pall Corporation, Port Washington, NY). Nanosensors were sized using dynamic light scattering (DLS) with a Brookhaven 90Plus (Brookhaven Instruments, Holtsville, NY). The nanosensors were concentrated approximately 30-fold for in vivo experiments and approximately 20-fold for in vitro photoacoustic imaging using Amicon Ultra centrifugal filters (0.5 mL volume, 50 kDa MWCO, Millipore Corporation, Billerica, MA). In vitro fluorescence experiments used unconcentrated nanosensors.

Fluorescent Nanosensor Characterization

Nanosensors were calibrated in vitro utilizing a Lumina II in vivo imaging system (Caliper Life Sciences, Hopkinton, MA). This plate was imaged with two channels. DiR: high lamp power, excitation filter centered at 745 nm (30 nm bandpass), emission filter from 810 nm to 875 nm, and a 1 second exposure. CHVII FRET: high lamp power, excitation filter centered at 640 nm (30 nm bandpass), emission filter from 810 nm to 875 nm, and a 1 second exposure. For data analysis regions of interest were drawn over each well using Living Image 4 software (Caliper Life Sciences) and total fluorescent intensity values were obtained for each well. The ratio of the intensities of the two channels (DiR / FRET) was calculated and then converted to alpha 23 by normalizing to the ratio obtained in the presence of 0.25 mM HCl and 0.25 mM NaOH, which set the maximum protonated and deprotonated states for the sensors. The data for the calibration curve was fit to a four parameter logistic curve using Prism 6 to determine EC50 and sensitivity. The bulk of nanosensor formulation and screening experiments examine only the chromoionophore fluorescence rather than a ratio of two fluorescent signals. Additionally, screening experiments used 10 mM HEPES, 6 mM TRIS as the buffer solution instead of PBS. This is to enable quantification of lithium selectivity over sodium. Nanosensors were added to a 96 well plate and lithium (or sodium) solutions were then added to the wells to final concentrations of 0 mM through 1M and the plate was scanned with a Spectramax M3 plate reader. Endpoint fluorescence values were obtained at an excitation wavelength of 660 nm, emission at 705 nm, and a cutoff filter at 665 nm. In experiments where DiR was included in the formulation, fluorescence was measured with excitation:emission:cutoff at 640:780:695 for FRET and 740:780:NA for DiR fluorescence. Fluorescence spectrums were acquired using similar settings. Absorbance spectrums were taken between 400 and 850 nm. To characterize nanoparticle diameter and morphology, scanning electron microscopy (SEM) was performed on fluorescent nanosensors. Nanosensors were dried under vacuum on conductive adhesive tape attached to 25 mm aluminum SEM mounts from Electron Microscopy Sciences (Hatfield, PA). SEM images were taken using a Hitachi S4800 HRSEM operating at an accelerating voltage of 5.0 kV. Nanoparticle diameter was measured using ImageJ.

Photoacoustic System

Two photoacoustic tomography systems were used for this research. For in vitro phantom experiments, photoacoustic computed tomography (PACT) was used as previously detailed 39 - 40 . Briefly, a tunable optical parametric oscillator laser (basiScan 120, Spectra-Physics) pumped by an Nd:YAG laser (Brilliant b, Quantel) with 10 Hz pulse repetition rate was used to excite photoacoustic waves. Light exiting the laser system was homogenized by an optical diffuser and then illuminates the sample. The photoacoustic waves were detected by a 512-element circular transducer array (Imasonic) with 5 MHz central frequency. The imaging system has 0.10 mm radial resolution and 0.10-0.25 mm tangential resolution 41 . The image was reconstructed based on back projection algorithm. For in vivo monitoring experiments, we used a deep reflection-mode photoacoustic tomography system as previously detailed 42 . Briefly, a dark-field focusing illumination was achieved through a series of optical elements from the same laser source as above. The optical fluence at the animal surface was controlled to be below the maximum permissible exposure set by the American National Standards Institute 43 . A photodiode (SM05PD1A, Thorlabs) was used for monitoring and compensating for the fluctuation of the laser energy. The excited ultrasound waves were detected by a 10 MHz central frequency focused ultrasonic transducer (V315, Panametrics-NDT) and the signals were then amplified by an amplifier (5072PR, Panametrics-NDT) and digitized by an oscilloscope (ZT4421, ZTEC Instruments). The imaging system was mounted on a linear translation stage (XY-6060, Danaher Motion) to acquire three-dimensional images. The spatial resolutions of the system are approximately 0.17 mm and 0.29 mm in the axial and lateral directions, respectively.

Photoacoustic In Vitro Characterization

Mixtures of nanosensors and different concentration of lithium were injected into silicone tubing. Then these pieces of tubing were sealed by rubber at both sides. The sealed tubing was embedded in agar gel and covered by chicken breast tissue for imaging. Based on these initial results, we chose imaging parameters for later in vivo imaging to minimize signal from vasculature and tissue to focus on the signal from nanosensors alone. In Vivo Studies All in vivo studies were approved by the institutional animal care and usage committee (IACUC) of Northeastern University. In vivo photoacoustic studies were also approved by the institutional animal care and usage committee of Washington University in St. Louis. In Vivo Photoacoustic Studies The mice used in this research were Hsd:Athymic Nude-Foxn1 nu from Harlan Laboratories. Photoacoustic imaging experiments were conducted using the system described above. Animals were anesthetized with 1.5% isoflurane in oxygen and 40 µL of nanosensors were injected i.d. into the mouse on the back. The animals were imaged with two channels at excitation wavelengths 515 nm and 660 nm (the peak wavelengths for the PA signal). Baseline raster scan images were acquired at both wavelengths and continuous monitoring was done with both wavelengths at the center of the nanosensor injection. Data points were acquired approximately every minute after an i.p. injection of 38 mg/kg lithium (administered as lithium chloride in PBS, experimental) or a matching volume of PBS (control). Images were acquired every minute for approximately one hour. Animals were imaged one at a time. All animals were sacrificed after experiments were completed. For data analysis of each experiment, an index of the two photoacoustic intensities was generated by dividing the 515 nm signal by the 660 nm signal. These ratios were normalized to that at the first time point after injection of lithium. These data were linearly interpolated to align time and intensity points before averaging. Error bars for lithium dataset represent the standard deviation of three animals.

Fluorescent in vivo Studies

The mice used in this research were male SKH1-E Nude mice from Charles River (Wilmington MA). Fluorescent imaging experiments were conducted using a Lumina II in vivo imaging system (IVIS). Animals were anesthetized with 2.25% isoflurane in oxygen and placed in the animal imager. 30 µL of nanosensors were injected i.d. into the mouse on the back. The animals were imaged with two channels. DiR: high lamp power, excitation filter centered at 745 nm (30 nm bandpass), emission filter from 810 nm to 875 nm, and a 1 second exposure. CHVII-FRET to DiR: high lamp power, excitation filter centered at 640 nm (30 nm bandpass), emission filter from 810 nm to 875 nm, and a 1 second exposure. Baseline images were acquired for approximately 30 minutes, followed by an i.p. injection of 12 or 38 mg/kg lithium (administered as lithium chloride in PBS, experimental) or a matching volume of PBS (control). Images were acquired every minute for approximately one hour. The imaging equipment could not image six animals simultaneously, so animals were imaged in pairs. All animals were sacrificed after experiments were completed. For data analysis of each experiment, a region of interest encompassing the injection area was selected and total fluorescent intensity for each channel was recorded. At each time point an index was generated by dividing the intensity from DiR channel by the CHVII-FRET channel. These ratios were normalized to the first time point before injection of lithium. This data was then averaged together across three experimental animals for each lithium concentration and six control animals using linear interpolation to align time and intensity points before averaging. Error bars represent the standard deviation of these animals.