Novel In Vivo Imaging Analysis of an Inner Ear Drug Delivery System in Mice: Comparison of Inner Ear Drug Concentrations over Time after Transtympanic and Systemic Injections

Objective Systemic steroid injections are used to treat idiopathic sudden-onset sensorineural hearing loss (ISSHL) and some inner ear disorders. Recent studies show that transtympanic (TT) steroid injections are effective for treating ISSHL. As in vivo monitoring of drug delivery dynamics for inner ear is lacking, its time course and dispersion of drugs is unknown. Here, we used a new in vivo imaging system to monitor drug delivery in live mice and to compare drug concentrations over time after TT and systemic injections. Methods Luciferin delivered into the inner ears of GFAP-Luc transgenic mice reacted with luciferase in GFAP-expressing cells in the cochlear spiral ganglion, resulting in photon bioluminescence. We used the Xenogen IVIS® imaging system to measure how long photons continued to be emitted in the inner ear after TT or systemic injections of luciferin, and then compared the associated drug dynamics. Results The response to TT and IP injections differed significantly. Photons were detected five minutes after TT injection, peaking at ∼20 minutes. By contrast, photons were first detected 30 minutes after i.p. injection. TT and i.p. drug delivery time differed considerably. With TT injections, photons were detected earlier than with IP injections. Photon bioluminescence also disappeared sooner. Delivery time varied with TT injections. Conclusions We speculate that the drug might enter the Eustachian tube from the middle ear. We conclude that inner-ear drug concentration can be maintained longer if the two injection routes are combined. As the size of luciferin differs from that of therapeutics like dexamethasone, combining drugs with luciferin may advance our understanding of in vivo drug delivery dynamics in the inner ear.


Introduction
Sensorineural hearing loss is mostly caused by inner ear disorders. Systemic injection of steroids is generally used in inner ear diseases, especially idiopathic sensorineural hearing loss. Recent studies show that an approach for drug delivery through the tympanic membrane to the round window (RW) into the inner ear is as effective as systemic injections [1]. However, there are a few studies examining the pharmacokinetics of inner ear drug delivery, and the time course remains elusive. Patients with hearing loss also exhibit a significant rate of RW obstruction, suggesting that drugs may not infiltrate the cochlea in these cases [2] [3]. When the RW is covered with connective tissue, we expect drug delivery into the inner ear can be facilitated.
Several animal experiment studies have been performed, but the inner ear has a very small volume of fluid, even when including both perilymph and endolymph (total 0.19 ml in mouse [4]). This makes reliable measurement difficult.
Microdialysis is an effective way for analyzing the cochlear fluid of animals. It has the advantage of making repeated measurements possible, allowing drug time course determination, preventing measurement artifacts arising through perilymph volume loss due to leaks, and lessens disturbance of perilymph due to the low volumes of drug recovery. However, there are several limitations associated with using microdialysis to analyze the pharmacokinetics of the inner: (1) technical difficulty of the measurement; (2) restricted locations available for measurement (i.e., basal turn of cochlea); (3) inability to quantify clearance values accurately for the inner ear; and (4) inability to measure the time course in the same animal. To solve these previously reported problems, we developed a new in vivo imaging technique and investigated whether RW obstructions block drug delivery into the inner ear.

GFAP-Luc transgenic mice
The transgenic GFAP-luc mice (FVB/N background; [5]) were obtained from Xenogen Corporation (Alameda, CA), and were backcrossed with CD1 (ICR, SLC Japan) for 5 or 6 generation. The mice contain a firefly luciferase gene expression cassette that is regulated by 12 kb of the murine GFAP promoter and the human b-globin intron 2 [5]. Luciferin delivered into the inner ear of these mice reacts with luciferase that is expressed in the GFAPexpressing cells in the cochlear nerve and spiral ganglion, and the resulting photons are detected by the camera. All experiments were approved by the Animal Care and Use Committee of Keio University (Permit Number: 08020) in accordance with the Guide for the Care and Use of Laboratory Animals (National Institute of Health, Bethesda,MD). All surgery was performed under ketamine and xylazine anesthesia, and all efforts were made to minimize suffering.

Round window niche obstruction surgery
Adult GFAP-Luc mice were anesthetized with an intraperitoneal (i.p.) injection of ketamine (100 mg/kg) and xylazine (10 mg/ kg). As previously described, an incision was made from the left post-auricular region to the submandibular region, and the tympanic bulla was exposed without damaging the facial nerve and vessels [6]. The bulla was perforated with a surgical needle, and the hole was expanded with forceps to approximately 2 mm diameter in order to cover over the RW with fascia (RW obstruction) without disturbing the auditory ossicular chain, mimicking clinical cases with RW false membrane. The hole in the cochlea and the opening in the tympanic bulla were sealed with carboxylate cement. Muscle and skin were sutured. It took approximately 20 min to complete this procedure. Left earlap was surgically removed to expose tympanic membrane for the injection of luciferin and to reduce the background.

Transtympanic or intraperitoneal injections
Just prior to imaging, we injected 0.1 ml or 0.5 ml of luciferin substrate (150 mg/kg) directly onto the round window membrane or intraperitoneally, respectively. Transtympanic injection was performed securely under the surgical microscope.

Statistics
Peak bioluminescence intensity and the time when the luminescence was maximum in individual groups (i.e., ip, TT injection and TT injection after round window closure) was examined by non-parametric t-test. All scores were averaged and analyzed with SPSS software 19.0 (Chicago, USA).

Bioluminescence imaging
An IVIS spectrum and CCD optical macroscopic imaging system (Xenogen, Alameda, CA) was used for spatiotemporal detection of the luciferase -luciferin reaction. In vivo bioluminescent images were captured immediately after i.p. injection (0.5 ml) or after transtympanic injection (TT, 0.1 ml) of D-luciferin (D-(2)-2-(6-hydroxy-2-benzothiazolyl) thiazone-4-carboxylic acid). The field-of-view was set at 10 cm. The photon count was analyzed between 0 and 40 min after the i.p. and TT injections of luciferin substrate. The integration time was fixed at 5 min for each image. All images were analyzed with Living Image software (Xenogen, Alameda, CA). The optical signal intensity was expressed as photon flux (photon count), in units of photons/s/cm 2 /steradian. Each image was displayed as a pseudocolored photon-count image superimposed onto a grayscale anatomic image. To quantify the measured light, we defined regions of interest (ROI) over the cellimplanted area and examined all values in the same ROI.

Immunohistochemistry
The animals were anesthetized with ketamine (100 mg/g) and xylazine (10 mg/g) and decapitated. The inner ear was removed and locally perfused with 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) followed by overnight fixation with the same fixative and wash with 0.1 M PBS. The otic capsules were embedded in paraffin and sectioned at 4 mm. For immunostaining, tissue sections were stained with anti-luciferase antibody (polyclonal goat, Promega) diluted at 1:75 for 60 min at room temperature, followed by an anti-goat secondary antibody (HistofineH Simple Stain MAX-PO, Nichirei Bioscience; 414351).

Characterization of the cochlea of GFAP-luc mouse
Previous bioimaging of GFAP-luc showed a signal around the ear [5]. We first investigated the origin of the signal. The signal was observed even after the local injection of luciferin thru tympanic membrane in the mouse of which earlap had been removed and the signal was seen over the temporal area (Fig. 1A). We then superimposed the biophotonic results over micro-CT image and the result clearly demonstrated that the signal was overlapped with temporal bone and not with superficial skins (Fig. 1B-D). We also dissected temporal bone of GFAP-luc and the luminescence of each part was measured: Strong signal was seen in the cochleae (Fig. 1E, arrows), while no signal was detected in the skin and ear cartridge (right). Collectively, we speculated that the signal in the ear area of the mice is derived from the cochlea. We then performed immnohistochemistry for luciferase in the cochlea, showing the expression was high in spiral ganglion (Fig. 1F,  arrows). Together, we conclude that the bioluminescence in the area of GFAP-luc reflects the reaction of luciferase and luciferin in the cochlear spiral ganglion and thus we monitor the delivery of luciferin to the ganglion by measuring the biophotonic signals using IVIS.

Bioimaging
In the TT injection group of mice, photon bioluminescence was observed five minutes after injection, reaching a peak 15-20 minutes after injection (Fig. 2B). Bioluminescence intensity gradually faded thereafter. In the i.p. injection group of mice, detectable bioluminescence appeared much later than that in the TT injection group. As expected, no bioluminescence was observed in the RW obstruction group (negative control). The lack of reaction in the negative control indicated that luciferase did not reach the inner ear.

Time curves
Trends of bioluminescence in GFAP-luc mouse and wild type control in the individual groups were examined (Fig. 3A, mean 6  SD). Peak bioluminescence intensity in the TT injection group was reached sooner after injection than in the i.p. injection group (Fig. 3C, mean 6 SEM p,0.05). Time course of the photon emission in the individual animals and their peak of TT injection group are in Fig. 4A and 4B, respectably.

Discussion
We found that RW niche obstructions, such as false membranes or fibrous connective tissue membranes, are critical for drug delivery into the inner ear. Therefore, it is very important to examine the RW before injecting drugs into the inner ear [3].
The present study also assessed the time needed for TT injections of drug delivery system (DDS) to reach the inner ear and traced DDS distribution in the inner ear. A previous report demonstrated that one hour after TT injection of methylprednisolone, dexamethasone, and hydrocortisone, drug concentration in the perilymphatic space becomes elevated [7]. However, a few studies report shorter drug elevation times after injection [8,9,10]. In this study, we detected bioluminescence in the inner ear earlier than expected; five minutes after TT injection of luciferin. A study in guinea pigs showed that significantly higher dexamethasone levels were detected in the inner ear just 30 minutes after its transtympanic administration, gradually dropping to zero after 6 hours [9]. The time differences may be due to differences in the sizes of animals (guinea pigs vs. mice), and/or due to differences in drug metabolism and drug molecular size (steroid vs. luciferin).
We observed variations in bioluminescence intensity among animals in the TT group. Individual variations between animals, especially in the local injection group, have been reported previously [7,8]. We speculated that the injected luciferin goes to the Eustachian tube of animals.
In Chandrasekhar et al., drug concentration in perilymph was 5.5262.80 mg/dL after intravenous (i.v.) injection of 0.45 mg/kg dexamethasone, but 13.2610.6 mg/dL after location injection of injection was approximately 10 times higher than that of the systemic injection, drug concentration in perilymph was nearly two times lower (5.5262.80 mg/dL vs. 13.2610.6 mg/dL) [8]. In the present study, although the dose of the systemic injections was five times higher than that of TT injections, drug concentration in perilymph was almost the same or slightly higher.
Here, we were able to observe, in real time, drug delivery and drug distribution in animals in vivo, enabling us to determine the time course of drug delivery. One limitation of this study, however, is that we investigated the delivery of only luciferin into the inner ear, not drugs commonly used to treat inner ear disorders. Differences in the molecular weights of luciferin and steroids could affect delivery and diffusion time. One future application of this system is drug tagging, in which a drug linked to luciferin can be traced. This drug-tagging system may help physicians determine which drug can be delivered into which portion of the inner ear.