Brain Targeting of Acyl-CoA:Cholesterol O-Acyltransferase-1 Inhibitor K-604 via the Intranasal Route Using a Hydroxycarboxylic Acid Solution.

An acyl-CoA:cholesterol O-acyltransferase-1 (ACAT-1/SOAT-1) inhibitor, K-604 is a promising drug candidate for the treatment of Alzheimer's disease and glioblastoma; however, it exhibits poor solubility in neutral water and low permeability across the blood-brain barrier. In this study, we report the successful delivery of K-604 to the brain via the intranasal route in mice using a hydroxycarboxylic acid solution. In cerebral tissue, the AUC of K-604 after intranasal administration (10 μL; 108 μg of K-604/mouse) was 772 ng·min/g, whereas that after oral administration (166 μg of K-604/mouse) was 8.9 ng·min/g. Thus, the index of brain-targeting efficiency was 133-fold based on the dose conversion. Even with intranasal administration of K-604 once per day for 7 days, the level of cholesteryl esters markedly decreased from 0.70 to 0.04 μmol/g in the mouse brain. Thus, this application will be a crucial therapeutic solution for ACAT-1 overexpressing diseases in the brain.

Introduction

Acyl-coenzyme A (CoA):cholesterol O-acyltransferase (ACAT), also known as sterol O-acyltransferase (SOAT), catalyzes the acylation of free cholesterol with long-chain fatty acids to form cholesteryl esters (CEs).[1] Two ACATs have been identified: ACAT-1 and ACAT-2.[2−6] ACAT-1 is the main isoenzyme in the brain, and in multiple neurodegenerative diseases including Alzheimer’s disease (AD), inhibition of ACAT-1 provides several benefits, such as the clearance of amyloid beta (Αβ) peptides and suppression of 24(S)-hydroxycholesterol (24S-OHC)-induced neuronal cell death.[7−12] Recent studies also reported that blocking cholesterol esterification via ACAT-1 inhibition is a promising therapeutic strategy to treat glioblastoma (GBM).[13,14] This evidence has shed light on the previous use of ACAT inhibitors for the treatment of brain disease. However, ACAT inhibitors have poor blood–brain barrier (BBB) permeability because they were developed with the goal of treating peripheral arterial disease. Therefore, these fascinating trends prompted us to apply our developed ACAT-1 inhibitor to brain disease. 2-(4-(2-((1H-Benzo[d]imidazol-2-yl)thio)ethyl)piperazin-1-yl)-N-(6-methyl-2,4-bis(methylthio)pyridin-3-yl)acetamide hydrochloride, (K-604),[15−19] shown in Figure , is the first potent and selective inhibitor of ACAT-1.

Figure 1

Chemical structure and properties of K-604.

The chemical structure of K-604 was designed by a solubility-driven structural optimization strategy;[20,21] four nitrogen atoms allow the molecule to dissolve in an acidic medium and to reach an aqueous solubility of 19 mg/mL in the first fluid [pH 1.2] used for the dissolution test. However, the solubility is only 0.05 mg/mL at pH 6.8. Therefore, K-604 requires an acidic pH, which allows for good oral absorption because of its high solubility in stomach fluid.[17] It remains uncertain whether K-604 can transport across the BBB, which often obstructs the development of brain-targeted therapeutic agents that are orally administered. The poor transport of therapeutic agents across the BBB requires a large systemically administered dose to reach the required pharmacological concentration in the brain. Moreover, nontargeted tissues other than the central nervous system are commonly exposed to these drugs, which may cause adverse effects. In this study, we explored the scope and limitations of methods to deliver K-604 to the brain across the BBB while bypassing the BBB. The intranasal delivery route from the nose to the brain along the olfactory and trigeminal pathways has been indicated as a promising approach.[22−24] Although numerous studies on nasal administration have been reported, their major purpose focused on the immediate pharmacological effect followed by the reduction in systemic exposure and accompanying side effects and improved bioavailability.[25−27] The representative examples are antifungal agents and bronchitis drugs. Furthermore, the efficacy of drug delivery to the brain and the mechanism underlying the conflicting results produced even with use of the same substance, as is the case for dopamine, have not been described.[28−32] Therefore, to confirm the feasibility and practice of intranasal administration, the present study focused on the following six proposed objectives: (1) to clarify the scope and limitation of oral K-604 administration by simulating the BBB penetration ratio and evaluating BBB permeability; (2) to confirm the validity of our protocol by intranasal delivery routes from the nose to the brain after administration of a fluorescein sodium salt (uranine) solution and to observe the stained tissue areas after 2 h of staining; (3) to determine the appropriate vehicle to increase the solubility of K-604 in the hydroxycarboxylic acid aqueous solution; (4) to monitor the pharmacokinetic (PK) profiles of K-604 in tissues, including the olfactory bulb and cerebrum, with different plasma-dependent vehicles and nasal volumes; (5) to evaluate the pharmacological efficacy of K-604 as an ACAT-1 inhibitor in the brain after intranasal administration in a single daily dose for 7 days; and (6) to investigate the impact of intranasal administration of acid solutions on olfactory tissue by histologically evaluating the respiratory system and olfactory epithelium.

Material and Methods

Materials

K-604 was developed for the treatment of acute coronary syndrome in this Phase II study by Kowa Company Ltd. (Tokyo, Japan) according to a previously reported method.[17] For use as an internal standard for the low-level quantification of K-604 in plasma and brain tissues, a highly deuterium-labeled compound, 2-(4-(2-((1H-benzo[d]imidazol-2-yl-4,5,6,7-d4)thio)ethyl)piperazin-1-yl)-N-(6-methyl-2,4-bis(methylthio)pyridin-3-yl)acetamide was also prepared by Kowa Company Ltd. Hyaluronic acid (HA), citric acid (CA), and d-gluconic acid (GA) were purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Sodium 2-(6-oxido-3-oxo-3,10-dihydroanthracen-9-yl)benzoate (uranine) and 5H-dibenzo[b,f]azepine-5-carboxamide (carbamazepine) were purchased from FUJIFILM Wako Pure Chemicals Corporation (Osaka, Japan). Water was purified by a Milli-Q Gradient system (Millipore, Milford, MA, USA). The BBB Kit (RBT-24) was purchased from PharmaCo-Cell Company, Ltd. (Nagasaki, Japan). All other solvents and chemicals were of HPLC or analytical grade.

Methods

Simulation of the BBB Penetration Ratio of K-604

The BBB penetration ratio of K-604 was simulated by StarDrop and ADMET Predictor software.[33−35]

In Vitro Evaluation of the BBB Permeability of the Drug

The BBB Kit (RBT-24) is a new in vitro model of the BBB composed of primary cultures of rat (Wistar) brain capillary endothelial cells, brain pericytes, and astrocytes,[36−41] and the BBB permeability coefficient was measured using the BBB Kit. The BBB Kit was stored at −80 °C and defrosted 4 days prior to the experiment using the following procedure: (1) 1000 and 200 μL of medium (10% PDS/DMEM F-12 warmed at 37 °C) was added to the brain and blood sides, respectively, as a thawing solution; (2) the BBB Kit was incubated for 2 to 3 h in a carbon dioxide incubator; (3) after the incubation, the medium was removed and 1200 and 300 μL of medium were added to the brain and blood sides; (4) 1 day later, the medium was removed and the same volume of medium was added; (5) the transendothelial electrical resistance of the BBB Kit was measured and confirmed to be more than 150 Ω × cm2; (6) the experiment was conducted within 4 to 7 days.

A solution of Dulbecco’s PBS-HEPES (DPBS-H, 100 mL) was prepared by combining 10× Dulbecco’s PBS (Ca+/Mg+) (10 mL), 1 M HEPES (pH 7.0–7.6) (1 mL), d-glucose (0.45 g), and distilled water (89 mL). A solution of DPBS-H with 0.1% bovine serum albumin (BSA) was prepared by dissolving a solution of 0.1% BSA (34.9 μL) in the DPBS-H solution (34.9 mL) using ultrasonic vibration and used as an assay-buffered solution. Carbamazepine and K-604 were dissolved in DMSO and diluted with the assay-buffered solution to make the test solution (1 μM). The drug test solution (1 μM) was added to the upper (luminal, blood-side) insert of a 24 well Millipore plate. After 30 min, the solution in the lower compartment (900 μL) was removed. The sample was centrifuged at 16000 × g for 5 min to obtain the supernatant. The prepared sample solution was centrifuged at 12000 × g for 5 min at 4 °C to obtain the supernatant, which was then added to the tube (20 μL) and stirred. The internal standard solution (K-604-d4, 100 ng/mL, 5 μL) and CH3CN (50 μL) were added to each tube. The resulting solution was stirred and then centrifuged at 12000 × g for 5 min at 4 °C to obtain the supernatant. The supernatant (50 μL) was collected and added to the measurement vial. Each resultant solution (2 μL) was injected into a liquid chromatography (LC: ACQUITY UPLC I-Class, Waters, Milford, MA, USA)–tandem mass spectrometry (MS, Xevo TQ-S, Waters) instrument for measurement.

Solubility Studies

To address the issue of low aqueous solubility, we examined the solubility of K-604 in various solutions, such as aqueous solutions of 0.05 N HCl, 0.01 N HCl, HA, GA, and CA. A known amount of K-604 was added to measured quantities of various solutions. Minimum amounts of various solutions required to solubilize K-604 were visually determined within 24 h. The end point of the solubility study was the formation of a clear solution.

Animal Study Design

In vivo drug absorption and brain uptake studies were performed as follows. Crl:CDI (Institute of Cancer Research; ICR) female mice (8 weeks old; body weight 25 to 29 g) were used for the in vivo studies. The mice (7 weeks) were obtained from Charles River Japan, Inc. (Kanagawa, Japan). All procedures performed on animals in this study were in accordance with established guidelines and regulations and reviewed and approved by the Committee on the Ethics of Animal Experiments, Kowa Company Ltd. The mice (n = 4 to 5 in each group) were housed in cages and had free access to standard chow pellets (CE2, CLEA Japan, Inc., Tokyo, Japan) and water under uniform housing and environmentally controlled conditions.

PK Studies on Oral Administration

Thirty female ICR mice (average body weight 27.7 g) were orally administered a 0.01 N HCl solution of K-604 (6 mg K-604/10 mL HCl solution)/kg. Plasma samples (0.5 mL blood/collection) were collected with heparin at 5, 15, 30, 45, 60, and 120 min after administration. After quick extraction of the brain, the extracted brain was separated into the cerebrum, cerebellum, and olfactory and homogenized with physiological saline.

PK Studies on Intranasal Administration

Before intranasal administration,[42] all mice were anesthetized by isoflurane inhalation for 1 min. The K-604 formulation was administered into each nostril via a polyethylene tube (ep. Dualfilter T.I.P.S. SealMax 2–100 μL) attached to a micropipette (Eppendorf). The process was performed gently, allowing the mice to inhale all of the loaded formulation with normal respiration. We established four categories representing the study objectives: purposes A, B, C, and D. Purpose A was to explore the efficacy of drug delivery from the nose to the brain with intranasal administration using uranine and K-604 in a 0.01 N HCl solution. Purpose B was to choose a dose of either 50 or 10 μL of a 0.01 N HCl solution of K-604 administered once per day for 7 days. Purpose C was to measure the drug concentrations of three formulations, HA, GA, and CA solutions, after one administration (10 μL). Purpose D was to measure cholesterol levels in the brain after intranasal administration of the HA, GA, or CA solution once per day (10 μL) for 7 days. To address purpose A, 8 groups of mice were established, all of which received a single dose (50 μL): group A1 (200 mg/mL uranine, 120 min), group A2 (0.01 N HCl solution as a control, 120 min), and group A3 to group A8 (3.24 mg/mL K-604 in 0.01 N HCl solution, 5, 15, 30, 45, 60, 120 min). To address purpose B, 17 groups of mice were established, all of which were administered a single daily dose for 7 days: groups B1 and B11 (0.01 N HCl solution as controls, 50 μL and 10 μL, respectively), group B2 to group B10 (5, 15, 30, 45, 60, 120, 240, 360 min, 24 h, 0.01 N HCl solution of K-604 (3.24 mg/mL), 50 μL), and group B12 to group B17 (5, 15, 30, 45, 60, 120 min, 0.01 N HCl solution of K-604 (3.24 mg/mL), 10 μL). To address purpose C, 13 groups of mice were established, all of which were administered a single dose (10 μL): groups C1 to C4 (5, 15, 30, 120 min, HA solution formulation); groups C5 to C8 (5, 15, 30, 120 min, GA solution formulation), groups C9 to C12 (5, 15, 30, 120 min, CA solution formulation), and groups C13 (nontreatment as a control). To address purpose D, 4 groups of mice were administered one dose per day for 7 days: groups D1 (nontreatment as a control), D2 (HA solution formulation of K-604, 10 μL), D3 (GA solution formulation of K-604, 10 μL), and D4 (CA solution formulation of K-604, 10 μL).

In Vivo Monitoring of Stained Tissue along the Intranasal Route

An aqueous solution of uranine (200 mg/mL) in water was prepared. This uranine solution (50 μL) or a 0.01 N HCl solution (50 μL) was intranasally administered to 2 groups of mice (female, 8 weeks old, n = 5). After 2 h, all mice were anesthetized by isoflurane inhalation and then euthanized by cervical dislocation. The brain was quickly harvested, separated into the cerebrum, cerebellum, and olfactory bulb tissue, and rinsed with physiological saline.

In Vivo PK Study

The plasma blood, olfactory bulb, and brain samples were prepared as follows. After administration of the treatments, all mice were anesthetized by isoflurane inhalation and then euthanized by cervical dislocation. Approximately 0.5 mL of mouse blood was collected from the postcaval vein using a heparinized syringe at different time points. The blood was centrifuged at 9100 × g for 5 min at 4 °C to obtain plasma. After quick extraction of the brain, the brain was separated into the olfactory bulb, cerebrum, and cerebellum and homogenized with physiological saline. Plasma, tissues, and homogenate were stored frozen at −30 °C until use. The concentrations of K-604 in the plasma, brain, cerebrum, cerebellum, and olfactory tissue samples were determined by liquid chromatography–tandem mass spectrometry (LC–MS/MS) after deproteinization with CH3CN.

Quantitative Analysis of K-604

The LC–MS/MS system comprised a 3133 HTS autosampler Z (Shiseido, Tokyo, Japan), an Agilent 1100/1200 series HPLC system (Agilent Technologies, Santa Clara, CA, USA), and an API 4000 mass spectrometer (AB Sciex, Framingham, MA, USA). Chromatographic separation of the analytes was achieved on a Kinetex C18 column (2.1 × 50 mm, 2.6 μm, Phenomenex, Torrance, CA, USA) using 10 mmol/L ammonium formate (pH 5) (mobile phase A) and CH3CN (mobile phase B) in system B. The mobile phase was delivered at a flow rate of 300 μL/min using the following multistep gradient elution program: linear gradient 20–95% B from 0 to 1 min, 95% B from 1 to 3.5 min, and 20% B from 3.5 to 7.5 min. Mass spectrometric detection of the analytes was accomplished using the TurboIonSpray interface operated in the positive ionization mode. The analyte response was measured by the multiple-reaction monitoring of selective mass transitions for each compound. The transitions of the protonated precursor ions to the selected product ions were from m/z 503 to m/z 353 for K-604 and from m/z 507 to m/z 353 for K-604-d4 (internal standard).

Analysis of the Brain Lipid Components after Intranasal Administration for 7 Days

The level of cholesterol was measured by gas chromatography–mass spectrometry (GC–MS) using a previously reported method.[8,12] The cerebrum was homogenized in saline and diluted with saline to 10 times the tissue sample volume. Lipids were extracted from the tissue homogenate solution (1 mL) with a mixed solution of CHCl3/MeOH (2:1, v/v, 3 mL) and water (1 mL) followed by vortexing for 1 min and centrifugation at 2330 × g for 10 min at room temperature. After removal of the aqueous layer, the organic layer was equally divided into two portions: one for total sterol (with saponification) quantification and the other for free sterol (without saponification) quantification. The organic layer was evaporated to dryness under a nitrogen stream. EtOH (1 mL) and a solution of 10 M KOH in aqueous 70% EtOH (300 μL) were added to the drying residue. The resulting mixture was incubated at 80 °C for 1 h. After saponification, CHCl3 (2 mL) and water (2.5 mL) were added to the mixture followed by vortexing for 1 min and centrifugation at 2330 × g for 10 min at room temperature. Then the organic layer was evaporated to dryness under a nitrogen stream. A mixture of i-PrOH/CH3CN (55:45, v/v, 50 μL) was added to the drying residue. A total of 10 μL of the sample solution was evaporated to dryness under a nitrogen stream. For the silylation of cholesterol, N,O-bis(trimethylsilyl)trifluoroacetamide (100 μL) was added to the dried residue. The solution was vigorously mixed by vortexing and incubated for 60 min at 60 °C followed by incubation for 48 h at 25 °C to obtain the trimethylsilyl esters and ethers. An aliquot of this sample was injected into a gas chromatograph (GC-2010 Ultra; Shimadzu, Kyoto, Japan) equipped with a quadrupole mass spectrometer (GCMS-QP2010 Ultra). A fused silica capillary column (DB-5MS, phenyl arylene polymer, 30 × 0.25 mm2; Agilent Technologies, Palo Alto, CA, USA) was used. Helium was used as the carrier gas at a flow rate of 1.41 mL/min. The temperature program increased the temperature from 50 to 250 °C at 20 °C/min and from 250 to 325 °C at 5 °C/min. The injector temperature was set to 280 °C, and the temperatures of the transfer lines to the mass detector and ion source were 280 and 200 °C, respectively. The electron energy was set to 70 eV. Cholesterol was identified based on retention times and mass patterns; ions with m/z 458 for cholesterol were selected for quantification. Cholesterol was quantitatively identified using cholesterol-d7 (Avanti Polar Lipids, Alabaster, AL, USA) as an internal standard.

Histological Evaluation of Respiratory and Olfactory Epithelia and Lung (with Bronchus)

The nasal cavity was fixed in 10% neutral buffered formalin and sectioned transversely at the anterior end of the olfactory bulb according to the literature.[43] Paraffin sections were cut to 3 to 5 μm thick and then stained with hematoxylin and eosin. The thickness of the respiratory and olfactory epithelia was examined using hematoxylin and eosin staining. The mean thickness was measured at four random points in the dorsal portions of both sides of the nasal septum (magnification, 400×). All measurements were performed by the same observer. Lung tissue (with bronchus) was fixed with 10% neutral buffered formalin, and paraffin-embedded lung sections (3 to 5 μm thick) were stained with hematoxylin and eosin.

Results and Discussion

Estimation of the BBB Penetration Ratio of K-604

To estimate the potential brain permeability of K-604, we simulated the BBB penetration ratio of carbamazepine and K-604 using StarDrop and ADMET Predictor software, respectively. Carbamazepine[44,45] is the brain-targeting agent used primarily for the treatment of epilepsy and neuropathic pain. Here, we considered the BBB permeability value of carbamazepine as a positive control. The BBB log(Cbrain/Cblood) values of carbamazepine and K-604 were calculated to be −0.063 and −0.879, respectively, by StarDrop, and the carbamazepine and K-604 values were estimated to be −0.181 and −0.631, respectively, by ADMET Predictor (Table ).

Table 1

Simulation of the BBB log(Cbrain/Cblood) Values of Carbamazepine and K-604

method	carbamazepine BBB log(Cbrain/Cblood)	K-604 BBB log(Cbrain/Cblood)
StarDrop	–0.063	–0.879
ADMET Predictor	–0.181	–0.631

As expected, carbamazepine exhibited a high BBB penetration ratio. In contrast, K-604 exhibited a moderate degree of BBB penetration; there was a significant overlap of both distributions (CNS– and CNS+)[33] with log(Cbrain/Cblood) values of −1 and 0 (BBB classification of StarDrop). These simulations suggested that K-604 might penetrate the BBB and potentially be delivered to the brain. Next, we evaluated the BBB permeability of K-604 using the in vitro BBB Kit.

In Vitro Evaluation of the BBB Permeability of K-604

The apparent permeability coefficient (Papp) (21.9 × 10–6 cm/s) of K-604 across the BBB was lower than that of carbamazepine (47.8 × 10–6 cm/s), which was used as a positive control agent in the BBB Kit evaluation system. The results of the experiment examining BBB permeability were almost consistent with the aforementioned simulation (carbamazepine BBB permeability = log(Cbrain/Cblood) = −0.063 to −0.181). Regarding the actual feasibility of oral administration, the delivery efficiency of K-604 is not expected to be as high as that of carbamazepine, and systemic adverse effects may result from high doses. Therefore, we aimed to develop an alternative drug delivery to replace oral administration and focused on intranasal administration to noninvasively deliver therapeutic drugs from the nose to the brain along the olfactory and trigeminal pathways, bypassing the BBB. Although several papers have introduced the potential utility of this method,[20−22] few cases have successfully applied intranasal administration of a drug to mice or rats and demonstrated sufficient exposure in the brain. Before performing intranasal administration, innovative modifications were required to obtain a high K-604 solubility in aqueous solution to ensure a high exposure level in the brain. We noted the utility of the chemical structures of HA,[46] GA, and CA as shown in Figure .

Figure 2

Chemical structures of hyaluronic acid, d-gluconic acid, and citric acid.

These hydroxycarboxylic acids enable the strong counterionic interaction between the carboxylate group and the amine moiety of K-604. Furthermore, the multiple hydroxyl groups allow for strong interaction with water, playing roles in solvating and hydrating the transamine moiety of K-604 and resulting in increased aqueous solubility. Accordingly, the solubility of K-604 could be enhanced despite the acidic pH and released from strong acidic invasion into the olfactory epithelial tissue. We next investigated the aqueous solubility of K-604 in the presence of these additives.

Solubility Studies on K-604 in Different Vehicles

The K-604 concentrations in different solution formulations are presented in Table .

Table 2

K-604 Solubility in Different Vehicles

vehicle	additive	K-604 solubility	pH
water		0.05 mg/mL	6.8
0.01 N HCl		3.24 mg/mL	2.3
0.05 N HCl		19 mg/mL	1.3
HA	5 mg	10.8 mg/mL	3.8
GA	4 mg (0.04 M)	10.8 mg/mL	3.6
CA	7.7 mg (0.04 M)	10.8 mg/mL	3.0

Although an extremely high aqueous solubility (19 mg/mL) of K-604 has been reported at pH 1.2,[17] this solution formulation is very invasive to nasal epithelial tissues because of its corrosive acidity. When using a 0.01 N HCl solution, the acidity was somewhat weakened to 2.3, but the solubility decreased to 3.24 mg/mL. The pH values of the aqueous K-604 solutions containing HA, GA, or CA were determined to be 3.8, 3.6, and 3.0, respectively. Notably, all three additives resulted in the same solubility (K-604 10.8 mg/mL, 0.02 M). As expected, these hydroxycarboxylic acids, namely, HA, GA, and CA, were found to be favorable solvation additives.

In Vivo Monitoring of Stained Tissues along the Intranasal Route

To monitor the drug delivery route from the nose to the brain along olfactory nerve bundles, we intranasally administered uranine to mice As shown in Figure , yellow-colored tissues (group A1, No. 1) stained by uranine were visually observable, while the group administered with 0.01 N HCl was not stained (group A2, No. 1).

Figure 3

Sliced yellow-colored tissues of the olfactory bulb, cerebellum, and cerebrum stained by uranine, which contrasted with the unstained tissue treated with a 0.01 N HCl solution.

This result visually indicated the existence of a drug delivery route from the nose to the brain. Before intranasally administering the K-604 solution formulation to mice, we investigated the scope and limitation of drug delivery to the brain by oral administration.

In Vivo PK Study after Single Oral Administration (K-604, 6 mg/kg) to Mice

The maximum plasma concentration (Cmax) of K-604 was 6.2 ng/mL at 15 min after administration. The time of peak concentration (Tmax) in the cerebrum and cerebellum tissue was 15 min after administration, and the Cmax values were 0.8 and 0.7 ng/g, respectively. These concentrations were less than the lower limit of quantification (LLOQ) at 60 min after treatment administration. It is noteworthy that the magnitude of these Cmax values was markedly different from the plasma Cmax value (6.2 ng/mL). The drug concentrations in the olfactory bulb tissue were less than the LLOQ at all points as shown in Figure .

Figure 4

Drug concentration–time profiles in the plasma, cerebrum, cerebellum, and olfactory bulb tissue after single oral administration (K-604, 6 mg/kg) to female mice. Each point represents the mean ± SD (n = 5). The LLOQ of the cerebrum and cerebellum samples was 0.5 ng/g, and the LLOQ of the olfactory bulb tissue sample was 1 ng/g.

Table illustrates the PK parameters after single oral administration of K-604 to mice at a dose of 6 mg/kg (166 μg of K-604/mouse). The AUC of the cerebral tissue was 8.9 ng·min/g, whereas that of the plasma was 197.5 ng·min/mL. Interestingly, the feasibility of drug detection in the cerebral tissue sample was well supported by not only the BBB permeability predictions of StarDrop and ADMET Predictor but also by evaluation using the in vitro BBB Kit. These tools were useful for estimating BBB permeability before conducting an in vivo PK study. Accordingly, we concluded that oral administration inefficiently delivered K-604 to the brain in practical applications.

Table 3

PK Parameters after Single Oral Administration of K-604 to Mice at a Dose of 6 mg/kga

single oral administrationat a dose of 6 mg/kg	plasma	cerebrum	cerebellum	olfactory bulb
Tmax (min)	15	15	15	N.C.
Cmax (ng/mL or ng/g)	6.2 ± 3.1	0.8 ± 0.4	0.7 ± 0.3	0 ± 0
T1/2 (min)	N.C.	N.C.	N.C.	N.C.
AUC0–t (ng·min/mL or ng·min/g)	197.5 ± 40.7	8.9 ± 7.4	5.4 ± 4.9	0 ± 0

N.C. not calculated. Data are expressed as mean ± SD (n = 5). The LLOQ of the cerebrum and cerebellum samples was 0.5 ng/g, and the LLOQ of the olfactory bulb tissue samples was 1 ng/g.

In Vivo PK Study after Single and Repeated Intranasal Administration of a 0.01 N HCl Solution (50 μL) of K-604 to Mice

Figure illustrates the PK profiles of plasma, cerebrum, and olfactory bulb tissue after single and repeated (7 days) intranasal administration of K-604 (162 μg/50 μL) to mice. With single administration, the Cmax values of the plasma, cerebrum, and olfactory bulb tissues were 574 ng/mL, 66 ng/g, and 406 ng/g, respectively. With repeated administrations, the Cmax values of the plasma, cerebrum, and olfactory bulb tissues were 870 ng/mL, 83 ng/g, and 285 ng/g, respectively. As shown in the overlapping drug concentration–time curves of the single and repeated administration effects, there is almost no difference between the two drug concentration–time profiles. We concluded that intranasal administration represents not only a stable and reproducible drug delivery method but also a safe method because the drug can be quickly eliminated from the tissues without side effects due to accumulation. In the initial time range (after 5 to 45 min) after administration of 50 μL of K-604, the drug concentrations (870, 449, 250, and 77 ng/mL) in the plasma were approximately 6- to 10-fold higher than those (83, 44, 32, and 12 ng/g) in the cerebral tissue. In fact, the discrepancy between the plasma AUC (20680 ng·min/mL) and the olfactory AUC (4806 ng·min/g) significantly expanded. This result suggested that the K-604 solution formulation might overflow from the olfactory route to the lung tissue and digestive tract and then absorb into and travel through the bloodstream to reach all parts of the body and partially reach the cerebral tissues. To suppress the overflow of the solution from the nasal route into the other tissues, we minimized the volume of the K-604 solution formulation from 50 to 10 μL and intranasally administered the reduced volume.

Figure 5

Drug concentration–time profiles in the plasma, cerebrum, and olfactory bulb tissue after single and repeated (7 days) intranasal administration of a 0.01 N HCl solution (K-604, 162 μg/50 μL) to mice. Each point represents mean ± SD (n = 5). The LLOQs of the cerebrum and olfactory bulb tissue samples were 0.5 and 1 ng/g, respectively.

In Vivo PK Study after Repeated Intranasal Administration of a 0.01 N HCl Solution (10 μL) of K-604 to Mice

Even with the reduced volume (32.4 μg/10 μL) of the 0.01 N HCl solution, the drug concentrations (21, 8.9, 4.5, 2.1, 0.8 ng/g) in the cerebral tissue could be clearly detected at 5, 15, 30, 45, and 60 min after intranasal administration (Figure ). The Cmax of K-604 (156 ng/g) in the olfactory bulb tissue exceeded that of K-604 (146 ng/mL) in the plasma. In the initial time range (from 5 to 45 min) after administration of 10 μL of K-604, the drug concentrations (142, 90, 33, and 10 ng/mL) in the plasma were slightly higher than those (156, 51, 15, and 6 ng/g) in the olfactory bulb tissue. Actually, the discrepancy between the plasma AUC (2989 ng·min/mL) and the olfactory AUC (2113 ng·min/g) markedly reduced. Thus, these results suggested that the overflow of the K-604 solution formulation from the olfactory route into the other tissues could be suppressed and improved. The cerebrum AUC was 367 ng·min/g with an intranasal dose of 32.4 μg per mouse, whereas the cerebrum AUC was 9 ng·min/g with an oral dose of 166 μg per mouse. Here, we demonstrated that intranasal administration efficiently delivered K-604 to the brain.

Figure 6

Drug concentration–time profiles of the plasma, cerebrum, and olfactory bulb tissue after repeated intranasal administration of a 0.01 N HCl solution (K-604, 32.4 μg/10 μL) to mice for 7 days. Each point represents mean ± SD (n = 5). The LLOQs for the cerebrum and olfactory bulb tissue samples were 0.5 and 1 ng/g, respectively.

Figure illustrates the drug concentration ratios of cerebrum-to-plasma and olfactory bulb to plasma (Kp value) as a function of time after repeated intranasal administration of K-604 (10 and 50 μL). In the olfactory bulb to plasma ratio, the Kp value of the 10 μL volume (1.18) was higher than that of the 50 μL volume (0.32) at 5 min after administration. In the initial time range (after 5 to 45 min), the drug concentrations in the olfactory bulb tissue were 3- to 7-fold and 1- to 3-fold higher than those in the cerebral tissue, respectively. These results revealed that a small volume (10 μL/mouse) could be more efficient than a large volume (50 μL/mouse). We established that K-604 could be efficiently delivered along the olfactory route to the brain via intranasal administration while minimizing systemic circulation in the lung and digestive tract.

Figure 7

Cerebrum and olfactory bulb to plasma drug concentration ratios as a function of time after 7 days of repeated intranasal administration (10 and 50 μL) to mice.

In Vivo PK Study after Single Intranasal Administration of Hydroxycarboxylic Acid Solution (10 μL) of K-604 to Mice

To further increase the drug exposure in the cerebral tissue, we tested highly concentrated solutions of K-604, which had solubilities that were increased by approximately 3-fold with the addition of HA, GA, or CA.

Figure illustrates the PK profiles of the plasma, cerebrum, and olfactory bulb tissue after single intranasal administration of the K-604 solution formulations containing HA, GA, or CA (108 μg/10 μL K-604) to mice. At 5 min after intranasal administration of 108 μg of K-604 per mouse, the Cmax values of the plasma were 274, 217, and 283 ng/mL, respectively, and the Cmax values of the cerebral tissue were 39, 32, and 47 ng/g, respectively. Additionally, the AUCs of the plasma with these formulations were 10220, 7263, and 9891 ng·min/mL, respectively, and those of cerebral tissue were 791, 579, and 772 ng·min/g, respectively.

Figure 8

Drug concentration–time profiles of plasma, cerebrum, and olfactory bulb tissue after single intranasal administration of HA, GA, or CA solution (108 μg/10 μL K-604) to mice. Each point represents mean ± SD (n = 5). The LLOQs of the cerebral and olfactory bulb tissue samples were 0.5 and 1 ng/g, respectively.

Table illustrates the PK parameters of the plasma, cerebrum, and olfactory bulb tissue after intranasal administration of the 0.01 N HCl, HA, GA, or CA solution formulation (10 μL) to mice. The Cmax values of the cerebrum and olfactory bulb tissues with the CA solution formulation were somewhat higher than those with the HA and GA solution formulations. To assess the brain-targeting efficiency (BTE) of nasal and oral administration, the BTE index of the cerebrum Cmax and AUC at an intranasal dose of 32.4 or 108 μg per mouse was calculated on the basis of the cerebrum Cmax and AUC at an oral dose of 166 μg of K-604 per mouse according to the following equations:

Table 4

PK Parameters after Intranasal Administration of the 0.01 N HCl, HA, GA, or CA Solution Formulation (10 μL) to Mice

parameter	plasma	cerebrum	olfactory bulb
10 μL 0.01 N HCl Repeated
Tmax (min)	5	5	5
Cmax (ng/mL or ng/g)	146 ± 20	21 ± 10	156 ± 131
T1/2 (min)	14.7 ± 1.1	15.6 ± 4.3	7.93
AUC0–t (ng· min/mL or ng·min/g)	2989 ± 395	367 ± 121	2113 ± 1380
10 μL HA Solution Single
Tmax (min)	5	15	15
Cmax (ng/mL or ng/g)	274 ± 30	39 ± 11	274 ± 100
T1/2 (min)	17.4 ± 2.1	N.C.a	N.C.a
AUC0–t (ng·min/mL or ng·min/g)	10220 ± 2767	791 ± 160	5200 ± 1036
10 μL GA Solution Single
Tmax (min)	5	5	5
Cmax (ng/mL or ng/g)	217 ± 51	32 ± 7	160 ± 24
T1/2 (min)	16.6 ± 2.9	N.C.a	N.C.a
AUC0–t (ng·min/mL or ng·min/g)	7263 ± 2138	579 ± 137	3406 ± 2127
10 μL CA Solution Single
Tmax (min)	5	5	5
Cmax (ng/mL or ng/g)	283 ± 121	47 ± 29	351 ± 112
T1/2 (min)	13.4 ± 1.7	N.C.a	N.C.a
AUC0–t (ng·min/mL or ng·min/g)	9891 ± 2556	772 ± 223	3976 ± 642

N.C. not calculated. Parameters were calculated from the average concentrations of five mice. Data are expressed as mean ± SD (n = 5). The LLOQs of the cerebral and olfactory bulb tissue samples were 0.5 and 1 ng/g, respectively.

Interestingly, the 0.01 N HCl solution showed the highest BTE index values in the Cmax and AUC categories among all solution formulations (Table ). The AUC BTE indices of the hydroxycarboxylic acid solutions were more than 100-fold but lower than that of the 0.01 N HCl solution. To our knowledge, this result was the first clear output to demonstrate the concrete feasibility of intranasally administering the drug.[26,47−52] The Cmax BTE indices of all the hydroxycarboxylic acids (HA, GA, and CA) solution formulations were relatively moderate and scattered compared to that of the 0.01 N HCl solution. We speculated that the hydroxycarboxylic acids might be prone to partial crystallization from the saturated solutions in the process of absorption from the mucous layer, epithelial membrane, and junctional barrier along the olfactory route. Moreover, the hydroxycarboxylic acid solution formulations precipitated as crystals after the solutions were allowed to stand for an entire day and night at room temperature (10–20 °C). The CA solution formulation was superior to the other solution formulations in both the Cmax and AUC BTE categories. Our final interest was to evaluate whether K-604 has the therapeutic potential to alter lipid profiles in the brain even with a short exposure time. Therefore, we selected the CA solution formulation to intranasally administer K-604 to mice for 7 days and measured the brain cholesterol level as an indicator of the pharmacological efficacy of this ACAT-1 inhibitor.

Table 5

Comparison of the BTE Indices with Each Solution Formulation

BTE index	0.01 N HCl	HA	GA	CA
intranasal Cmax BTE	134	75	61	90
intranasal AUC BTE	211	137	100	133

Brain Lipid Profiles after Intranasal Administration of K-604 to Mice for 7 Days

Notably, K-604 dramatically decreased the CE levels from 0.70 to 0.04 μmol/g in the brains of mice even after a single daily dose administration for 7 days (Figure ). This was an outstanding effect, that is, 94% reduction in the CE level in vivo using only a trace amount of K-604, that we never observed in other tissues, including the systemic plasma, liver, adrenal, and intestine, when evaluating ACAT inhibitors in atherosclerosis models.[17] In the brain, a large pool of cholesterol exists as free cholesterol in the myelin sheath to support the saltatory conduction of action potentials, and a small portion of cholesterol is found in the plasma and subcellular membranes of neurons and glial cells. Because myelin has a very slow turnover rate, myelin-associated cholesterol is virtually immobilized. Brain cholesterol also plays roles in controlling synapse formation.[53] Brain cholesterol cannot be supplied from systemic plasma cholesterol because the BBB prevents cholesterol transportation between the blood and the brain. Therefore, brain cholesterol is locally produced, and its level is maintained by converting surplus cholesterol (1) into 24S-OHC (3) via the enzyme cholesterol 24-hydroxylase (CYP46A1). 24S-OHC can be excreted from the brain to the systemic circulation through the BBB. A fraction of CE (2) was also present in the brain. This cholesterol metabolism[10] is depicted in Figure .

Figure 9

Change in total cholesterol (TC) and CE levels in the brain after intranasal administration of the CA solution (108 μg/10 μL K-604) to mice with a single daily dose for 7 days. The TC and CE levels represent the average of five mice.

Figure 10

Metabolism from cholesterol (1) to cholesteryl ester (2) and 24(S)-hydroxycholesterol (3) in the brain.

Focusing on CEs, lipid droplets are subcellular organelles that store large amounts of neutral lipids, triglycerides, and/or CEs and are interestingly found in some aggressive tumor tissues. Geng et al. have recently reported that lipid droplets are prevalent in GBM, the most malignant glioma, but are not detectable in benign brain tumors and normal brain tissues.[54,55] The authors suggested that the suppression of CE production by ACAT-1 inhibitors should be a therapeutic target in GBM. Furthermore, Bryleva et al. reported that ablation of the ACAT-1 gene in an AD model mouse ameliorates amyloid pathology by increasing 24S-OHC content in the brain.[56] The physiological range of 24S-OHC has been shown to have an inhibitory effect on Aβ production.[57] Because intranasal administration of K-604 demonstrated efficacious ACAT-1 inhibition even in a short exposure period, less than 2 h, K-604 should provide promising results for the treatment of GBM and AD.[58]

Histological Evaluations of Respiratory and Olfactory Epithelium and Lung (with Bronchus) after Intranasal Administration of K-604 Using Acidic Solutions

We evaluated the histological findings of respiratory and olfactory epithelium (Figures and 12) and lung with bronchus after intranasal administration of the acidic solutions of K-604 with a single daily dose for 7 days.

Figure 11

Hematoxylin and eosin staining of respiratory epithelium (bar = 50 μm). The observation grades (−: normal, ±: minimal, 1+: slight)[59] were comprehensively determined by the depth degree and number of desquamations that progressed from degeneration and necrosis, as shown by the representative histological findings regarding the respiratory epithelia from groups D1, D2, and B7.

Figure 12

Hematoxylin and eosin staining of olfactory epithelium (bar = 50 μm). The observation grades (−: normal, 0: within normal limits, ±: minimal, 1+: slight, 2+: moderate) were comprehensively determined by the depth degree and number of desquamation that progressed from degeneration and necrosis as shown by the representative histological findings regarding the olfactory epithelia from groups D1, B17, and B7.

As anticipated, intranasal administration of 50 μL of the 0.01 N HCl solution formulation caused an overflow from the olfactory route to the lung tissue and induced interstitial inflammation and basophilic bronchiole epithelia (Figure ). However, administration of a minimized volume (10 μL) made it possible to avoid the abovementioned adverse effects. Even more surprisingly, the olfactory epithelia (from five mice) were not injured at all despite the strong acidic medium (pH 2.3) as shown in Table (group B17). We presumed that the mucus layer might play a role in protecting the direct invasion of epithelial cells by a trace quantity of HCl solution. Importantly, drug absorption did not proceed or accelerate as a result of olfactory epithelium disruption and was simply followed by absorption via the paracellular, transcellular, and/or neuronal pathway.

Figure 13

Hematoxylin and eosin staining of the lung (with bronchus).

Table 6

Histological Evaluation of the Respiratory Epithelium, Olfactory Epithelium, Lung (with Bronchus), Basophilic of Bronchiole Epithelium, and Interstitial Inflammation after Intranasal Administration of the 0.01 N HCl Solution Formulations (50 and 10 μL) to Mice

group number	group B1					group B7					group B9					group B11					group B17
time	0					120 min					360 min					0 min					120 min
vehicle	0.01 N HCl					0.01 N HCl					0.01 N HCl					0.01 N HCl					0.01 N HCl
test						K-604					K-604					K-604					K-604
conc. (mg/mL)	0					3.24					3.24					0					3.24
volume (μL/mouse)	50					50					10					10					10
observation grade	0	±	1+	2+	3+	0	±	1+	2+	3+	0	±	1+	2+	3+	0	±	1+	2+	3+	0	±	1+	2+	3+
respiratory epithelium (degeneration/necrosis/desquamation)	0	3	2	0	0	0	0	5	0	0	0	0	5	0	0	0	4	1	0	0	0	4	1	0	0
olfactory epithelium (degeneration/necrosis/atrophy)	2	1	1	1	0	0	1	2	2	0	0	2	2	1	0	5	0	0	0	0	5	0	0	0	0
lung (with bronchus) (degeneration/necrosis/atrophy)	4	0	1	0	0	1	2	2	0	0	1	4	0	0	0	5	0	0	0	0	5	0	0	0	0
basophilic of bronchiole epithelium	4	0	1	0	0	0	1	4	0	0	1	2	2	0	0	5	0	0	0	0	5	0	0	0	0
interstitial inflammation	5	0	0	0	0	0	1	3	1	0	3	0	2	0	0	5	0	0	0	0	5	0	0	0	0

Although the acidic pH values of the hydroxycarboxylic acid solution formulations (pH 3.0 to 3.8) were weaker than that of the 0.01 N HCl solution (pH 2.3), they unexpectedly resulted in minimal (±) and slight (1+) invasion observation grades in both epithelial tissues assessed (Table ). To elucidate the epithelial invasion of these formulations, we investigated the relationships between the thickness (shrinkage damage) of the olfactory epithelium[60,61] and the drug concentration of K-604 in each tissue at 120 min after intranasal administration for 7 days (Supporting Information). Hence, we verified the working hypothesis that the drug concentration of K-604 in the olfactory bulb could be markedly increased by decreasing or disrupting the thickness of olfactory epithelial cells, leading to enhanced drug permeability. In the scatter plots of the three hydroxycarboxylic acid solution groups (D2-HA, D3-GA, and D4-CA), the relationships between the olfactory epithelial thickness (μm) and the drug concentration of K-604 (ng/g) are presented (Figure ).

Table 7

Histological Evaluation of Respiratory Epithelium and Olfactory Epithelium after Intranasal Administration of the HA, GA, or CA Solution Formulation (10 μL) to Mice

group number	group D1					group D2					group D3					group D4
time	0					120 min					120 min					120 min
vehicle						HA					GA					CA
test						K-604					K-604					K-604
conc. (mg/mL)	0					10.8					10.8					10.8
volume (μL/mouse)	0					10					10					10
observation grade	–	±	1+	2+	3+	0	±	1+	2+	3+	0	±	1+	2+	3+	0	±	1+	2+	3+
respiratory epithelium (degeneration/necrosis/desquamation)	5	0	0	0	0	0	4	1	0	0	0	5	0	0	0	0	4	1	0	0
olfactory epithelium (degeneration/necrosis/atrophy)	5	0	0	0	0	2	2	1	0	0	3	2	0	0	0	3	2	0	0	0

Figure 14

Relationships between the olfactory epithelial thickness (μm) and the drug concentration of K-604 (ng/g) in the hydroxycarboxylic acid solution groups (D2-HA, D3-GA, and D4-CA).

As shown in Figure , the three hydroxycarboxylic acid solution formulations interestingly distributed a variety of drug concentrations irrespective of the olfactory epithelial thickness. The characteristic results of the three vehicles are summarized as follows. (1) The HA formulation (D2-HA group) fell into the relatively higher drug concentration category. (2) The CA formulation (D4-CA group) fell into the significantly lower drug concentration category despite showing the highest Cmax of the olfactory bulb among the three groups. (3) The GA formulation (D3-GA group) formed a population at the median in both drug concentration and epithelial thickness categories. Given the reasons for these physiologically different results, HA played a role in retaining the drug in the nasal cavity for a longer period, as reported in the literature.[62,63] In contrast, CA seemed to behave as an absorption enhancer[64] and thus might reversibly modify the structure of the epithelial barrier. We assume that GA, which has multiple hydroxyl groups, can strongly interact with multiple alcohols and the amine moiety of K-604 to develop a tight GA and K-604 complex, thus minimizing its range of distribution. Pujara et al. investigated the effect of pH, osmolarity, type (acetate, adipate, citrate, and phosphate), and concentration of buffers on the nasal mucosal epithelium in rats using an in situ nasal perfusion technique.[65] In their report, phosphate buffers (0.07 M) with pH values between 3 and 10 exerted very low or essentially similar impacts, while buffers with pH values above 10 and below 3 seemed to result in both membrane and intracellular damage. Furthermore, the effect of buffer concentrations on the rat nasal mucosa was studied using acetate buffer (pH 4.75) at three different concentrations (0.07, 0.14, and 0.21 M). The damage to the nasal mucosa by acetate buffers was concentration-dependent. When considering the above results together, we suspect that a high drug concentration (108 μg/10 μL K-604) of the hydroxycarboxylic acid solution formulation might damage the respiratory and olfactory epithelial cell integrity more than a low concentration of the 0.01 N HCl solution (32.4 μg/10 μL K-604) irrespective of the acidic pH of vehicles. As a dramatic reduction (94%) of the brain CE level was attained during a short period, even a low concentration of K-604 could be pharmacologically expected and would be clinically promised.

Conclusions

The present study revealed several issues that we raised in the Introduction. Intranasal administration has long been expected to be a potentially useful method to deliver therapeutic agents to the brain. However, until now, there has been no substantially good example of this method. In our study, we first clarified the olfactory route used to deliver the drug from nose to the brain by observing the yellow-stained tissues from the olfactory bulb to the cerebellum and cerebrum using uranine. Second, we conducted a simulation of the brain penetration ratio of K-604 using the software StarDrop and ADMET Predictor and evaluated the drug BBB permeability of K-604 using an in vitro BBB Kit. The simple prediction of the BBB permeability was consistent with the in vitro BBB permeability and was a useful tool before the conduction of in vivo experiments. Third, we succeeded in improving the aqueous solubility of K-604, making use of hydrogen bonding using hydroxycarboxylic acid solutions as vehicles. Fourth, we demonstrated that intranasal administration enhanced the drug delivery efficiency as measured by an AUC that was by approximately 100- to 211-fold higher than that achieved with oral administration. Fifth, K-604 was very effectively targeted to the mouse brain where it decreased CE levels from 0.70 to 0.04 μmol/g with only a single daily dose for 7 days. Sixth, we clarified that the olfactory epithelium damage was not directly related to the enhancement of K-604 delivery to the olfactory bulb and cerebrum. Based on the above results, we have established a potential intranasal administration method to efficiently deliver K-604 into the brain for the treatment of GBM and AD. We still need to explore the feasible dosage of K-604 with a combination of optimal vehicles in practical applications in the future.