Intranasal Diazepam Outperforms Midazolam and Lorazepam in Early Seizure Control in Pilocarpine-Induced Mice Model
Anson Hei-Ka Tong, Jiajia Zhao, Jiahao Li, Weikang Ban, Rilla Jiansu Wang, Yuning Xie, Yufeng Zhang, Zhong Zuo
TL;DR
This study finds that intranasal diazepam is more effective than midazolam and lorazepam for early seizure control in a mouse model of epilepsy.
Contribution
The study is the first to compare intranasal benzodiazepines in a pilocarpine-induced seizure model and identifies diazepam as the most effective.
Findings
Intranasal diazepam showed the highest brain-to-plasma ratio and best seizure control at 10 minutes.
Diazepam reduced TNF-α mRNA and protein levels more effectively than midazolam and lorazepam.
All three drugs had similar brain binding and receptor affinity, but diazepam had the lowest plasma protein binding.
Abstract
Epilepsy is a central nervous system disease characterized by the sudden onset of seizures, loss of consciousness, or confusion. In recent years, many non-intravenous routes of administration of benzodiazepines have been developed for seizure control, with intranasal administration being an attractive route of choice. However, for such a route of administration, there is a lack of evidence on the choice of the epilepsy drug. This study aims to compare the intranasal formulations of three first-line drugs for seizure control, namely midazolam, diazepam, and lorazepam, via an ideal intranasal treatment. A pilocarpine-induced seizure model in mice was used to compare drug efficacy. The three drugs were administered intranasally to 36 C57 mice at a single dose of 1 mg/kg, followed by inducing the seizures via intraperitoneal injection of pilocarpine. The subsequent seizure scores were…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsNeuroscience and Neuropharmacology Research · Epilepsy research and treatment · Sleep and Wakefulness Research
Introduction
Epilepsy is a central nervous system (CNS) condition characterized by a sudden onset of neurological symptoms ranging from seizures and loss of consciousness to confusion. It can be especially disabling for a patient in their everyday life when a sudden epileptic episode occurs. Defined as 5 min or more of continuous clinical and/or electrographic seizure activity, or recurrent seizure activity without recovery between seizures, status epilepticus (SE) is a medical emergency that requires prompt intervention. Serious complications may follow an episode of SE, including cardiac, respiratory, and neurological problems. If uncontrolled, it may even progress to prolonged refractory status epilepticus that could cause permanent damage to the brain [1]. Currently, benzodiazepines, including midazolam, diazepam, and lorazepam are commonly indicated in first-line treatments of status epilepticus.
For the treatment of status epilepticus in a hospital setting, the intravenous route of administration is often the best choice. However, outside the hospital, in situations where intravenous access is not possible, alternative administration routes may be utilized, such as rectal, buccal, intramuscular, intraosseous, or intranasal routes. Early recognition and treatment in prehospital or out-of-hospital settings are crucial for managing status epilepticus [2]. In recent years, there has been an increasing spotlight on intranasal (IN) administration as an alternative route for anti-epileptic drugs. Most notably, the FDA approved intranasal formulations of midazolam and diazepam in May 2019 and January 2020, respectively. Both approved intranasal formulations are indicated for “acute treatment of seizure clusters, acute repetitive seizures that are distinct from a patient’s usual seizure pattern in patients with epilepsy” [3, 4]. Clinically, this indication underscores the importance of “early intervention”, treating the initial signs of a seizure cluster to prevent progression to status epilepticus [5]. Intranasal administration is favorable in acute epileptic episodes due to its ease of use, especially in out-of-hospital environments. Compared to intramuscular administration, it requires less training [6] and facilitates fast absorption of drugs [7]. Compared to rectal administration, it is more acceptable to patients, particularly adults. Compared to buccal administration, nasal administration avoids problems such as excessive buccal solutions or vomiting [8]. Intranasal delivery of CNS drugs has the benefit of direct and rapid delivery to the brain through extensive blood networks in the nasal mucosa, the olfactory pathway and the trigeminal nerve pathway [9]. Compared to oral administration, intranasal administration also avoids first-pass metabolism. The respective metabolisms of midazolam and diazepam produce active metabolites. Diazepam produces an active metabolite, N-desmethyldiazepam, which has an extended half-life of around 100 h. In contrast, midazolam produces an active metabolite of 1-hydroxy-midazolam [10]. Therefore, direct and rapid delivery from the nose to the brain could lead to lower required drug dosages due to higher bioavailability [11].
Previous studies have compared the onset of intravenous (IV) benzodiazepines from a pharmacodynamic perspective and suggested comparable efficacy of IV lorazepam and IV diazepam for seizure cessation [12–14]. Similarly, there are no clinical differences between the effects of IV lorazepam and IV midazolam, and IV midazolam and IV diazepam for seizure cessation [14]. Comparisons have also been made between different routes of administration for benzodiazepines. IN midazolam has been compared with rectal diazepam [15–17] and IV diazepam [18–20]. Between IN midazolam and rectal diazepam, IN midazolam performed better in seizure cessation [15, 16]. Between IN midazolam and IV diazepam, the time from drug administration to seizure cessation was comparable if the time for intravenous insertion was excluded, however, since additional time was needed to establish an IV line, IN midazolam was therefore more favorable for seizure control from the time of onset of seizure to cessation [18, 19]. In another study, intranasal lorazepam was found to have a similar seizure control effect to that of IV lorazepam [21]. While pharmacodynamic studies indicated similar drug effects, some pharmacokinetic comparisons have suggested differences in onset. Studies that investigated the uptake of IN midazolam and IN diazepam showed that midazolam reached the maximum plasma concentration (T_max_) at a faster rate than diazepam [22]. Between IV diazepam and IV lorazepam, the former was taken up into the brain at a faster rate than the latter [23, 24]. Comparison of the T_max_ for intramuscular administrations of these three drugs from different studies suggested that midazolam reaches plasma T_max_ the quickest, followed by diazepam and lorazepam [25].
Although many studies compared different modes of administration of benzodiazepines for seizure control, limited studies directly compare the intranasal delivery of these three most used benzodiazepines. Therefore, the current study is designed to focus on the comparison of intranasal formulations of the three first-line drugs for seizure control, namely midazolam, diazepam, and lorazepam using a pilocarpine-induced acute status epilepticus (SE) model. Crucially, we employed prophylactic administration design for both clinical and experimental reasons. Clinically, this mimics the early intervention strategy for managing seizure clusters or prodromal auras, where intranasal benzodiazepines are administered on demand to prevent the progression of early symptoms into status epilepticus [5]. Experimentally, this design is necessitated by the technical constraints of the intranasal route. Unlike parenteral injections, intranasal delivery requires to temporarily restrict the activity range of animal. Drug administration during active generalized convulsions would introduce significant dosing variability and aspiration risks. Moreover, given the rapid mortality associated with this model [26], prophylactic dosing can prevent the lethal outcomes. We aim to compare the efficacies of these three intranasally administered drugs using a pilocarpine-induced seizure mice model to propose a recommendation for the ideal intranasal treatment for acute seizure control.
Materials
Chemicals and Solvents
Midazolam, diazepam and lorazepam were purchased from the USP Pharmacopeia (USA). Polyethylene glycol 400 and Propylene glycol were purchased from Merck Limited, an affiliate of Merck KGaA, Darmstadt, Germany. Scopolamine methyl bromide and pilocarpine hydrochloride were purchased from Sigma-Aldrich Chem. Co. (Milwaukee, WI, USA). Formic acid was supplied by BDH laboratory Supplies Ltd. (Kampala, Ukraine). Phosphate buffer and DMSO were supplied by Sigma-Aldrich Chem. Co. (Milwaukee, WI, USA). The internal standard, Berberine, was purchased from MeilunBio (Dalian). Acetonitrile was HPLC grade and obtained from Merck (Darmstadt, Germany). All reagents and chemicals were analytical grade and used without further purification. Distilled and deionized water was prepared from the Millipore water purification system (Millipore, Milford, MA, USA).
Animals
Male C57 mice with body weights between 20 and 30 g were provided by the Laboratory Animal Services Centre at the Chinese University of Hong Kong. The mice were housed in a controlled environment (25 ± 2 °C; 50% ± 5% humidity; 12-h dark-light cycle) and were provided with a standard diet, allowing them to drink water ad libitum. Throughout the study, every procedure was executed in strict accordance with the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health in USA [27]. The animal experiments were conducted under the approval of the Animal Ethics Committee of the Chinese University of Hong Kong (Animal Ethical Approval number: 21–053-MIS).
Methods
Selection of Clinically Relevant Drug Doses in Mice
The same drug concentration was prepared for midazolam, diazepam, and lorazepam tests. Literature and market-available doses of the three drugs were reviewed, and the smallest average dose among these three drugs is 5 mg for a 60 kg human. Based on the scale factor of 12.3 for human to mice recommended by the FDA [28], a drug dose of 1 mg/kg was investigated on mouse models. Given a body weight of 0.03 kg per mouse, 0.03 mg per mouse was needed. Each mouse was to receive a volume of no more than 0.01 ml of drug per nostril, so a concentration of 3 mg/ml was prepared for midazolam, diazepam, and lorazepam.
Preparation of Midazolam, Diazepam, Lorazepam, Scopolamine Methyl Bromide, and Pilocarpine Hydrochloride for Mouse Administration
Midazolam, diazepam, and lorazepam solution were prepared by dissolving 3 mg of the powdered drug into 200 µl polyethylene glycol 400 (PEG 400) followed by sonicating for 15 min, and subsequent addition of 200 µl of propylene glycol (PLG) per time for four times to reach a 3 mg/ml solution for each tested drug. The prepared drug formulations were stored at 4 °C and protected from light. Based on previous report on successful induction of status epilepticus in the C57BL/6 mice [29], scopolamine methyl bromide, used for protection, and pilocarpine hydrochloride, used for induction of seizure, were prepared in saline to produce a solution of 0.15 mg/ml (to achieve a target dose of 1 mg/kg) and 52.5 mg/ml (to achieve a target dose of 350 mg/kg), respectively, for intraperitoneal (i.p.) injection.
Animal Grouping and Drug Treatment in the Pilocarpine-Induced Acute Seizures Mice Model
The mice were divided into nine groups to evaluate both the early-stage (Groups I to V) and late-stage (Groups VI to IX) anti-epileptic effects of the selected intranasal administered benzodiazepines. Specifically, mice in Group I (negative control group, n = 6) received Vehicle (intranasal, i.n.) followed by saline (intraperitoneal, i.p., at 5 min and 25 min post Vehicle dosing), while mice in Group II (seizure model control group, n = 6) received Vehicle (i.n.) followed by 1 mg/kg scopolamine methyl bromide (i.p.) and 350 mg/kg pilocarpine (i.p.). For those mice in treatment Group III (midazolam treatment group, n = 12), Group IV (diazepam treatment group, n = 12) and Group V (lorazepam treatment group, n = 12), they received 1 mg/kg midazolam (i.n.), 1 mg/kg diazepam (i.n.) and 1 mg/kg lorazepam (i.n.), respectively, followed by 1 mg/kg scopolamine methyl bromide (i.p.) and 350 mg/kg pilocarpine (i.p.). For the evaluation of late-stage anti-epileptic effects, mice in Group VI (seizure model control group, n = 6) received Vehicle (i.n.) followed by 1 mg/kg scopolamine methyl bromide (i.p.) and 350 mg/kg pilocarpine (i.p.), while mice in treatment Group VII (midazolam treatment group, n = 6), Group VIII (diazepam treatment group, n = 6) and Group IX (lorazepam treatment group, n = 6) received 1 mg/kg midazolam (i.n.), 1 mg/kg diazepam (i.n.) and 1 mg/kg lorazepam (i.n.), respectively, followed by 1 mg/kg scopolamine methyl bromide (i.p.) and 350 mg/kg pilocarpine (i.p.).
As shown in Fig. 1, two sets of mouse experiments were conducted to evaluate the early anti-epileptic effects and late anti-epileptic effects at 10 min (Groups II to V) and 100 min (Groups VI to IX) after seizure induction, respectively. According to the dosing schedule shown in Fig. 1, drugs/vehicle were administered intranasally to mice from all groups at 0 min, followed by i.p. administrations of scopolamine/saline at 5 min and pilocarpine/saline at 25 min. Intranasal administration was performed on conscious mice to avoid anesthetic interference [30]. Mice were manually restrained in a supine position. A total volume of about 10 µL was delivered into both nostrils (5 µL each) using a micropipette, with a 10–20 s interval to ensure absorption. After behavior test observation, mice were sacrificed for sample collection.
Fig. 1. Animal grouping and treatment for evaluation of early anti-epileptic effect of Groups I to V at 10 min (n = 6 ~ 12/group) and late anti-epileptic effect of Groups VI to IX at 100-minute (n = 6/group)
Measurement of Seizure Score
Seizure score was evaluated for early anti-epileptic effects and late anti-epileptic effects using a modified Racine’s scale [31]. Assessment was performed using a time-point sampling strategy. Specifically, the seizure behavior of each mouse was assessed in real-time at 10-minute time point for early anti-epileptic effects assessment and each 10-minute time point (immediately at 10, 20, 30. up to 100 min post-administration) for late anti-epileptic effects assessment as indicated in Fig. 1. A score of 0 describes normal non-epileptic activity. Score 1 is immobilization, 2 is head nodding and partial myoclonus, 3 is continuous whole-body myoclonus, 4 is rearing and tonic seizure, and 5 is death.
Plasma and Brain Sample Collection, Treatment, and Analyses Using LC-MS/MS
At the end of the experiment, mice were deeply anesthetized with an intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg) prior to sample collection. Once the animals were unresponsive to toe pinch reflexes, the thoracic cavity was opened to expose the heart. Blood samples were immediately collected via cardiac puncture for plasma analysis. Plasma was obtained by collecting blood samples in a centrifuge tube containing heparin, followed by centrifugation at 8000 rpm for 3 min. The plasma was then stored at −20 °C until analysis. Subsequently, a needle was inserted into the left ventricle, and an incision was made in the liver to allow venous drainage. The whole brain was obtained after intracardiac perfusion with 20 mL of ice-cold saline until the effluent ran clear to remove blood residue from the brain [32]. After removing the whole brains from the skulls, they were rinsed in saline and wiped with tissue paper to remove excess saline. For pharmacokinetic studies, the whole brain was collected and weighed followed by storing at −20 °C until analysis. For molecular analyses (mRNA and protein levels), the hippocampus and cortex were rapidly dissected on ice, frozen in liquid nitrogen and stored at −80 °C until further analysis. All whole brain tissues for pharmacokinetic studies were homogenized with 4 times their volume of water with 0.2% formic acid for diazepam and lorazepam, or phosphate buffer for midazolam, using an Omni homogenizer (Omni International Inc., Kennesaw, GA). To extract lorazepam, diazepam, and midazolam from brain homogenates and plasma samples, 50 µL of the collected samples was mixed with 150 µL of a 200 µg/mL berberine solution (internal standard, IS), followed by vortexing for 1 min. After centrifugation at 13,000 rpm for 10 min at 4 °C, an aliquot of 10 µL of supernatant was injected into the LC-MS/MS for analysis. The LC-MS/MS system consisted of an Agilent triple quadrupole mass spectrometer equipped with an electrospray ionization source (ESI), two Agilent 1290 series pumps, and an autosampler (Agilent Technologies Inc., Santa Clara, CA). Data acquisition and analysis were performed using Agilent MassHunter Quantitative Analysis software (version B.03.01). The four analytes, together with their internal standard, berberine, were chromatographically separated on a Waters ACQUITY UPLC BEH C_18_ (2.1 × 50 mm, 1.7 μm) column with gradient elution. The mobile phases, consisting of (A) H_2_O with 0.1% formic acid and (B) acetonitrile, were eluted at a flow rate of 0.2 ml/min. The gradient for lorazepam and midazolam started at 10% B from 0 to 1 min, increased to 75% B from 1 to 2 min, and was maintained at 75% B from 2 to 5 min. It was then decreased to 10% B from 5 to 6 min and maintained at 10% until 7 min. The temperature of the autosampler and analytical column was set at 8℃ and ambient, respectively. The gradient for diazepam started at 10% B from 0 to 1 min, increased to 65% B from 1 to 2 min, and was maintained at 65% B from 2 to 5 min, followed by a decrease to 10% B from 5 to 8 min. The detection was performed using the positive ESI in multiple reaction monitoring mode with precursor-to-product ion transitions at 336.1→320.1 (fragment/collision energy, 143 V/32 eV) for berberine (IS), 285.2→154.0 (125 V/26 eV) for diazepam, was 326.2→284.0 (150 V/30 eV) for midazolam and 321.1→275.0 (90 V/10 eV) for lorazepam. The linearity was R^2^ > 0.999 for diazepam (5.0–80.0 ng/mL), midazolam (5.0–160.0 ng/mL) and lorazepam (2.5–160.0 ng/mL) in both plasma and brain. The lower limit of quantification (LLOQ) was defined as the lowest concentration of the calibration standards. The accuracy was consistently within ± 15% of the nominal values (Mean accuracy as 99.8–100% with SD < 6%), confirming the reliability of the quantification method.
Tested Drug Stability in Plasma Protein/Brain Tissue and Their Binding Towards Plasma Protein/Brain Tissue Using the Ultrafiltration Method
About 399 µl of blank plasma or blank brain homogenate from C57 mice was spiked with 1 µl stock solution (10 mM drug in DMSO) to obtain a plasma or brain homogenate mixture containing 25 µM of each tested drug. Approximately 50 µL of these mixtures were incubated at 37 °C for 20 min, followed by concentration quantification to evaluate the drug stability in plasma or brain homogenate. Another 300 µL of such mixtures were incubated at 37 °C for 20 min, followed by addition to the top chamber of an ultrafiltration centrifuge tube. After centrifugation at 13,000 rpm for 5 min at 25℃, both top plasma/brain homogenate and bottom ultrafiltrate were collected for drug concentration analyses using the above-mentioned LC-MS/MS method for evaluation of drug binding to plasma protein or brain homogenate.
GABARα1, TNF-α and IL-1β mRNA Level Measurement
To evaluate the effect of the three tested drugs on the neuroinflammatory mRNA level, total RNA was extracted from the collected cortex and hippocampus using an animal total RNA isolation kit (FOREGENE, Chengdu, China) according to the manufacturer’s instructions. The total RNA was reverse-transcribed into cDNA using RT EasyTM II (with gDNase, FOREGENE, China). For quantitative analysis, reverse transcription–polymerase chain reaction (RT-PCR) was conducted using RT-PCR EasyTM-SYBR Green Ⅰ (FOREGENE) in the Light Cycler PRO system (Roche) under the following thermal cycling conditions: initial denaturation and enzyme activation at 95 °C for 3 min (one cycle), followed by 45 cycles of denaturation at 95 °C for 10 s, annealing at 65 °C for 10 s, and extension at 72 °C for 20 s. Subsequently, melting curve analysis (one cycle) was performed after PCR to ensure the specificity of the amplified products. Relative mRNA expression was normalised to GAPDH as an internal control to standardize the difference. The following primer sequences reported previously were used in the study:
GAPDH [33]: F: GGAGCGAGATCCCTCCAAAAT; R: GGCTGTTGTCATACTTCTCATGG, GABARα1 [33]: F: TGAGCACACTGTCGGGAAGA; R: CAGCAGTCGGTCCAAAATTCT,
TNF-α [34]: F: TAACGTTTCCCTCGTGAGTA; R: CCGCAGTCTAGAACGCAGAT,
TNF-α [33]: F: GGCAGGTCTACTTTGGAGTCATTGC; R: ACATTCGAGGCTCCAGTGAATTCGG,
IL-1β [34]: F: GGATGCTACGAAGTTAGCAC; R: GTTAGTGATGGTGGTGATAT,
IL-1β [33]: F: CACTACAGGCTCCGAGATGA; R: TTTGTCGTTGCTTGGTTCTC.
TNF-α Protein Level Measurement
Biomarkers with significant differences between the model control and the negative control in the neuroinflammatory mRNA level measurement were selected for further observation of their protein expression levels. To evaluate the effect of the three tested drugs on the neuroinflammatory protein level, total proteins from the collected cortex and hippocampus were extracted using RIPA lysis buffer (CWBIO, China) containing a protease inhibitor cocktail (Beyotime, China) according to the manufacturer’s instructions. The total protein was separated using a 4%–12% FuturePAGE gel (ACE Biotec, China) and then transferred to polyvinylidene fluoride (PVDF) membranes (Merck Millipore, Germany). The PVDF membranes were blocked with 5% non-fat milk in phosphate-buffered saline with Tween 20 (PBST) for 1 h. To monitor the neuroinflammatory protein levels, the membranes were incubated with primary antibodies against TNF-α (1:1000, ABclonal, China) and beta-actin (1:10,000, ABclonal) overnight at 4 °C. The membranes were then incubated with HRP-conjugated goat anti-rabbit secondary antibody (1:5000, ABclonal) at room temperature for 1 h. The protein bands on the PVDF membranes were visualised using a ChemiDoc MP imaging system (Bio-Rad, Singapore) and quantified using ImageJ software (http://rsb.info.nih.gov/ij/).
Molecular Docking for GABAA Receptors and the Human Serum Albumin (HSA)
Molecular docking procedures were performed to analyse the binding mode within GABA_A_ receptors and the human serum albumin (HSA) binding site-1 and site-2 for midazolam, diazepam, and lorazepam by using AutoDock Vina. Crystal structures of GABA_A_ (PDB ID: 6 × 3X) and HSA (PDB ID: 2BXD) were downloaded from the Protein Data Bank. The molecular structures of the studied ligands were prepared and optimised using the Molecular Operating Environment (MOE). The centre of the docking search box was determined using the co-crystal ligands for GABAA (D: DZP:404) and site-1 of HSA (A: RWF:2), and the search spaces were 151515 Å for diazepam and lorazepam; 161616 Å for midazolam. Since no co-crystal ligands were available in the site-2 of HSA, the residues (A: ASN:391) around HSA site-2 were utilised to define search spaces for the studied ligands (151515 Å for lorazepam, 161616 Å for midazolam and diazepam). The conformation with the lowest binding energy was selected for analysis, while visualisation was performed using PyMOL (Version 2.1.0).
Data Analyses
Experimental results were presented as mean ± one Standard Error of the Mean. To assess statistical significance, two-way ANOVA followed by Tukey’s post hoc test was used for multiple group comparisons in seizure scores and brain-to-plasma concentration ratios, while one-way ANOVA followed by Tukey’s post hoc test was used for multiple group comparisons in mRNA level and protein level. Protein binding was obtained by calculations described by Wang and Williams [35]. Data analysis was conducted using GraphPad Prism (version 9.5.1; GraphPad Software, La Jolla, CA). A p-value below 0.05 was considered statistically significant.
Results
Comparison of Seizure Scores in Different Drug Treatment Groups
The seizure scores over 10 min and 100 min after intranasal administration of vehicle (control), midazolam (1 mg/kg), diazepam (1 mg/kg), and lorazepam (1 mg/kg) to pilocarpine-induced C57 mice was shown in Fig. 2. At 10 min as shown in Fig. 2A, the mean seizure score of control, midazolam, and lorazepam were 2.7, 2.4 and 2.8, respectively, which were all significantly higher than that of diazepam (mean score of 1.4) (p < 0.01). Starting from 30 min to 100 min, the seizure scores of midazolam, diazepam, and lorazepam were all significantly lower than that of the control, but no difference was observed between the three drug-treated groups. Furthermore, mortality analysis revealed that 100% (n = 6) of the mice administered the vehicle died within 60 min due to lack of seizure control. In contrast, 100% (n = 6/treatment group) of the mice administered the benzodiazepine drugs (midazolam, diazepam, and lorazepam) survived until the end of the experiment.
Fig. 2. Comparison of seizure scores after intranasal administrations of vehicle (control), midazolam (1 mg/kg), diazepam (1 mg/kg) and lorazepam (1 mg/kg) for A early anti-epileptic effect and B late anti-epileptic effectin pilocarpine-induced C57 mice
Comparison of Drug Plasma and Brain Concentration from Different Treatment Groups
The mean concentrations of midazolam, diazepam, and lorazepam in plasma and brain homogenate after intranasal administration to pilocarpine-induced C57 mice were measured at 10 and 100 min. At 10 min, the mean plasma concentration was 80.2 ± 30.4 ng/ml, 49.1 ± 18.1 ng/ml, and 117.9 ± 42.9 ng/ml, respectively. Meanwhile, the mean brain concentration was 127.9 ± 54.4 ng/g, 156.3 ± 39.5 ng/g and 193.9 ± 39.7 ng/g for midazolam, diazepam and lorazepam, respectively. At 100 min, the mean plasma concentration was 119.0 ± 25.6 ng/ml, 21.5 ± 4.6 ng/ml, and 128.2 ± 79.0 ng/ml, and the mean brain concentration was 145.7 ± 27.3 ng/g, 58.4 ± 17.9 ng/g, and 177.9 ± 77.7 ng/g for midazolam, diazepam and lorazepam, respectively. The brain-to-plasma concentration ratios of the three treatment groups were calculated and compared as shown in Fig. 3. Among the three tested drugs, diazepam had the highest brain-to-plasma concentration ratio at both 10 min and 100 min, at 3.4 ± 1.0 and 2.7 ± 0.6 respectively, with no significant differences in brain-to-plasma concentration ratio between midazolam 1.6 ± 0.3 at 10 min, and 1.2 ± 0.2 at 100 min and lorazepam 1.8 ± 0.5 at 10 min and 1.6 ± 0.4 at 100 min treatment groups.
Fig. 3. Comparison of mean brain-to-plasma concentration ratios of midazolam, diazepam and lorazepam after intranasal administration of lorazepam (1 mg/kg), diazepam (1 mg/kg) and midazolam (1 mg/kg) to pilocarpine-induced C57 mice at 10 and 100 min
Comparison of Tested Drugs Stability in Plasma/Brain Tissue and Their Bindings Towards Plasma/Brain Tissue
After incubation, the percentages of drugs remaining in plasma were 102.76%, 100.08%, and 98.64% and those in the brain were 99.48%, 99.20%, and 101.78% for midazolam, diazepam, and lorazepam respectively, suggesting good stability of these tested drugs in plasma and brain tissue at physiological temperature. The bindings of the three drugs to plasma proteins and brain tissue were all above 80% with plasma protein bindings of 98.26%, 90.22%, and 92.96%, and the brain tissue bindings of 95.14%, 88.41%, and 84.62% for midazolam, diazepam, and lorazepam, respectively. It is noted that mouse plasma protein binding was most significant for midazolam, followed by lorazepam, then diazepam. At the same time, brain tissue binding was most significant for midazolam, followed by diazepam, then lorazepam.
Comparison of Neuroinflammatory mRNA and Protein Levels for the Three Intranasally Delivered Drug Treatments
Among the tested inflammation marker mRNA levels in the cortex shown in Figure S1, TNF-α was the only one that significantly differs at 10 min (approximated six times higher) in model control when comparing with those in negative control (p < 0.001), indicating that TNF-α could serve as an inflammation marker in the acute seizure model. Further comparisons demonstrated that the significantly increased TNF-α mRNA (Fig. 4A) and protein levels (Fig. 4B) in both the cortex (Fig. 4A (i) & Fig. 4B (i)) and hippocampus (Fig. 4A (ii) & Fig. 4B (ii)) at 10 min in model control than that from negative control (p < 0.0001), and these increases were significantly reduced by intranasal treatment with midazolam group, diazepam group, or lorazepam group. Among these drug treatments, intranasally administered diazepam exhibited the greatest efficacy in reducing TNF-α mRNA (p < 0.05 vs. Lorazepam, p < 0.05 vs. Midazolam) and protein levels (p < 0.05 vs. Lorazepam, p < 0.01 vs. Midazolam) in both cortex and hippocampus.
Fig. 4. Effect of intranasally administered midazolam, diazepam and lorazepam on A TNF-α mRNA level and B TNF-α protein level in the i cortex and ii hippocampus of pilocarpine-induced acute seizures mice at 10 min
Comparison of GABAA Receptor and HSA Binding of the Three Selected Drugs via Computer Docking
For the GABA_A_ receptor, computer docking of midazolam, diazepam, and lorazepam revealed that the docking pose of diazepam matched reasonably with the experimental pose (RMSD = 1.913 Å), with a binding affinity of − 9.4 kcal/mol. RMSD refers to the “root-mean-square deviation of the heavy atoms which is a parameter to evaluate the closeness of two poses in space. Usually, a RMSD value of less than 2 Å indicates a fairly good match and the acceptability of the docker procedure [36]. The binding affinities for midazolam and lorazepam were − 9.8 kcal/mol and − 10.0 kcal/mol, respectively. Overall docking results suggested a similar affinity of the three drugs to the benzodiazepine binding site, with lorazepam showing the highest affinity, followed by midazolam and diazepam, as shown in Fig. 5A.
For HSA, all three benzodiazepines showed a relatively low affinity toward site-2 of HSA (affinity < − 5.0 kcal/mol) and a moderate affinity (− 8.9 to − 8.0 kcal/mol) to site-1 of HSA. Further analysis of protein-ligand interactions between the three benzodiazepines and site-1 of HSA shown in Fig. 5B indicated the existence of arene-cation interactions with no hydrogen-bond interactions. Among the three tested drugs, diazepam has a relatively lower affinity (− 8.0 kcal/mol) to site-1 of HSA than those of midazolam (− 8.6 kcal/mol) and lorazepam (− 8.9 kcal/mol), which could serve as a partial explanation of its lower plasma protein binding in mice.
Fig. 5. Further mechanistic illustrations via A overlapping of the docking poses with lowest binding energy for the studied three benzodiazepines in GABA_A_ receptor and B arene-cation interactions between residues (Lys199, Arg257) of HSA and the studied three benzodiazepines (i: diazepam; ii: midazolam; iii: lorazepam)
Discussions
Our current study for the first time directly compares the efficacy of intranasal administration of three benzodiazepines used for first-line seizure control, namely midazolam, diazepam, and lorazepam, using a pilocarpine-induced acute status epilepticus (SE) model. A pilocarpine seizure model of mice was used for this pilot investigation, as it is commonly used to research temporal lobe epilepsy or status epilepticus [37]. Our results demonstrated the effectiveness of intranasal delivery of the three seizure drugs to control acute SE. When comparing midazolam, diazepam, and lorazepam, it was noted that diazepam performed better in terms of onset time and seizure control compared to midazolam and lorazepam. To further explain such difference in drug onset time and seizure control, we have considered the following factors: (1) difference in brain permeability; (2) difference in GABA_A_ receptor affinity; (3) stability in plasma and brain tissue; (4) binding toward plasma protein and brain tissue, and (5) effects on neuroinflammation.
Brain uptake is favorable in drugs with higher lipophilicity, as it allows better penetration through the blood-brain barrier [38]. The favorable performance of diazepam can be attributed to its lipophilic nature, so that it can be more rapidly distributed into the brain [39]. Previous studies have shown a difference in the time to reach maximum brain concentration between midazolam, diazepam, and lorazepam, attributed to their respective lipophilicities. Of the three drugs, midazolam is most lipophilic, followed by diazepam and then lorazepam [23, 40, 41]. Lipophilicity also affects the intranasal absorption of drugs, since it needs to diffuse through mucous membranes that cover the nasal passage [42]. After that, the drugs are transported to the brain, either through the direct pathway: the olfactory nerves or the trigeminal nerves, or the indirect pathway, the nasal vasculature [9]. Compared to midazolam and diazepam, lorazepam is less effectively absorbed intranasally due to its comparatively lower lipophilicity [42, 43].
It should be noted that whole brain homogenates were utilized for pharmacokinetics analysis to assess total drug delivery and ensure detection sensitivity, whereas region-specific dissections (hippocampus and cortex) were employed for molecular analysis to capture localized inflammatory changes without dilution from unaffected brain regions. Inflammatory responses within brain tissue have been documented in both human epilepsy patients [44] and animal seizure models [45]. Recent studies indicate that these inflammatory processes might be pivotal in both the onset and progression of epilepsy [46]. Among the tested drug treatments, intranasally administered diazepam was most effective in controlling acute seizures and reducing TNF-α mRNA and protein levels in both the hippocampus and cortex - two brain regions critically involved in seizure initiation [47] and strongly affected by pilocarpine [48]. The significant reduction of TNF-α in the hippocampus by intranasal benzodiazepines, especially diazepam, suggests their potential to mitigate early epileptogenic processes. Moreover, the suppression of this response by diazepam may explain its superior efficacy against motor seizures.
Upon reaching the brain, midazolam, diazepam, and lorazepam allosterically bind to specific subunits on GABA_A_ receptors to modulate the binding of GABA, thus producing anti-epileptic effects [10]. Hence, receptor affinity was also investigated using computer-docked models of GABA_A_ receptors. The receptor affinity of the three drugs was similar, but lorazepam had a slightly higher affinity to GABA_A_ receptors, followed by midazolam, then diazepam. This aligns with previous studies measuring benzodiazepine receptor binding affinities using receptor Ki. In those experiments, lorazepam and midazolam had higher binding affinities, whereas diazepam had a comparatively lower affinity [23, 49, 50]. As evidenced by a previous study [23] and our current findings suggest that receptor affinity plays a less significant role in determining onset compared to lipophilicity.
Considering the similar GABA_A_ receptor affinity and lack of degradation in plasma and brain for all tested drugs, the greater onset of intranasally delivered diazepam could be attributed to its highest brain-to-plasma ratio among all treatment groups. The plasma protein and brain tissue bindings of midazolam, diazepam, and lorazepam were also investigated in the present study. Drug-to-plasma-protein and drug-to-brain binding can deduce the unbound fractions of free drug available for distribution, and therefore anti-epileptic effects [51]. Our findings showed that diazepam had the lowest plasma protein binding, followed by lorazepam, and then midazolam. This may be partially explained by our computer docking of the three drugs to HSA, which revealed that diazepam had a relatively lower binding affinity to site-1 of HSA compared to the other two drugs. Similar observations of plasma protein binding were made in previous studies, in which diazepam had a higher unbound drug fraction in plasma than midazolam [41]. In addition to protein binding, much lower brain tissue binding of diazepam than midazolam was noticed, suggesting more free diazepam available in the brain, which could contribute to its more efficient onset and better seizure control.
Our study design adopted prophylactic administration of benzodiazepines in order to mimic the early intervention of seizures. The high survival rates in our study demonstrated that intranasal administration deliver benzodiazepine drugs to the brain fast enough to stop the seizure process before it becomes severe. This is highly relevant for patients with seizure clusters, who often take ‘rescue medication’ when they feel a warning sign to prevent status epilepticus [5]. We recognize that clinical administration of benzodiazepines usually follows the onset of seizures in status epilepticus instead of prophylactic administration, however, precise intranasal dosing is difficult to achieve in a mouse undergoing convulsions. Administering the drug beforehand ensured that every animal received the exact correct dose, avoiding data variability caused by drug loss during seizures.
Our preclinical findings are consistent with clinical epilepsy management. Notably, a recent real-world study by Li and Benbadis reported that while both intranasal agents are effective, intranasal diazepam achieved a higher rate of seizure cessation after the first dose (83%) compared to intranasal midazolam (52%) in adult patients [5]. This clinical observation replicates our current data, where intranasal diazepam demonstrated superior acute seizure suppression compared to midazolam. Our pharmacokinetic analysis suggests that this clinical advantage may be driven by the favorable lipophilicity and brain-to-plasma distribution profile of diazepam, enabling it to rapidly reach therapeutic thresholds in the early stage of seizure clusters.
Our study is not without limitations. The pilocarpine seizure model was selected in this pilot study as it is commonly used to mimic acute SE, a common form of seizure in humans [37]; however, this model administers antiepileptic drugs prophylactically before seizure induction, instead of after seizure onset like in a real seizure episode. Furthermore, a mouse model may not be representative of a human model, as the typical GABAA receptor modulation in humans may differ slightly from that of mice. As this was a pilot study, an observational model for seizure patterns was more suitable. However, for further investigations, mice could be objectively monitored for an episode of status epilepticus, such as using an EEG, followed by administration of drugs, and then continued monitoring of drug seizure control. Another limitation is that due to our sampling strategy (assessing behavior at 10-minute intervals) employed to monitor the long-term duration of drug action, continuous parameters such as precise seizure latency and frequency were not recorded. Future studies utilizing continuous video-EEG monitoring could be adopted to capture these details.
In summary, the benzodiazepines midazolam, diazepam, and lorazepam are effective and useful drugs in the control of acute seizures. The intranasal administration of these anti-epileptic drugs can provide a rapid intervention in non-hospital settings. In particular, diazepam was superior in brain uptake due to its higher brain-to-plasma concentration ratio compared to midazolam and lorazepam.
Supplementary Information
Below is the link to the electronic supplementary material.Supplementary material 1 (DOCX 170.1 kb)
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1I. UCB (2023) NAYZILAM® (midazolam) nasal spray Highlights of Prescribing Information, U.S. Food and Drug Administration