Minimally Invasive Nasal Depot (MIND) Technique for Direct BDNF AntagoNAT Delivery to the Brain

Abstract

The limitations of central nervous system (CNS) drug delivery conferred by the blood-brain barrier (BBB) have been a significant obstacle in the development of large molecule therapeutics for CNS disease. Though significantly safer than direct CNS administration via intrathecal (IT) or intracerebroventricular (ICV) injection, the topical intranasal delivery of CNS therapeutics has failed to become clinically useful due to a variety of practical and physiologic drawbacks leading to high dose variability and poor bioavailability. This study describes the minimally invasive nasal depot (MIND) technique, a novel method of direct trans-nasal CNS drug delivery which overcomes the dosing variability and efficiency challenges of traditional topical trans-nasal, trans-olfactory strategies by delivering the entire therapeutic dose directly to the olfactory submucosal space. We found that the implantation of a depot containing an AntagoNAT (AT) capable of de-repressing brain derived neurotrophic factor (BDNF) expression enabled CNS distribution of ATs with significant and sustained upregulation of BDNF with efficiencies approaching 40% of ICV delivery. As the MIND technique is derived from common outpatient rhinological procedures routinely performed in Ear, Nose and Throat (ENT) clinics, our findings support the significant translational potential of this novel minimally invasive strategy as a reliable therapeutic delivery approach for the treatment of CNS diseases.

Introduction

Neurodegenerative diseases represent a significant risk to global health. Alzheimer’s disease and Parkinson’s disease are the most prevalent, affecting millions of Americans [1–3]. These diseases are primarily characterized by progressive neuronal degeneration within the central nervous system (CNS), specifically in brain regions such as the hippocampus and substantia nigra, eventually leading to disruption of motor function and cognitive decline [4–6]. The majority of treatment strategies are expensive [3,7], and yet only directed towards symptom management. Disease modifying therapeutics remain elusive, largely as a result of restrictions on CNS delivery imposed by the blood-brain barrier (BBB).

Neurotrophic factors are capable of stimulating both pro-survival and pro-functional neuronal activity, which makes them attractive candidates for treating several neurodegenerative disorders [8,9]. Brain derived neurotrophic factor (BDNF), specifically has shown to support neuronal survival and differentiation. Furthermore, BDNF reductions within the cortex, hippocampus, and substantia nigra have been reported to result in neuronal loss contributing to both Alzheimer’s disease and Parkinson’s disease [10–13]. These findings have led to numerous strategies for augmenting BDNF levels in an effort to mitigate disease progression in these patients [12,13]. Nevertheless, CNS delivery of neurotrophins remains challenging due to their unique physico-chemical properties, such as high molecular weight and polarity, which effectively render them BBB impermeant [8,14]. Additional complexities have arisen with respect to the engineering of recombinant neurotrophic factors, such as toxicity concerns, dosing challenges, potential immunogenicity, and off target effects owing to improper post-translational alterations and mis-localization resulting from manufacturing processes [15–17].

The human BDNF gene has a highly intricate structure due to the presence of 11 distinct exons, regulated by 9 different functional promoters which causes cells to generate many BDNF transcripts. The initial synthesis as pre-pro BDNF precursor, presence of alternative splice sites and transcription stop sites, post-translational modification and existence of a conserved noncoding natural antisense transcript (NAT), termed BDNF-AS, all further contribute to its complexity [13,18–20]. Such distinct features governing the translational and post-translational aspects of BDNF activity pose great challenges to the engineering of a recombinant protein capable of mimicking its endogenous counterpart.

Of high relevance to neurotrophin therapies, oligonucleotide-based therapeutic approaches for treating neurodegenerative disorders are gaining interest, particularly due to their high target specificity, reduced systemic exposure, limited toxicity, and relatively prolonged half-life within the CNS as compared to the majority of the small molecule drugs [21]. Moreover, the free uptake of naked oligonucleotides by cells alleviates the requirement of any kind of potentially toxic and/or immunogenic carriers, making them good therapeutic options [21].

Oligonucleotide-based therapies have significantly benefited from the discovery of NATs abundantly present in mammalian genomes [21]. NATs are a subclass of endogenous long non-coding RNAs transcribed specifically from the opposite strand of the coding gene loci. NATs possess the ability to regulate their corresponding sense gene expression [22]. For example BDNF-AS was reported to repress BDNF sense RNA transcription by changing the BDNF locus chromatin composition, thereby decreasing the levels of endogenous BDNF protein [23]. Inhibition of BDNF-AS was found to significantly upregulate BDNF mRNA expression resulting in increased BDNF protein levels [23]. This “de-repression” can be achieved using AntagoNATs (ATs) which are synthetic, short, chemically modified, single-stranded oligonucleotide-based compounds, complementary to a particular NAT. In particular, BDNF AntagoNAT (BDNF AT) can inhibit BDNF-AS activity. Upregulation of endogenous BDNF by AntagoNATs was found to induce neuronal differentiation in vitro in neuronal progenitor cells and enhance the neuronal proliferation in vivo [23]. In addition to their single protein locus specificity, ATs are also active only in cells with active expression in their target locus [24,25]. Moreover, the low prevalence of nucleases in cerebrospinal fluid (CSF), high potency, and ease of neuronal uptake highlight the utility of AT-based approaches for therapy of neurodegenerative diseases [21,26,27].

Despite these evident advantages, the clinical adoption of AT-based strategies remains hindered by their inability to cross the BBB [21]. AT administration to the CNS currently requires invasive techniques such as intracerebroventricular (ICV) and intrathecal (IT) delivery [21,27,28]. These routes have the potential to induce trauma, brain edema, catheter failure, infections, and even death [29,30].

The morbidity and risks of invasive CNS drug delivery techniques have driven considerable investigation into the direct trans-nasal pathway to the CNS. This method utilizes a highly permeable nasal mucosal membrane with a direct anatomic conduit into the brain via the olfactory neurons which innervate the olfactory epithelium (OE) which constitutes the olfactory mucosa of nasal cavity [31,32]. Proponents of this method suggest that trans-nasal delivery may bypass the BBB, thereby avoiding the need for invasive delivery methods [33] while minimizing systemic side effects [34]. Though promising, the topical intranasal delivery of CNS therapeutics has not been widely clinically adopted due to a variety of practical and physiological limitations. Specifically, the human OE comprises only 3–5% of human nasal surface area and is positioned within a narrow cleft facilitating only smaller volumes of drug administration and limited topical drug distribution [35–39]. Any drug which achieves contact with the OE then has an effective residence time of only 15–20 minutes before it is cleared by mucociliary action [40]. During this period, the nasal mucus can further exert a degradatory effect through an array of secreted proteases and nucleases [41–43]. Finally, the residual drug must diffuse through the semipermeable epithelial cell layer in order to gain access to the olfactory nerve sheathes [44]. Additionally, the majority of studies employ a separate nasal delivery component to permit the delivery of drugs via nose to the brain [45,46]. These factors conspire to create intrinsically variable dose uniformity for topical intranasally delivered agents which has, in turn, restricted clinically successful delivery programs for CNS therapeutics.

In this study, our team demonstrates for the first time, an alternative approach of trans-nasal drug administration for direct CNS delivery. Herein, the drug suspended in a gel carrier is directly injected into the submucosal compartment of the OE, in order to overcome the intrinsic limitations associated with topical intranasal administration to the surface of OE. In doing so, the entire dose is implanted within the tissue directly surrounding the olfactory neurons thereby ameliorating the concerns regarding distribution, retention, trans-epithelial diffusion, and dose uniformity. The idea of “minimally invasive nasal depot (MIND)” technique is based on routine intranasal procedures performed by otorhinolaryngology specialists in Ear, Nose and Throat (ENT) clinics using commonly available endoscopic instrumentation in awake patients with minimal discomfort [42]. Using this technique, the therapeutic depot can be directly injected to the submucosal space of OE facilitating its direct CNS delivery. Considering the inherent ease and safety of the procedure, the MIND technique holds significant translational potential and can be performed in the clinic in both the adult and paediatric populations.

In light of the unique therapeutic potential of BDNF ATs in neurodegenerative disease, our team hypothesized that the MIND technique could be used to enable direct AT delivery to the brain with efficiencies approaching those of direct invasive ICV injections. The purpose of this study was to therefore: 1) develop a novel rodent model of the MIND technique and 2) investigate the pharmacokinetics (PK) and pharmacodynamics (PD) of BDNF AT delivered through this method in naïve animals.

Materials and methods

Study Design

We utilized a sample size of 4 Sprague Dawley rats for each formulation type (BDNF AT-Liposome-in Gel and BDNF AT-in Gel) under 7 different time points (2, 6, 12, 24, 48, 72 and 96 hours) for investigating the safety and efficacy of MIND technique. The intracerebroventricular BDNF AT solution dosed arm comprised of n= 4 rats for 3 time points such as 2, 12 and 24 hours. n = 4 untreated naïve rats were used as controls. The in vivo CNS distribution of BDNF AT when delivered by MIND technique or ICV was studied by quantifying the AT levels in different brain end target regions such as olfactory bulb, striatum, hippocampus, substantia nigra and cerebellum by BDNF AT hybridisation assay. BDNF protein levels in all of the above samples were quantified by ELISA and normalised to the protein content measured by BCA assay. Noncompartmental analysis (NCA) of the concentration-time curves of both AT levels and BDNF protein levels for MIND and ICV groups were performed using the SimBiology application within the MATLAB software (version 2018b) and the results were compared for studying the delivery efficiency of MIND relative to ICV.

Design and Validation of BDNF ATs

The first step in the design of AntagoNATs used for this study was identification of the potential regulatory NATs in the BDNF locus on rat chromosome 3 (accession numbers: CN544668, AI030286 and BF391266 in the UCSC genome browser https://genome.ucsc.edu/cgi-bin/hgGateway). Single stranded, phosphorothioated ATs to these transcripts were designed with the help of Oligo Design and Analysis Tools, IDT Inc., San Diego, CA, http://www.idtdna.com/calc/analyzer). A selection of those AT sequences with no other substantial homologies in the rat genome was done using NCBI BLAST and such ATs were thereafter synthesized by IDT. In the in vitro screen in RT4D6P2T cells (ATCC® CRL2768^™) the AT #CUR-2219 termed ‘BDNF AT’ with the sequence C*A*T*A*G*G*A*G*A*C*C*C*T*C*C*G*C*A*A*C, against the transcript CN544668 was found to show the highest rate of BDNF mRNA upregulation as determined by real time PCR as described in Hsiao et al. [47]. The molecular weight of BDNF AT is 6.361 kDa. Melting point is 57.4 °C and GC content is 60 %.

Synthesis of BDNF AT cationic liposomes

Lipids for synthesizing liposomal formulations such as cationic DOTAP (1, 2 dioleoyl-3-trimethylammonium propane (chloride salt) and anionic DPPC (1,2 dipalmitoyl-snglycero-3-phosphocholine) were purchased from Avanti Polar Lipids, Alabaster, AL. Cholesterol salt serving as stabilising neutral lipid was procured from Sigma Aldrich, St. Louis, MO. Liposomes encapsulating BDNF AT with the lipids, DOTAP: cholesterol: DPPC mixed in the molar ratio of 5:3:5 were synthesized by the thin film hydration protocol mentioned previously. Briefly, 1 mL from stock solutions of each of these lipids dissolved in the organic solvent chloroform (Fisher Scientific, Fair Lawn, NJ) were taken and mixed in a round bottom glass flask which was subsequently subjected to rotary evaporation using a rotary evaporator (RV Control 10, IKA, Staufen, Germany) rotating at 100 rpm. This was connected to a water bath set at room temperature, to facilitate the removal of chloroform and subsequently form a thin lipid film. The flask was thereafter removed from the apparatus and then vacuum dried overnight for the complete removal of residual solvent. Hydration of the film was done with the addition of 1 mL of saline containing BDNF AT (2 mg/mL AT concentration) and vortexed till a white suspension was formed. This was followed by five repetitive freeze thaw cycles by alternately placing the flask with suspension in ice bath and warm water bath, for two minutes each separated by vortexing for 30 seconds. The suspension was then probe sonicated on ice for about 5 minutes, following which it was centrifuged at 20000 g for 15 minutes at 4 °C using ultra centrifugal filters of 100000 molecular weight cut off (Amicon® Ultra-15, Millipore Sigma, Burlington, MA). The concentrated supernatants (with a volume of 300–350 μL) were collected and used for all experiments henceforth.

Determination of BDNF AT loading efficiency of liposomes

The encapsulation efficiency of BDNF AT in the concentrated supernatant of liposomal formulation was determined by the indirect method of quantification. This gives a measure of the amount of AT which was not encapsulated within the liposomes. Nanodrop spectrophotometer (NanoDrop 2000/2000c, Thermo Scientific, Waltham, MA) was utilized to measure the AT content in the centrifuged supernatant. The amount obtained was subtracted from the initial amount loaded into the liposome (during synthesis) for determination of BDNF AT encapsulation efficiency.

Characterisation of morphology of liposomes

The morphology of liposomes was analysed using transmission electronic microscopy (TEM).

1.5% uranyl acetate was used to negatively stain the liposomal particles. To 10 μl of sample, a 300-mesh carbon coated copper grid was applied. Thereafter, the grid was briefly touched by filter paper to wick off the excess sample. Following three rinses with deionised water, the grid was touched 3–5 times with 1.5% uranyl acetate stain. Sample grids were finally visualized with TEM instrument (JEOL JEM-1010, Tokyo, Japan) at an operating voltage of 80 kV.

Preparation of BDNF AT-Liposome-in Gel and AT-in Gel formulations

Thermosensitive liposome-in gel (LiG) system was formed by suspending the liposomes in a gel formed by Pluronic F-127 (BASF Corp., Florham Park, NJ). Owing to the gelling property of Pluronic at room temperature and flow retention as liquid at 4 °C, the 30 w/v% solution was formed by dissolving Pluronic F-127 in saline upon continuous stirring at 4 °C. BDNF AT encapsulated cationic liposomes (300–350 μL concentrates with 200 μg AT) were added to 300 μL of homogenous Pluronic solution (30 w/v%) at 4 °C and stirred thoroughly to form the BDNF AT-LiG formulation (0.25–0.35 mg/mL AT, final). BDNF AT (2 mg/mL) dissolved in 300 μL Pluronic gel solution directly (AT-G) served as the formulation for comparisons with respect to LiG group.

Characterisation of size and charge of liposomes and liposome-in F-127 formulations

Liposome characteristics such as hydrodynamic diameter, size distribution and charge were measured with the help of dynamic light scattering (Nano-ZS90, Malvern Instruments, Inc., Westborough, MA). Prior to the measurement, formulations were diluted with aqueous phase appropriately. Values obtained as three independent measurements from each sample were averaged and expressed as mean ± SD. Similarly, BDNF AT liposomes dispersed in Pluronic F-127 solutions were also tested for the above attributes under different dilutions with the help of dynamic light scattering.

In vivo studies

The design of animal experiments was based on guidelines of Institutional Animal Care and Use Committee (IACUC) of Northeastern University. Male Sprague-Dawley rats (250–300 g) used for this study were purchased from Charles River Laboratories (Wilmington, MA) and were housed in pairs under controlled laboratory conditions in an automatic 12 h/12 h light/dark cycle. Rats were provided with sterilised food and drinking water ad libitum.

Animal surgery by Minimally Invasive Nasal Depot (MIND) approach

Rats were anesthetized with 2.5 % isoflurane and placed on stereotactic frame with body temperature being maintained at 37 °C using a heating pad. The nasal dorsum surgical site was shaved and aseptically prepared with povidone iodine and alcohol. A midline sagittal incision was then made with a scalpel blade from the naso-frontal suture line to the nares. Bilateral skin flaps were elevated and laterally undermined to expose the underlying nasal bone while creating a subcutaneous pocket of 250–300 μL. A high speed drill (Dremel, Mt. Prospect, IL) was used to thin and remove the nasal bones while preserving the underlying basolateral olfactory mucoperiosteum. The skin incisions were then closed with a running locking stitch using 5–0 nylon sutures (Med-Vet International, Mettawa, IL).

Dosing of BDNF AT formulations in Pluronic F-127 gel

After the closure of skin incisions, an 18-gauge needle was used to make a separate puncture providing direct access to the submucosal potential space. Appropriate volumes of both formulations (BDNF AT-LiG and AT-G maintained on ice) corresponding to the weight of rat and AT dose of 0.15 mg/kg were injected using this syringe to form a submucosal depot of the respective formulation in gel. The syringe was left in the same position for few minutes-post dosing to obviate the sudden efflux or leakage of formulation and to facilitate the gelling of entire volumes of formulations at 37 °C.

Intracerebroventricular dosing of BDNF ATs in solution form

Intracerebroventricular (ICV) dosing of BDNF AT solution into the lateral ventricles was performed using the stereotaxic coordinates, anterio-posterior (AP) = 1 mm, mediolateral (ML) = 2 mm, dorsoventricular (DV) = 4 mm (Paxinos and Watson rat brain atlas, 1997) using a programmable syringe pump (Remote Infuse/Withdraw Pump 11 Elite Nanomite Programmable Syringe Pump, Harvard Apparatus, Holliston, MA) connected to a microlitre syringe (Hamilton®, Reno, NV). BDNF AT in water (0.15 mg/kg) was infused at a flow rate of 1.0 μL/min. The syringe was left in position for an additional 5 minutes prior to withdrawal to ensure the complete influx of free AT into the injection site.

Pharmacokinetic evaluation of BDNF AT formulations administered through different routes of delivery

Animals undergoing MIND surgery with BDNF AT-LiG and AT-G formulations were randomly assigned to different groups corresponding to time points such as 2, 6, 12, 24, 48, 72 and 96 hours (n = 4 animals/group/time point). Similarly, animals dosed with ICV BDNF AT solution were also grouped under time points 2, 12 and 24 hours (n = 4 animals/group/time point). Blood was collected by cardiac puncture prior to sacrifice according to experimental group and plasma was collected as supernatant after its centrifugation at 2000 g for 10 minutes at 4 °C. Following sacrifice, brain tissues were collected. End target regions of interest were isolated according to rat brain atlas coordinates and 3 mm tissue biopsy punches were used to collect uniform samples from the striatum, hippocampus, substantia nigra, and cerebellum. Punches were homogenized using ice cold tissue lysis buffer (composition: 10 mM Tris-HCl (pH 7), 100 mM NaCl, 0.4 mM EDTA, 2 g BSA, 2 % Triton X-100 and 1.54 mM sodium azide in 100 mL water, with EDTA-free Protease Inhibitor tablet (Sigma-Aldrich, St. Louis, MO)) added and centrifuged at 20,000 g for 20 minutes to extract the total protein as supernatants. The protein content in each of these extracted samples was quantified using Pierce BCA assay kit (Thermo Fisher Scientific, Waltham, MA). Plasma as well as brain tissue extracts were subjected to a hybridization assay conducted as described previously [47] to quantify BDNF AT levels.

Capture and detection probes used for the hybridization assay were designed with capture probe complementary to the 3’ whereas detection probe complementary to the 5’ end of the BDNF AT as shown below:

BDNF AT sequence - 5′-C*A*T*A*G*G*A*G*A*C*C*C*T*C*C*G*C*A*A*C-3′

Capture probe sequence - (5AmMC12//iSp18/iSp18//G*+T*+T*+G*+C*+G*+G*+A*+G)

Detection probe sequence -(+G*+G*+T*+C*+T*+C*+C*+T*+A*+T*+G/iSp18//iSp18//iBiodT//3BioTEG) in which * corresponds to phosphorothioate bond, + corresponds to LNA modifications, 5AMmc12 : 5’-amino modifier C12m, iSp18 : internal 18-mer spacer, iBiodT : internal biotin-dT and 3BioTEG : 3’ biotin-TEG). Probes were synthesized by Qiagen Inc (Germantown, MD).

Briefly, 40 μL of 5000 pmole/mL capture probe was added to 19.96 mL of capture probe buffer, and used to coat a 96 well white NuncTM plate (Thermo scientific, Waltham, MA). The plate was then blocked with BSA and incubated with a solution of detection probe (200 μL of 5000 pmole/mL of detection probe in 19.8 mL of 4X SSC / 0.5 % sarkosyl buffer (Fisher Scientific, Fair Lawn, NJ) thermally annealed with extracted sample. After wash, a streptavidin-HRP conjugate (Jackson ImmunoResearch, West Grove, PA), at a dilution of 1:50,000 was added and the plate was incubated for 30 minutes at 37 °C. Following the washing step, 150 μL of the Elisa Femto Solution Mix (Thermo Scientific, Waltham, MA) was added into the wells and luminescence was measured immediately by plate reader.

BDNF AT concentrations were calculated based on the standard curve obtained with values from AT standards made in homogenisation buffer and expressed as pg AT/μg protein, normalising with respect to protein content in the tissue sample. Average values are calculated from the samples of all animals of each group and represented as mean ± SEM. AT concentrations at different regions of brain were plotted with respect to different time points for different formulations and delivery routes.

Evaluation of pharmacodynamic response to BDNF AT formulations administered through different routes of delivery

Extracted protein samples of brain sub-regions were subjected to quantification of BDNF protein de-repression. This was performed using a commercially available BDNF sandwich ELISA kits (ChemiKine Brain Derived Neurotrophic Factor, Sandwich ELISA, Millipore Sigma, Burlington, MA). BDNF protein concentrations were normalized to total protein and expressed as pg BDNF/μg protein. Average values were calculated from the samples of all animals in each group and represented as mean ± SEM. Comparisons were made across different delivery routes as well.

Pharmacokinetic analysis

BDNF AT amounts per gram tissue were converted to concentration of tissue equivalents (i.e., pg/mL) after normalization to the tissue mass using reported values for rat brain density [48]. For all data sets, inconsistent terminal data points, likely a reflection of intrinsic tissue variability and of assay variability, resulted in poor data fits. Therefore, terminal slope analysis was not included, and compartmental analysis was not performed. Instead, noncompartmental analysis (NCA) of the concentration-time curve was performed using the SimBiology application within the MATLAB software (version 2018b) [49]. The maximum concentration (C_max) and time of maximum concentration (t_max) were determined using SimBiology, and verified graphically. The area under the tissue concentration-time curve (AUC) from time zero to the last measured time (AUC_0-last) was determined using the linear trapezoidal method. The area under the first moment concentration-time curve (AUMC) was calculated from time zero to the last measured time (AUMC_0-last) using the linear trapezoidal method. Due to the lack of reliable terminal data points, AUC or AUMC analysis was not extended to time infinity. The mean residence time (MRT) was calculated as the ratio of AUMC and AUC (i.e., MRT = AUMC/AUC).

Pharmacodynamic analysis

The tissue levels of BDNF in treated animals were baseline corrected using BDNF levels from naïve animals. NCA of the concentration-time curve of BDNF protein was performed using SimBiology/MATLAB to determine the maximal concentration (C_max) and time of maximal concentration (t_max). The area under the effect curve (AUEC) analysis was restricted from time zero to the last measured time (AUEC_0-last).

Statistical analysis

All experiments were performed with duplicates or triplicates of samples and the experimental data is represented as mean ± SEM. Statistical significance was assessed by using 2-way ANOVA with post-hoc Tukey and Sidak tests for multiple comparisons. The analyses were carried out with GraphPad Prism (version 6.01) and statistical significance was set at p < 0.05 for all analyses.

Results

Development of MIND model in Sprague Dawley rats

The first goal of our study was to create and validate a reproducible surgical model to recapitulate the human MIND technique in healthy Sprague Dawley rats. In humans, the OE submucosa may be directly accessed using an endoscopically guided trans-nasal injection without requiring sedation or general anaesthetic (Fig. 1A and B). However as the rat snout is too small to access this space trans-nasally, we developed the described open surgical approach with skin incision and removal of the nasal bones (NB) (Fig. 1C–F). The nasal bones of rats were first accessed through the creation of bilateral skin flaps which simultaneously created a subcutaneous pocket facilitating the depot implantation (Fig. 1E). A reproducible method to atraumatically expose the olfactory submucosa without penetrating into the nasal cavity or rupturing of blood capillaries was provided by repetitive thinning and subsequent drilling of a small portion of the nasal bone (of elliptical dimensions approximately 0.5 cm length × 0.3 cm diameter). Histologic assessment of the surgical site confirmed the preservation of epithelial barrier layer within the OE even after nasal bone removal (Fig. 1F). Ultimately however, the open surgical technique creates access to precisely the same OE submucosal space as the injection as depicted in Fig. 1C–F. The overall procedure is therefore minimally invasive with regard to the other invasive procedures performed for CNS delivery [29,30,50].

Fig.1. Rat Model of Minimally Invasive Nasal Depot (MIND) technique.

A. Illustration of placement of the minimally invasive depot injection in the human nose within the submucosal space of the olfactory epithelium (OE, dotted line, adapted from Servier Medical Art). B. Performance of an intranasal procedure in an awake patient during an outpatient visit to the Otolaryngology (ENT) clinic guided by a high-definition nasal endoscope with the image projected onto a real time high definition monitor indicated by black arrow. OE labelled on screen as seen during an endoscopic exam. C. Cross sectional illustration of rat snout demonstrating location of the depot in contact with the basolateral aspect of the olfactory epithelium (OE). The position of this depot is directly anatomically analogous to the submucosal space occupied by the depot in the human nose depicted in Fig. 1A. D. Rat surgical model demonstrating reflection of skin flaps with exposure of nasal bones (NB). E. Removal of nasal bones reveals the basolateral aspect of the olfactory epithelium bilaterally (white arrows). F. Histologic section (H&E) of rat surgery depicted in Fig. 1E with nasal bone (NB) removed and depot in place (inset is higher magnification view comparing depot site before and after NB removal over olfactory epithelium).

The rat model of MIND was successfully reproduced in all 56 rats used for the entire study and all animals were found to tolerate the depot procedure with no significant attrition throughout the study period. There were no signs of infection or inflammation at the surgery site and animals did not display any signs of pain or distress, further confirming the safety of the procedure.

Synthesis and characterisation of BDNF AT-Liposome-in Gel (AT-LiG) and AT-in Gel (AT-G) formulations

Having established the MIND surgical model in rats, our next goal was to administer BDNF ATs through this approach. We formulated liposomes-in gel as delivery vehicles for this payload and characterized them. BDNF ATs were embedded within liposomes at an entrapment efficiency of ~100 %. The structure and inner architecture of liposomes were characterised by imaging using transmission electron microscopy (TEM) (Fig. 2A). For the next step of synthesizing gel formulations, we made a concentrated and homogenous 30 % (w/v) Pluronic F-127 solution which is liquid at 4 °C and solidifies as a gel at body temperature (Fig. 2B). Dispersion of BDNF AT liposomes and BDNF AT into this solution at 4 °C resulted in the formation of AT Liposomes-in-gel (AT-LiG) and AT-in gel (AT-G) formulations respectively. Cationic liposomes entrapping BDNF ATs possessed an average hydrodynamic diameter of 157.91 ± 7.78 nm with an average polydispersity index of 0.246 ± 0.05 and a zeta potential of 38.67 ± 0.96 mV (Fig. 2C). The characteristics of these intact BDNF AT liposomes were retained even after their dispersion in Pluronic F-127 solution (Fig. S1 and Fig. 2C).

Fig. 2. Liposome-in Gel formulation and characterization.

A. TEM images showing the morphology of BDNF AT liposomes at 100 nm magnification. B. Thermosensitivity of Pluronic F-127 gel. C. Characteristics of BDNF AT Liposomes and Liposomes-in F-127 formulation (n = 3, values represented as mean ± SD).

CNS distribution after BDNF AT delivery by ICV infusion

Prior to studying the efficacy of BDNF AT delivery by MIND surgery, we investigated the brain distribution of BDNF AT delivered via ICV infusion. Since the ICV route delivers drug directly into the CSF, it is an appropriate positive control for evaluating alternate modes of delivery.

BDNF AT in saline was delivered directly to the cerebral ventricles of rats by ICV infusion (Fig. 3A). Relevant end target brain tissues, such as olfactory bulb (OB), striatum (STR), hippocampus (HC), substantia nigra (SN) and cerebellum (CB), were collected at 2, 12 and 24 hours, and BDNF AT levels were quantified by using AT hybridization assay (Fig. 3B). Mean AT amounts 2 hours post-infusion ranged from 4 to 6 pg/μg and declined to 3 to 4.5 pg/μg by 24 hours.

Fig. 3. ICV infusion of BDNF AT into rat brain and CNS distribution of AT levels.

A. ICV infusion of BDNF AT solution to lateral ventricles of rat brain by Hamilton syringe connected to a pump. B. BDNF AT levels in rats ICV-dosed with AT, detected in different regions of rat brains such as olfactory bulb (OB), striatum (STR), hippocampus (HC), substantia nigra (SN) and cerebellum (CB) (hybridization assay data represented as mean ± SEM, n = 4 animals/time/group, *p < 0.05, 2-way ANOVA with Tukey post-hoc test for multiple comparisons).

Pharmacokinetic analysis of BDNF AT delivery by ICV infusion

We further characterised the time course of BDNF AT brain tissue concentrations following ICV AT delivery using noncompartmental analysis (NCA) (Table 1). Maximum tissue concentration was attained within 2 hours for all tissues, and it ranged from 82.5–107 pg*mL⁻¹. The extent of tissue distribution, as related by area under the curve (AUC) values, ranged from 1729 pg*mL⁻¹*hr to 2273 pg*mL⁻¹*hr. Consistent mean residence time (MRT) values within the range of 11.2–11.6 hours were obtained for all brain tissues.

Table 1.

Pharmacokinetic parameters of BDNF AT following administration of free AT via intracerebroventricular infusion.

Parameter	Unit	OB	STR	HC	SN	CB
tmax	Hr	2	2	2	2	2
Cmax	pg*mL−1	107	88.1	93.8	82.5	85.6
AUC†	pg*mL−1*hr	2273	1743	1969	1729	1803
AUMC†	pg*mL−I*hr2	26481	19462	22597	20013	20507
MRT†	Hr	11.6	11.2	11.5	11.6	11.4

^†Estimated from time zero to last time point; AUC – Area under the curve, AUMC – Area under the first moment curve, MRT- Mean Residence Time

CNS distribution of BDNF AT after delivery by MIND technique

Subsequently, we explored the CNS distribution of BDNF AT delivered as AT-LiG and AT-G formulations via the MIND approach. Both formulations were found to gel in situ once in direct contact with the rat olfactory submucosa. Following administration, the BDNF AT levels in relevant end target brain regions were quantified at 2, 6, 12, 24, 48, 72 and 96 hours (Fig. 4 & Fig. S2).

Fig. 4. BDNF AT levels in brain regions of rats dosed with BDNF AT formulations using MIND.

BDNF AT levels in rats dosed with BDNF AT in the liposome-in-gel (AT-LiG) and gel (AT-G) by MIND surgery, detected in different regions of rat brains such as olfactory bulb (OB), hippocampus (HC) and substantia nigra (SN) at different time points (hybridization assay data represented as mean ± SEM, n = 4 animals/time/group, *p < 0.05, **p < 0.001, ##p < 0.001, ***p < 0.0001, 2-way ANOVA with Sidak test for multiple comparisons; * for comparisons between time points for same formulation, # for comparisons of AT-LiG with naïve levels)

Both formulations resulted in a similar distribution of AT in all brain regions except for the OB. Administration of AT-G resulted in higher mean AT concentrations in the OB at 6 h when compared to the AT-LiG formulation (p = 0.0083) (Fig. 4). Moreover, these OB AT levels at 6 h were significantly higher than that of mean AT levels at later time points for AT-G such as 48, 72 and 96 hours in the OB. In contrast, LiG increased the extent of AT exposure compared to AT-G in other brain regions, particularly in HC as noted by increased mean AT levels at 2, 6 and 12 h, and in SN at 12, 24 and 48 h respectively. The SN AT levels of Li-G at 12 hours was higher than that for 48, 72 and 96 hours and a similar trend was noted in CB (Fig. S1). Importantly, BDNF AT was not detected in plasma samples retrieved from all of these animals.

Pharmacokinetic analysis of BDNF AT delivery by MIND technique

We further characterized the time course of BDNF AT tissue concentrations attained through MIND delivery using NCA to determine differences in PK between the two formulations (Table 2). The AT-G formulation demonstrated distinct kinetics in the OB compared to deeper brain tissue. The extent of distribution, as determined by AUC values, varied from 210–556 pg*mL⁻¹*hr in deeper brain tissue, which is substantially lower than the distribution in the OB (Table 2). Similar trends were observed for C_max with a range of 11.9–34.0 pg*mL⁻¹ for deep brain tissue and 93.9 pg*mL⁻¹ for the OB. The AT-LiG formulation demonstrated a more uniform pharmacokinetic profile between the OB and deeper brain tissues. For all tissues, after AT-LiG administration t_max was 12 hours, C_max ranged from 21.7 – 46.7 pg*mL⁻¹, and AUC ranged from 457 – 2975 pg*mL⁻¹*hr. MRT was generally increased in AT-LiG compared to AT-G.

Table 2.

Pharmacokinetic parameters of BDNF AT following administration of AT-G and AT-LiG via MIND Surgery

AT-G MIND administration
Parameter	Unit	OB	STR	HC	SN	CB
tmax	hr	6	2	2	2	2
Cmax	pg*mL−1	93.9	34.0	11.9	19.6	28.7
AUC†	pg*mL−1*hr	2923	556	210	228	514
AUMC†	pg*mL−I*hr2	82014	13592	2002	1914	5144
MRT†	hr	28.1	24.4	9.51	8.41	10.0
AT-LiG MIND administration
tmax	hr	12	12	12	12	12
Cmax	pg*mL−1	46.7	35.8	38.9	21.7	30.4
AUC†	pg*mL−1*hr	2975	860	713	721	457
AUMC†	pg*mL−I*hr2	128404	24406	9771	18764	6433
MRT†	hr	43.2	28.4	13.7	26.0	14.1

^†Estimated from time zero to last time point

In vivo BDNF protein upregulation in rat brain after ICV infusion of BDNF AT and analysis of response kinetics

Having measured the PK of BDNF AT following ICV administration, we next quantified BDNF protein levels in each brain sub-region using naïve rats as a negative control (Fig. 5).

Fig. 5. BDNF protein levels in brain regions of rats after ICV delivery of BDNF AT.

Comparisons of BDNF protein levels detected in different brain regions of naïve (untreated) rats and those dosed ICV with free BDNF AT in water at different time points such as 2, 12 and 24 hours (ELISA data represented as mean ± SEM, n = 4 animals/time/group, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, 2-way ANOVA with Tukey post-hoc test for multiple comparisons).

BDNF protein levels were found to be significantly upregulated in all brain regions of ICV infused rats relative to the corresponding naïve protein levels particularly at 2 and 12 hours. The highest mean BDNF protein concentrations were noted at 12 hours for ICV infused animals in all brain regions. Specifically, the mean BDNF protein levels in HC and SN at 12 hours for ICV group were significantly higher than those attained in OB and STR at the same time point.

NCA was further used to characterise the time course of BDNF protein brain tissue concentrations following ICV delivery (Table 3). A relatively homogenous BDNF response to AT treatment was noted for all brain regions based on t_max, C_max and AUC.

Table 3.

Pharmacodynamic parameters of BDNF response following administration of free BDNF AT via intracerebroventricular infusion

Parameter	Unit	OB	STR	HC	SN	CB
tmax	hr	12	12	12	12	12
Cmax	pg*μg protein−1	1.81	2.16	3.21	3.08	2.13
AUEC†	pg*μL−1*hr	28.7	31.3	46.2	44.8	36.5

^†Estimated from time zero to last time point; AUEC – Area under the effect curve.

In vivo BDNF protein upregulation in the brain after BDNF AT delivery by MIND technique

We next quantified BDNF protein levels in each brain subregion of rats subjected to MIND technique using naïve rats as a negative control (Fig. 6 & Fig. S3).

Fig. 6. BDNF protein levels in rats dosed with BDNF AT via MIND approach.

Comparisons of BDNF protein levels of naïve (untreated) rats (average value of 4 rats represented as dotted line) and those dosed with BDNF AT in the liposome-in-gel (AT-LiG) and BDNF AT in gel (AT-G) by MIND surgery in different brain regions such as olfactory bulb (OB), hippocampus (HC) and substantia nigra (SN) at different time points (ELISA data represented as mean ± SEM, n = 4 animals/time/group, *p < 0.05, #p < 0.05, **p < 0.01, ##p < 0.01, φφp < 0.01, ***p < 0.001, ###p < 0.001, 2-way ANOVA with Tukey-post hoc test for multiple comparisons; * for comparisons between time points for same formulation, # and φ for comparisons of AT-LiG and AT-G respectively within the same time point with naïve levels)

BDNF protein levels were substantially upregulated in all brain regions for both formulations in comparison to the naïve BDNF levels. Generally, BDNF levels increased over time before peaking and declining towards baseline values. However, brain sub-regions exhibited tissue-specific kinetics, as suggested by the differences in t_max values. There was no statistically significant difference between BDNF levels for both AT-G and LiG formulations at all time points, despite the distinct differences noted in the mean AT concentrations between the two groups.

We also compared the mean BDNF protein levels in brain tissues of rats dosed with ICV AT to those of MIND formulations (Fig. S4). There was a general upregulation of BDNF after AT administration, regardless of route or formulation, as early as 2 h post-administration in different brain regions, especially SN and CB. Additionally, CB BDNF levels at 2 hours after ICV as well as MIND AT dosing were significantly higher than the levels at other brain sub-regions. At 12 hours, the ICV administration demonstrated higher mean BDNF protein concentrations compared to both MIND formulations. However, the mean BDNF protein levels of all groups were found to be similar at 24 hours in brain regions such as OB, STR and CB. Nevertheless, the mean BDNF protein levels at 24 hours, particularly after MIND AT-G administration were significantly elevated in HC and SN compared to naïve BDNF values.

Analysis of BDNF protein response kinetics after BDNF AT delivery by MIND technique

The kinetics of the pharmacodynamic response were also evaluated via NCA (Table 4). Unlike the pharmacokinetic parameters of AT, most BDNF-associated parameters were similar between both formulations. Specifically, C_max values for AT-G and AT-LiG (Table 4) ranged from 0.51–1.2 pg*μg⁻¹ and 0.76–1.0 pg*μg⁻¹, respectively. AUEC values for AT-G and AT-LiG, which represent the extent of effect over this time-frame, ranged from 25.2–53.9 pg*μg⁻¹*hr and 24.5–59.5 pg*μg⁻¹*hr, respectively. HC and SN were found to have the highest AUEC values of all tissues for both formulations, indicating an intrinsic tissue difference in the degree of response to BDNF AT treatment.

Table 4.

Pharmacodynamic parameters of BDNF response following administration of AT-G and AT-LiG via MIND technique

AT-G MIND administration
Parameter	Unit	OB	STR	HC	SN	CB
tmax	hr	6	24	24	24	2
Cmax	pg*μg protein−1	0.524	0.509	1.05	0.941	1.23
AUEC†	pg*μg protein−1*hr	28.4	25.2	53.9	52.0	28.7
AT-LiG MIND administration
tmax	hr	6	12	12	48	2
Cmax	pg*μg protein−1	0.995	0.763	0.931	1.03	0.796
AUEC†	pg*μg protein−1*hr	32.3	24.5	38.3	59.5	28.3

^†Estimated from time zero to last time point

Comparison of kinetics of BDNF AT delivery and BDNF protein response for the MIND approach relative to ICV infusion

We next compared the kinetic responses of MIND technique relative to ICV infusion in order to inform on relative delivery efficiency. A subset of the MIND formulation data, including time points matched to the ICV infusion data, was re-analysed by NCA. Subsequently, the C_max and AUC or AUEC ratios for MIND approach were calculated relative to ICV, for both AT and BDNF protein levels (Table 5). This comparative analysis was restricted to delivery of free AT (AT-G), since liposomal AT was not administered via ICV infusion.

Table 5.

Pharmacokinetic and pharmacodynamic parameter ratios for BDNF AT delivery and BDNF response following AT-G administration via MIND technique compared to free AT delivered via ICV infusion.

Pharmacokinetic parameter ratios
Ratio (MIND AT-G/ICV)	Unit	OB	STR	HC	SN	CB
Cmax	%	68.1	38.6	12.7	23.7	33.5
AUC†	%	61.8	23.9	10.2	12.2	29.8
Pharmacodynamic parameter ratios
Cmax	%	22.7	15.9	58.1	43.6	40.1
AUEC†	%	15.6	20.3	54.3	43.3	36.2

^†Estimated from time zero to last time point

As shown in Table 5, the MIND/ICV pharmacokinetic ratios for C_max varied substantially among tissues. The C_max ratio ranged from a lowest value of 12.7 % in the HC and a highest value of 68.1 % in the OB. Similarly, the AUC ratio ranged from 10.2–61.8 %. The C_max and AUEC ratios for the BDNF protein kinetics also varied substantially among different tissues. The C_max ratios ranged from 15.9 % in the STR to 58.1 % in the HC. The AUC ratios ranged from 15.6–54.3 %. Overall, these results suggest that MIND provides > 10 % efficiency of both AT delivery and BDNF expression compared to ICV administration.

Discussion

The purpose of this study was to explore the MIND approach as a modified intranasal strategy for CNS drug delivery by exploiting the unique connection between the CNS and olfactory epithelium within the nasal cavity. This technique overcomes the practical issues and physiological challenges associated with traditional topical methods of intranasal dosing, which often leads to poor drug uptake, significant dose non-uniformity and limited CNS distribution. Through the direct application of therapeutics to the submucosal space, the MIND technique ensures the accurate and precise delivery of the entire intended dose directly to the olfactory epithelial submucosa and avoids concerns regarding mucociliary clearance and restricted trans-epithelial diffusion.

Our rat model of MIND successfully recapitulated the anatomy of proposed clinical drug delivery [51,52], while avoiding the pitfalls traditionally associated with rodent models of trans-nasal drug delivery. These include a reliance on the relative abundance of olfactory epithelium in the rodent relative to the human and the lack of controlling for secondary CNS uptake from the peripheral vasculature [51,53]. The human olfactory mucosa comprises only 3–5% area of the entire nasal mucosa in humans whereas the rat olfactory mucosa occupies up to 50% of nasal area. Therefore the existing preclinical studies in rodents tend to vastly overestimate the relative amount of olfactory mucosa surface area which is available for drug uptake clinically within the human nose [35–39]. Conversely, the described model of the MIND technique exposes only a small surface area of the mucosa through the surgical approach thereby mitigating the traditional drawback of overestimating the efficacy of trans-olfactory delivery in rat relative to humans. Additionally, we avoided the use any additional nasal delivery devices to facilitate nose to brain dosing unlike previous studies [45,46].

Considering the extremely small size of the rat snout, we developed the described open surgical approach to access the OE submucosal space. This was achieved by the transcutaneous removal of the nasal bones while preserving the intact olfactory epithelium. We observed that the surgery was well tolerated, reproducible, and also provided direct delivery of BDNF ATs throughout the brain. As the MIND technique can be performed in humans using a simple trans-nasal injection without requiring any incisions or tissue removal, our preclinical findings in Sprague Dawley rats validate the potential clinical utility of the MIND approach as a safer alternative to the standard invasive BBB-penetrating CNS delivery procedures with their attendant morbidities and safety concerns [29,30,50].

In order to enhance the translatability of the MIND approach, we chose to use Pluronic F-127 gel as the basis of the AT depot. The thermo-gelling properties of Pluronic F-127 hydrogels enable precise administration via injection and enhance the residence time of the payload within the desired submucosal space. In addition to evaluating the MIND technique for AT delivery, we also explored the ability of a liposomal formulation to enhance delivery of BDNF AT. Cationic liposomes have been traditionally used to protect oligonucleotide cargo from circulating nucleases [54,55], and have been shown to enhance CNS uptake from periphery. The liposomal formulation of BDNF AT was able to provide complete entrapment of AT within a homogenous phospholipid vesicle population. This formulation was readily incorporated into the Pluronic F-127 hydrogel and the liposome characteristics were found to be retained.

In order to understand the utility of the MIND approach relative to standard CNS delivery procedures, ICV infusion directly into the lateral ventricles was used as a positive control. Noncompartmental analysis of the tissue concentration-time curve emphasized the homogenous distribution of BDNF AT within the brain after ICV administration. All of the calculated pharmacokinetic parameters were relatively similar across the various sub-regions of brain, which suggests that the kinetics of AT distribution from convective CSF flow and uptake into tissues were relatively homogenous. Analysis of the kinetics of BDNF protein response shows a similar homogeneity amongst brain tissues. The time of maximal BDNF protein concentration was 12 hours for all regions, which appears to be reasonable considering that AT acts on BDNF levels via a complex response mechanism [23]. All tissues also showed a relatively similar extent of response based on AUEC values.

Although ICV administration successfully delivered BDNF AT and upregulated BDNF protein expression in all brain regions of interest, it required a highly invasive procedure to physically breach the BBB. Moreover, this approach was severely limited in the ability to instill drug solution over a short period of time without significantly elevating intracranial pressure [56]. In fact, this volume limitation precluded our ability to test ICV administration of the liposomal formulation of BDNF AT at meaningful doses.

Evaluation of PK properties by NCA indicates that the brain distribution following the MIND approach differs from that of ICV infusion. For both depot formulations such as AT-G and AT-LiG, there is a greater degree of variability between tissues, and most notably, distinctly higher tissue concentration in the olfactory bulb. This is an unsurprising observation, as the OB is the first intracranial structure along the trans-olfactory delivery route [57], which suggests that diffusion contributed substantially to AT distribution into the OB, in addition to the expected convective distribution attributed to CSF flow [58]. Additionally, the lack of measurable BDNF AT concentrations in plasma samples from animals in the MIND groups strongly suggests that CNS uptake occurs predominantly via direct uptake through the olfactory epithelium and is not reliant on secondary peripheral distribution. Collectively, our PK studies validate that the MIND technique can successfully deliver the otherwise BBB-impermeant BDNF AT to all desired end target regions of brain.

Additionally, the pharmacokinetic analysis showed a clear differentiation between the two formulations. Compared to free AT-Gel, encapsulation in liposomes resulted in a greater and more uniform extent of BDNF AT distribution. This is suggested by the apparently narrower range of peak AT concentrations across all tissues, and the slightly elevated C_max and AUC values for deep brain tissues of particular interest (STR, SN, HC). Compared to free AT, the liposomal formulation appeared to delay the time to peak concentration and appeared to increase MRT in brain tissues. It is important to note, however, large variability in MRT in different tissues, indicating tissue-specific and formulation-dependent MRT of AT. Collectively, these data suggest the ability of liposomes to alter both the absorption and elimination kinetics of AT, and likely favor a slower release compared to free AT.

In order to inform on the biological effects of AT administration, the kinetics of the BDNF protein response were evaluated. The time course of BDNF protein levels indicates that both AT-Gel and AT-LiG formulations effected an increase in BDNF levels across the duration of the study compared to naïve animals. Notably, this kinetic analysis shows that with regard to the response to BDNF AT, as highlighted by AUEC values, there are negligible differences between the two formulations which is also supported by the BDNF protein concentration values. Not only are the magnitudes of the AUEC values very similar for each tissue, but the differences between tissues are comparable for both formulations. Given the observed differences in the extent of AT distribution for the two formulations, this analysis suggests a decoupling of the concentration-response relationship. This observation is consistent with the proposed mechanism for ATs, whereby the binding of BDNF AT to NAT relieves the NAT-mediated repression of BDNF mRNA transcription [23]. Based on this mechanism, it stands to reason that intrinsic levels of NAT, repressive complexes, and BDNF mRNA will all exert more control over the AT response than AT concentration itself. Indeed, it is known that brain tissues express varying amounts of BDNF mRNA available for translation [28], and can have limited and variable copy numbers of BDNF NAT [23], all of which may contribute to the variable BDNF response in various tissues [23].

In order to evaluate how effectively the MIND approach delivered AT relative to ICV, the ratios of the kinetic parameters were compared. This comparison suggests that the MIND technique could provide effective CNS delivery of BDNF AT, as evidenced by an AUC ratio of > 10 % for the MIND technique relative to ICV. This is highly clinically relevant since widespread implementation of direct CNS delivery is limited by safety risks [29,30] and peripheral delivery of BBB-impermeant drugs (e.g., adacanumab) is limited by very low efficiencies (< 1 %) [59,60]. Furthermore, our comparison of BDNF protein response kinetics showed the highest AUEC ratios for the HC and SN (> 40 % of ICV), which are the brain tissues most relevant in neurodegenerative diseases such as Alzheimer’s and Parkinson’s diseases.

There are several limitations of our study which bear discussion. With respect to the animal model, the limited time points in the ICV arm narrowed the analysis of terminal slope values for PK studies. Therefore, the noncompartmental analysis used to compare the kinetic parameters between the ICV and MIND delivery was confined to the overlapping time points. Similarly, the NCA was performed with averaged data as opposed to readouts from individual animals. Nevertheless, our overall results demonstrated the utility of the MIND approach for achieving efficient CNS therapeutic delivery with minimal morbidity.

These results are highly translatable as the MIND technique is directly derived from endoscopic endonasal procedures already commonly performed in the outpatient ENT clinic. Therefore, the feasibility of performing a submucosal depot injection on an awake patient with minimal discomfort using endoscopic guidance has already been validated. Furthermore, our data demonstrated that desirable levels of BDNF upregulation could be obtained with very low doses of AT due to the saturable mechanism of upregulation ascribed to BDNF ATs. This implies that the volume of depot required for any given administration would not exceed the capacity of the submucosal space in patients. Finally, ATs and similar oligonucleotide based compounds such as nusinersen have already been approved by the FDA for intrathecal use indicating their safety for use within the CNS.

Conclusions

The present work demonstrates a novel, reliable and highly translatable strategy for trans-nasal delivery of blood brain barrier (BBB) impermeant drugs directly to the central nervous system (CNS). This technique called as MIND, can be easily performed in the outpatient clinic in both adult as well as pediatric populations by direct depot injections of therapeutics to submucosal space, and therefore has enormous translational potential. We have also established the minimally invasive nature of this procedure which possess an improved safety profile relative to the highly invasive clinically adopted routes of CNS drug delivery such as intrathecal or intracerebroventricular (ICV) modalities. Our in vivo results demonstrate the effective utility of MIND technique using BDNF AntagoNATs (ATs), a class of BBB-impermeant oligonucleotide therapeutics. An efficient CNS uptake and brain distribution of BDNF ATs attained through MIND delivery of ATs suspended in thermogelling polymers led to significant BDNF upregulation in various brain sub-regions, particularly in hippocampus and substantia nigra, which are relevant areas of interest for those neurodegenerative diseases such as Alzheimer’s and Parkinson’s. An efficiency approaching 40 % of direct ICV delivery emphasizes the promising potential of MIND technique as a highly effective therapeutic delivery strategy for CNS diseases.

Highlights.

• Minimally Invasive Nasal Depot (MIND) is a novel nose-to-brain delivery technique

• MIND obviates challenges of olfactory distribution, residence, and dose-uniformity

• MIND gives efficient CNS delivery for BBB-impermeant drugs such as BDNF AntagoNATs

• AntagoNAT delivery via MIND affords BDNF expression in different brain sub-regions

• BDNF upregulation has high potential for treatment of neurodegenerative diseases

Abbreviation list

AT

AntagoNAT

AT-G

AT-in-gel

AT-LiG

AT Liposomes-in-gel

AUC

Area under the curve

AUEC

Area under the effect curve

AUMC

Area under the first moment curve

BBB

Blood-brain barrier

BDNF

Brain derived neurotrophic factor

CSF

Cerebrospinal fluid

CNS

Central nervous system

ENT

Ear, Nose and Throat

ICV

Intracerebroventricular

IT

Intrathecal

MRT

Mean residence time

NAT

Natural antisense transcript

NCA

Noncompartmental analysis

OE

Olfactory epithelium

PD

Pharmacodynamics

PK

Pharmacokinetics

TEM

Transmission electron microscopy