Parkinson's disease (PD) is a neurological disorder that develops after the irreversible deterioration of the dopaminergic neurons in the substantia nigra due to the accumulation of α-synuclein inside cytoplasmic bodies called lewy bodies. The formation of α-synuclein involves toxic intermediates that can hinder the mitochondrial and lysosomal functions and cause membrane disruption [1]. It is more common in males, smokers, and old people (above 50 years) with higher incidence for patients with previous family history [2]. PD is clinically manifested by some motor and non-motor symptoms. The main clinical motor symptoms are postural instability, bradykinesia, and extremities' tremors especially during rest, whereas; the non-motor symptoms may include sleep disorders, depression, and cognitive impairment [3]. Diagnosis of PD depends mainly on the appearance of these clinical symptoms with no clear radiological or laboratory indicators [4]. Although the early diagnosis of the disease improves its prognosis and clinical symptoms, but it is also challenging as the clinical symptoms usually appear after the loss of 60 % of the dopaminergic neurons. Transcranial sonography has been recently applied as a one of the useful tools in the early diagnosis to detect the hyperechogenicity of the substania nigra observed in PD patients [5]. On the other hand, recent studies have also shown some advances in the early diagnosis by measuring the metabolites of dopamine in the plasma and cerebrospinal fluid (CSF) [6]. It is worth noting that the loss of dopaminergic neurons is irreversible without a definite cure to decrease or stop the disease progression; thus, the treatment plan usually aims to relieve the symptoms and prevent further dopaminergic neurons loss [7]. Monoamine oxidase B (MAO-B) inhibitors such as Selegiline and Rasagiline are commonly used in combination with levodopa to decrease dopamine degradation, stop neuronal deterioration and relieve the symptoms [8].

Rasagiline mesylate (RM) is used as one of the first line treatments in the management of PD symptoms. It exerts its effects by inhibiting MAO-B enzyme irreversibly to restore sufficient levels of dopamine inside the CNS [8]. Besides its MAO-B inhibitory activity, RM, at its therapeutic dose, possesses a free radical scavenging activity, which contributes to its neuroprotective effect [9]. However, the hydrophilic nature of RM limits its ability to cross the blood brain barrier (BBB) and reach the CNS to exert its effect. In addition, it suffers from low oral bioavailability, extensive first pass metabolism and rapid systemic elimination [10]. All these factors have contributed to the limited therapeutic outcomes accompanying RM utilization and urged the need for a more efficient multiple dose regimen [11].

Drug administration via the intranasal (IN) route has been efficiently employed to deliver drugs for the systemic circulation and target the central nervous system (CNS) as well. The high permeability and vasculature of the nasal mucosa have contributed to the rapid and almost complete absorption of the intranasally administered drugs [12]. In addition, the ease of IN administration and their ability to overcome the first pass metabolism have added more interest in their application to improve drug delivery and enhance patient compliance [13]. The IN route has also been applied for the systemic delivery of proteins, peptides and hormones that cannot withstand the acidity and degradation through the gastrointestinal tract. So, it was employed in the administration of cyanocobalamin to treat vitamin B12 deficiency in children [14] and as alternative route for the injectable insulin [15]. On the other hand, the utilization of the IN route for brain targeting has received special interest lately because of the direct connection between the olfactory mucosa located at the roof of the nasal cavity to cribriform plate in the cranial cavity through olfactory bulb. This direct nose to brain drug delivery may result in faster onset of action with better efficacy which may permit dose reduction and subsequently lower peripheral side effects of the centrally acting drugs [12].

Nanotechnology has gained a significant interest in the biomedical field as a way to improve the targeting behavior of certain drugs. Nanocarriers are characterized with their small size which allows more efficient drug delivery with higher therapeutic efficacy and/or better diagnostic performance. Their nano-scaled size permits the effective permeation and delivery of most of the therapeutic agents in higher concentration to systemic circulation and different tissues. Similarly, they can easily permeate through the tiny junctions of the blood brain barrier (BBB) achieving higher and better localization in the brain [16]. Additionally, nanosystems can be designed to control the drug release rate; hence, they are widely used for sustaining the action of certain therapeutics [17].

The Blood brain barrier (BBB) is characterized by tight junction complexes that act to keep the adjacent endothelial cells together. These tight junctions impart an electrical resistance to the BBB, which limits the passage of pathogens, toxic substances, and most therapeutic compounds [18]. This barrier also tends to limit the therapeutic effect of many CNS-acting drugs. Most of the brain targeting systems depend on the passive influx from the circulating system to the extravascular spaces in brain which is greatly restricted by the surface area of the components and their net surface charge [19]. One of the techniques used to overcome the tightness of BBB is to formulate nanoparticles with small particle size (PS), high elasticity and deformability that are able to squeeze themselves, pass through BBB and target the various brain tissues. It was previously reported that the absence of the fenestrations in the BBB reduces the chance of nano-vesicles with size more than 200 nm to cross BBB [20]. The surface charge of nanoparticles is another factor that may affect their passage through biological membranes [16]. The brain endothelial lining cell membrane is negatively charged; thus, the passage of the charged therapeutic compounds will be affected by their interaction with the negatively charged endothelial membrane [21]. Positively charged nanoparticles may penetrate into the brain by adsorption transcytosis as they usually got attracted to the negatively charged endothelial membrane at the blood side which forms a vesicle enclosing the positively charged particle then pass through the brain parenchyma to be integrated with the endothelial membrane of brain side. On contrary, negatively charged particle will be repulsed from the endothelial cells resulting in longer systemic circulation time [22,23]. This claim was also supported by Fenart et al. [24] who found that the positively charged nanosystems have better brain targeting performance than the negatively charged and uncharged systems. Another study by Gu et al. [25] showed that the positively charged chitosan coated nanoparticles with size range 150–250 nm were effective in delivering small interfering RNA (siRNA) through BBB. Meanwhile, other studies have claimed that the anionic small colloidal dispersions show higher BBB penetration than neutral large ones [26]. Similar results were obtained by Lockman et al. [27] and Tavitian et al. [28] who concluded that the anionic nanoparticles have demonstrated an enhanced BBB permeability. As observed in the literature, the effect of the surface charge of the formulated nanoparticles on their penetrability through BBB is contradictory.

Transfersomal vesicles are elastic nanovesicles of an inner aqueous core surrounded by one or more lipid bilayer. They can be synthetized using natural phospholipids such as phosphatidylcholine together with a proper edge activator (EA). Their ultra-deformability, elasticity, and ability to squeeze themselves and pass through the tight junctions of the BBB have supported their application for the intranasal delivery of different therapeutic agents to the brain [29,30].

In the last decades, it has been more convenient in the biomedical field to rely on an objective un-biased statistical tool while evaluating and optimizing the different formulation variables to produce an optimized system with optimum performance and favorable properties. The full replicate factorial designs were mainly applied for this purpose until recently [31] where many researchers switched to the less time-consuming and comparably efficient partial replicate designs. And thus, Response Surface Methodology (RSM) was utilized. RSM is a designed regression analytical methodology that is characterized by its ability in the optimization and prediction of data for different combinations of levels other than those used in constructing the design [32]. It collects the mathematical and statistical aspects for building and evaluating different processes variables and analyze it to discover the optimum combination of the independent factors required to achieve the desired responses [33]. RSM shows a great ability in investigating the interactions and main effects between the independent variables and the response of interest in fewer runs than the conventional designs; thus, it saves time and reduces costs [34]. Different RSM designs have been employed in different experimental fields namely, Box-Behnken design, Central Composite design, and Optimal designs (I and D-optimals). Optimal designs are characterized by having a constrained design area that allows the inclusion of qualitative and quantitative variables, combined with great economic benefits owing to their power in analyzing the experimental variables with the least number of trials [35]. I-optimal design criterion is characterized by generating more runs in the center of the design and fewer runs in the extremes, whereas D-optimal design generates more points in the edges than the core. I-optimal design reduces the average prediction variance over the whole design space than D-optimal design, so the I-optimal design criterion focuses more on prediction while the D-optimal design is more precise in estimating the experiment parameters; and thus, more frequently used in the screening stage [36]. It was previously claimed that the application of I-optimal design, which is more oriented towards prediction, is more preferred for data optimization [37]. Surprisingly, the D-optimal design is more commonly used to evaluate and optimize the effect of the different formulation variables in many pharmaceutical studies [[38], [39], [40]], while the I-optimal design is not usually used.

Based on these pieces of evidence, the main aim of this study was to enhance the IN delivery of RM to the brain through its incorporation into elastic vesicles with high brain targeting power. Thus, RM-loaded transfersomal vesicles were prepared applying thin film hydration technique. In the course of our study, we managed to formulate negatively and positively charged vesicles using different EAs to study the effect of the particle size and vesicles’ surface charge on their brain targeting behavior. A secondary objective in our study was to assess and compare the performance of the two statistical optimal designs namely, I-optimal and D-optimal; in terms of their power in evaluating and optimizing the investigated independent formulation variables. The formulated transfersomal vesicles were characterized and the results were analyzed and optimized using the 2 statistical approaches to select an optimum negatively charged (NT) and positively charged transfersomes (PT). The optimized systems were then characterized in terms of in-vitro release profiles, deformability index (DI), transmission electron microscope (TEM) and pH. Finally, the in-vivo performance and brain targeting behavior of the optimized NT and PT were evaluated upon IN administration compared to an intravenous RM solution in rat model.