1. INTRODUCTION
Peptides, short amino acid monomer chains linked by peptide bonds, have gained interest as therapeutic agents over the past decade. Although peptides exert their pharmacological effects in various ways, the process typically begins with their binding to cell surface receptors or other proteins, thereby triggering an intracellular or extracellular mechanism. Most peptides have predictable metabolism and high selectivity, with reasonable safety and efficacy [1]. Thus, they have relatively short retention times in the body, high potency, and good patient tolerability. The major barriers in the widespread use of peptides as drug candidates is their low bioavailability arising from the chemical, physical and enzymatic instability of the peptide structure and low membrane permeability [2]. These problems lead to short half-lives, rapid elimination from the body, and usually a lack of orally bioavailability [3].
The most common route of administration to deliver peptide drugs to their desired sites of action is parenteral injection (e.g., neurotensin via intravenous injection). This route is considered the conventional method because of the predictable onset of action and satisfactory bioavailability. However, the parenteral route is profoundly invasive, painful, and inconvenient; moreover, peptides require repeated administration because of their extremely short half-lives, thus placing a major burden on patients [4, 5]. This problem is more severe for the patients with severe neurological disorders, in whom repeated doses must be administered over extended periods of time [4]. Oral delivery of peptide drugs to the brain requires the drugs to cross the intestinal epithelial membranes, enter the peripheral bloodstream, and then cross the blood-brain barrier (BBB) into the brain [5]. Formulation scientists have attempted to address the issues of stability and delivery of peptide drugs through non-invasive approaches, such as buccal, intranasal, sublingual, rectal, ocular, and oral delivery. Oral delivery is the most commonly used route because of its simplicity, convenience, cost effectiveness, and patient acceptability and consequently compliance [6]. However, the progress made in developing orally administered pharmaceutical peptide drugs to cross the BBB has been inadequate to date because of poor gastrointestinal tract (GIT) stability, as well as low membrane permeability across both intestinal epithelial membranes and the BBB [7]. These shortcomings can be addressed by using novel bi-ligand formulation technologies (e.g., conjugation of bi-ligands to liposomes, nanoparticles, and niosomes) [8]. Two ligands have been used to deliver peptide drugs to the brain: the non-specific cell-penetrating peptide poly-L-arginine (targeting peptide delivery across the GIT) and a transferrin or lactoferrin ligand specific to the transferrin receptors abundantly found along the BBB [9, 10]. These two ligands are used to improve the cellular uptake and transport of peptide drugs to the brain across the GIT and BBB, by protecting them inside nanoparticular vesicles. Encapsulation with a drug delivery system can also prevent degradation and increase peptide permeation and delivery to the brain [11]. Other studies have indicated that drug delivery systems can be functionalized by using various ligands to specifically and maximally target drugs to the BBB for efficient uptake and transport [12]. The aim of this review is to evaluate the various receptors, ligands, and drug delivery systems used to deliver peptide drugs to the brain via oral administration. The chemical, enzymatic, and physical barriers that must be overcome, and the various strategies used to penetrate these barriers, are discussed in detail.
1.1. The gastrointestinal tract barrier
The key role of the GIT is to digest and absorb nutrients while protecting the human body against harmful agents [13]. To achieve this, the GIT has several specialized routes of absorption and many protective mechanisms against potential pathogens, antigens, or toxins. The main pathways of drug absorption are paracellular, transcellular, and receptor-mediated transcytosis ( Figure 1 ) [14]. The paracellular pathway transports molecules across the epithelium through their passage through intercellular spaces [15]. It is classified as a passive transport mechanism that relies on a concentration gradient and is usually restricted to small (<100–200 Da) hydrophilic molecules [16]. The transcellular pathways that transport molecules through the cell across the apical and basolateral membranes can be either active or passive [17]. Receptor-mediated transcytosis is an active molecular transport process in which intestinal epithelial cells adsorb various proteins, metabolites, or hormones through receptor recognition [18]. Any ligand or chemical modification used to overcome the GIT can be classified into three categories: the paracellular pathway, lipophilic transcellular pathway, and receptor-mediated transcytosis pathway. Both the paracellular pathway and transcellular lipophilic pathway are dependent on the drug and/or ligand interaction with the cell wall, on the basis of lipophilicity. The ability of a drug to overcome the GIT barrier via one of the above pathways does not necessarily translate to the utilization of the same pathway to cross the BBB, owing to the large structural differences between these barriers, as highlighted later in this review. Transport through the pathway of receptor-mediated transcytosis usually results from a receptor’s strong affinity toward a given drug or ligand, thus enabling uptake ( Figure 1A ).
A) Schematic diagram of the various pathways of drug transport across the GIT. B) Schematic diagram showing the different cell types of single-layer intestinal epithelium.
The two main challenges in GIT absorption are physical and enzymatic (biochemical) barriers, and drug metabolism and drug elimination are additional contributing factors.
Cell membranes and tight junctions make up the cell lining; along with the mucus layer and efflux systems, they create the GIT physical barrier [19]. The presence of a stagnant aqueous layer consisting of water, mucus, and glycocalyx inside the GIT gives rise to the unstirred water layer. This layer prevents large peptides from reaching the epithelium but otherwise has limited importance in the absorption of molecules [20]. A combination of enterocytes, goblet cells, endocrine cells, and Paneth cells ( Figure 1B ) make up the single layer of the intestinal epithelium [21, 22]. These cells are connected by tight junctions and create a rigid physical barrier that is semi-permeable because of the cellular phospholipid bilayer [23]. Tight junctions, adherens junctions, and desmosomes are three types of intercellular junctional complexes [24]. Tight junctions are the only occluding junctions, and they contain fenestrae with widths between 3 and 10 Å [25]. These tight junctions are selectively permeable and are regulated by several compounds that alter their permeability [26]. On the apical surface, P-glycoprotein efflux pumps actively transport molecules back into the GIT lumen, thus decreasing the overall absorption of multiple drugs. Thus, efflux systems in combination with intracellular metabolism can markedly alter the oral bioavailability of a compound [27].
Enzymatic barriers in the GIT cause peptide instability because of the presence of high pH, enzymes, and/or microorganisms [28]. Proteolytic enzymes are abundant throughout the body and degrade unprotected peptides [29–31]. The pH-dependent hydrolysis of drugs occurs throughout the GIT, because the pH of the intestinal fluids varies substantially [21]. In addition, microorganisms in the colon can perform deglucuronidation, decarboxylation, reduction, hydrolysis, and dihydroxylation reactions, all of which can lead to peptide degradation [32–35]. Specifically, a range of aminopeptidases and carboxypeptidases facilitate the breakdown of the peptide at the terminal ends, and endopeptidases cleave peptides at specific sites. This review focuses on drug delivery systems specifically crafted to overcome the physical and enzymatic barriers of the GIT. After overcoming the GIT barriers, peptides must then be delivered across the BBB to target the brain.
1.2. The blood-brain barrier
Physical and enzymatic barriers present in the BBB prevent many compounds, including peptides, from crossing [36, 37]. The BBB is composed of brain capillary endothelial cells, pericytes, astrocytes, and neuronal cells. The key functions of the BBB include maintaining ion homeostasis, preserving neural connectivity, preventing neurotoxic molecules from crossing the BBB, and transporting essential nutrients and molecules to the brain [36, 37]. The asymmetric arrangement of the membrane-bound transport systems on the apical and basolateral surfaces of the endothelium is essential to barrier function. In particular, P-glycoprotein efflux pumps on apical surfaces protect the brain from harmful or unwanted compounds [36, 37].
Brain capillary endothelial cells are characterized by continuous tight junctions that connect cells [36, 37]. These tight junctions, as well as adherens junctions between adjacent endothelial cells on the apical side of the BBB, create a physical barrier that prevents paracellular transport to the brain [36, 37]. The tight junctions between brain capillary endothelial cells have extremely high transendothelial electrical resistance; consequently, these membranes are highly resistant to passive diffusion [2, 36]. Almost 100% of macromolecular drugs and 98% of small lipophilic drug candidates (500–1,000 Da) are unable to cross the BBB, but small lipophilic (<500 Da) nutrient molecules are an exception [38]. An additional BBB metabolic barrier arises from the presence of various intracellular and extracellular proteolytic enzymes [2, 37]. This combination of characteristics makes the BBB very difficult to cross. Therefore, researchers are using ligands to target specific transporters or carrier systems to facilitate the passage of drugs across the BBB ( Table 1 ) [2, 37].
Brief list of methods for peptide delivery across the BBB
| Peptide | Formulation | Route of administration | Therapeutic effect | Ref. |
|---|---|---|---|---|
| Leucine-enkephalin | N-palmitoylN-monomethylN,N-dimethylN,N,N-trimethyl-6-O-glycolchitosan ligand | Oral | Antinociception | [39, 40] |
| Neurotensin | Angiopep-2 ligand | Intravenous | Allodynia | [40, 41] |
| O-palmitoyl tyrosinate ester-dalarg | Nanofibers | Intravenous | Antinociception | [40, 42] |
| Dalargin | Polymeric nanoparticles | Intravenous | Antinociception | [40, 43] |
| Opioid peptide | VH0445 peptide conjugates | Intravenous | Antinociception | [40, 44] |
| 4-Demethyl penclomedine | Derivative of 4-demethyl penclomedine conjugate | Intravenous | Brain neoplasms | [45] |
| Aurimmune | Tumor necrosis factor conjugate | Intravenous | Solid tumors | [46] |
The main pathways of drug transport across the BBB are the transcellular pathway, receptor-mediated transcytosis, adsorptive transcytosis, and the transporter pathways ( Figure 2 ) [47]. Any ligand or chemical modification used to overcome the BBB can be categorized into paracellular, transcellular lipophilic, or adsorptive transcytosis pathways, depending on how the drugs and/or ligands interact with the BBB cell wall through lipophilicity or ionization. Transport with receptor-mediated transcytosis, transport proteins, and efflux pumps usually results from a receptor’s strong affinity toward a drug or ligand, thus enabling uptake. The enzymatic barrier hinders delivery of certain drugs to the brain, because many proteolytic enzymes are present on the surface. Some success has been found in delivery of compounds across the BBB by using receptors that are overexpressed on the BBB [48]. The most common receptors used for brain delivery include transferrin receptors, insulin receptors, low-density lipoprotein receptor (LDL-R)–related proteins (LRPs), diphtheria toxin receptors, heparin-binding EGF-like growth factors, and leptin receptors [36]. Glutathione and choline transport proteins are also prevalent on the BBB and have been used for drug delivery [36]. Two categories of strategies are used to improve BBB penetration: 1) active targeting ligands and 2) cell-penetrating peptides (CPPs). To maximize brain uptake and targeting, researchers have used multiple ligands in single delivery systems [48, 49].
2. BRAIN-TARGETING LIGANDS
One of the most promising ways to non-invasively increase oral drug delivery to the brain involves using active targeting ligands specific to the BBB. These ligands have moieties that attach to and form complexes with receptors on endothelial brain cells, thus facilitating transport across the BBB. Various transport mechanisms can be used, such as receptor-mediated transport, carrier-mediated transport, or adsorptive-mediated endocytosis. Previous studies have demonstrated that dual ligand modification with a drug delivery system results in significantly greater penetration than the use of just a single ligand. The use of a drug delivery system, and hence the protection of the peptide, lessens the need for using specific enzyme inhibitors that can potentially prevent the breakdown of peptides in the body. Only a few ligands with good potential and data have been selected for discussion in this review. Any single method or ligand is typically insufficient for transport across both the GIT and the BBB, owing to the large differences between these two barriers and the need for specificity. Even if a receptor were present on both the GIT and BBB, using that receptor for both pathways with a single ligand would not be ideal. For a given receptor, the specificity to different parts of the body and the receptor density within a tissue may vary. Even with direct intravenous injection, uptake into the brain is minimal; consequently, future attempts to develop oral delivery to the brain should include methods that overcome the GIT and BBB separately.
2.1. Receptor-mediated transport
Receptor-mediated transport mainly includes five categories of receptors: transferrin receptors, lactoferrin receptors, insulin receptors, LRPs, and heparin-binding EGF-like growth factors ( Table 2 ).
Summary of various types of receptors for receptor-mediated transport across the BBB
| Receptor | Structure | Function | Mechanism | Application | Refs. |
|---|---|---|---|---|---|
| Transferrin receptor |
 | Regulation of intracellular iron levels | Receptor for BBB targeting and transport in transcytosis | MAbs increasing brain uptake | [50] |
| Lactoferrin receptor |
 | Anti-inflammatory, antimicrobial, and immunomodulatory effects | Two binding sites with an adsorptive-mediated mechanism | Lactoferrin in Alzheimer’s disease, Parkinson’s disease, and Huntington’s disease | [51, 52] |
| Insulin receptor |
 | Delivery of insulin to regulate blood glucose levels | Tyrosine kinase receptors via transcytosis | Peptidomimetic MAb inhibition | [53, 54] |
| LRP | Apolipoprotein E | Binding to angiopep-2 ligand | LDL-R and LRP transport via endocytosis | Arylsulfatase A transport to the basolateral membrane of the BBB | [55] |
| HB-EGF |
 | Stimulation of BBB growth and differentiation | Diphtheria toxin binding and transport via endocytosis | CRM197 as a carrier protein for delivery of human vaccines | [56] |
2.1.1. Transferrin receptor
The transferrin receptor is a transmembrane glycoprotein consisting of two linked 90-kDa subunits, which bind a transferrin molecule and transport iron via receptor-mediated transcytosis [36, 57]. Transferrin is important for cell proliferation and essential for iron homeostasis; it strongly but reversibly binds iron at two high-affinity Fe (III)-binding sites. Two isotopes of the transferrin receptor exist: TfR1 and TfR2. TfR1 regulates intracellular iron levels with a greater affinity toward iron than TfR2, which only maintains iron levels within the body. TfR1 delivers iron into cells through an endocytic process known as the transferrin cycle. After binding and releasing iron into the cell, the iron-depleted TfR recycles itself to the cell surface. The structures and mechanisms of TfR1 and TfR2 are explained in Table 2 .
The transferrin receptor is the most widely researched receptor for BBB targeting and is highly expressed in erythroid cells, placental tissue, and highly proliferating cells, such as brain endothelial cells. However, concerns exist regarding the use of transferrin as a ligand, because of the high endogenous transferrin concentrations that nearly saturate the transferrin receptors, thus resulting in competition with endogenous transferrin [36]. Nonetheless, success has been reported by Mishra et al., who have used transferrin-PEGylated nanoparticles to deliver approximately 20% of an injected dose to the brain [50]. Because of the limitations of using transferrin, monoclonal antibodies (MAbs) with higher affinity toward transferrin receptors than endogenous transferrin are used instead [58]. The basic structure of MAbs consists of two heterodimeric transmembrane glycoproteins bound to a transferrin receptor, to which transferrin binds and transports iron via transcytosis. The most common MAbs with high affinity to the transferrin receptors are OX26, 8D3, and RI7 [59]. OX26, the best-known MAb, specifically binds the rat transferrin receptor but is thought to be active only in rat species [36, 60]. In a mouse study, approximately 3.1% of the injected dose reached the brain when the 8D3 MAb was used, and only 1.6% of the dose reached the brain when RI7 was used, whereas OX26 resulted in a negligible 0.06% uptake [36, 60]. Both 8D3 and RI7 MAbs are suitable for mouse brain uptake studies, but the RI7 MAb has higher brain selectivity, whereas the liver and kidneys show less ligand retention [36, 60]. Alata et al. have confirmed the unidirectional and fully saturable mechanism of the RI7 transferrin MAb across brain capillary endothelial cells in mice [60]. Paris-Robidas et al. have used quantum dots with the RI7 ligand to show that a therapeutically relevant concentration of target molecule can be transported inside brain capillary endothelial cells, thus indicating that the ligand can overcome the BBB to enhance brain delivery [61]. The RI7 ligand appears to be the most promising, because it demonstrates good penetration and the highest selectivity, thus minimizing the release of neuropeptide drugs in other parts of the body and potentially decreasing any adverse effects [60, 61].
2.1.2. Lactoferrin receptor
The lactoferrin receptor consists of two identical transmembrane proteins (in a homodimer) that facilitate the transport of lactoferrin, which is found throughout the body, including the BBB [52]. Lactoferrin is a single-chain cationic-iron-binding glycoprotein belonging to the transferrin family, which has anti-inflammatory, antimicrobial, and immunomodulatory functions [51, 52]. Two binding sites are present on the lactoferrin receptor: a high-affinity binding site and a low-affinity binding site [62]. Lactoferrin receptors have varying sizes and characteristics in different cell types, including the BBB, thus indicating the possibility of developing a selective lactoferrin-receptor ligand [63]. Transferrin and lactoferrin have similar qualities, but lactoferrin has two main advantages: a lower plasma concentration and a unidirectional brain uptake mechanism. Lower endogenous concentrations of lactoferrin lead to less competition and interference with the lactoferrin receptors in the body. Ji et al. have compared lactoferrin to transferrin and OX26 MAb, and found that lactoferrin has superior brain uptake [64, 65]. Huang et al. have demonstrated that lactoferrin-modified nanoparticles are transported with both receptor and adsorptive-mediated mechanisms that efficiently cross the BBB [66]. Lactoferrin may be of interest in the treatment of neurodegenerative diseases, because the lactoferrin receptor on the BBB is overexpressed in several neurological conditions such as Alzheimer’s disease, Parkinson’s disease, and Huntington’s disease [67, 68].
2.1.3. Insulin receptor
The insulin receptor present on the BBB is a transmembrane receptor belonging to a large class of tyrosine kinase receptors, which delivers insulin to the brain via transcytosis [67]. The insulin receptor is approximately ˜300 kDa and has two alpha-subunits and two beta subunits with a disulfide bond between the alpha and beta subunits, thus creating a cylindrical structure. When insulin binds, a conformational change occurs, thereby enabling tyrosine kinase activity and receptor internalization [53, 54]. Using insulin as a ligand results in two major problems: insulin degradation in the bloodstream, with a half-life of 10 minutes, and interference with natural insulin, thus causing hypoglycemia. These limitations can be addressed by using a peptidomimetic MAb against the human insulin receptor [53, 54]. Wu et al. have reported relatively high drug uptake (delivery of approximately 4% of the injected dose to the brain) with the ligand MAb 83-14, targeting the human insulin receptor [53, 54]. However, this MAb is not effective in rats and has been evaluated only in Old World primates (rhesus monkeys). Because of the mouse origin of MAb 83-14, utilization in humans could lead to various immunogenic responses. To overcome the immunogenicity, a chimeric antibody has been constructed with 85% human origin and 15% mouse origin, and a fully humanized form has also been developed. Humanization of the antibody results in a 27% decrease in affinity but enables direct administration to human patients and has lower adverse effects [69]. Kuo et al. have developed promising solid lipid nanoparticles (SLNs) modified with MAb 83-14, poloxamer 407, and Tween 80, which not only successfully enhance BBB permeability but also promote endocytosis into human brain microvascular endothelial cells [70]. Utilization of the insulin receptor may have potential in overcoming the BBB. The widespread expression of the insulin receptor on peripheral organs may limit its applicability and result in potential toxicity [71].
2.1.4. LRPs
Lipoproteins bind both LDL-R and LRPs, which facilitate lipoprotein transport via endocytosis [72, 73]. On the surface of the BBB endothelium, compared with the peripheral endothelium, LDL-R is upregulated and thus may be a potential target for drug delivery. Low-density lipoprotein (LDL) particles have a highly hydrophobic core with a hydrophilic shell exhibiting multimodal loading capacity in incorporating drugs with differing affinity into its structure. Angiopep-2 is a 19-amino acid ligand for the LRP-1 receptor, which exhibits high transcytosis capacity and can bypass the P-glycoprotein efflux pump [12, 36]. Demeule et al. have determined that angiopep-2 has transcytosis capacity and parenchymal accumulation superior to those of transferrin, lactoferrin, and avidin [74]. Current research has indicated that the utilization of angiopep-2 also depends on the cargo, because attachment to the drug molecule results in good penetration, whereas attachment to a nanoparticular delivery system results in little penetration [7, 75]. A comparative study performed by van Rooy et al. has indicated that both angiopep-2 and COG133 conjugated to liposomes do not significantly target the mouse brain in vivo [76]. Böckenhoff et al. have found that transactivator of transcription (TAT)- and angiopep-2-conjugated arylsulfatase A (ASA) show insignificant transport to the basolateral membrane of the BBB, whereas apolipoproteins B and E (apoB and apoE, respectively) significantly transport ASA to the basolateral side. Both apoE and apoB facilitate lipid transport from the plasma to the central nervous system by using LDL-R and LRPs [55]. Native LDLs, such as apoB, are not suitable, because they are difficult to isolate in large quantities, and they vary in their size and composition. ApoE and apoB have demonstrated efficacy but have inherent protein instability and compete with LDL. Therefore, Laskowitz et al. have developed COG133 as an apoE-mimetic peptide to provide an attractive smaller alternative [77]. These ligands’ lack of brain specificity and their stability issues are major concerns in orally delivery of compounds to the brain.
2.1.5. Heparin-binding EGF-like growth factor
The heparin-binding epidermal growth factor-like growth factor (HB-EGF) is naturally found in the BBB, neurons, and glial cells, in which it stimulates growth and differentiation [36, 56, 75]. This transmembrane protein has no known endogenous ligands, but the diphtheria toxin can bind it and enter cells through endocytosis [12, 36]. Diphtheria toxin enzymatically blocks protein synthesis, thus causing cell death; therefore, it cannot be directly used as a ligand [75]. Cross-reacting material-197 (CRM197) is a non-toxic mutated diphtheria toxin that lacks the enzymatic function of diphtheria toxin but retains the ability to bind HB-EGF [75]. CRM197 has been extensively used as a carrier protein to safely and effectively deliver human vaccines. Consequently, a large pool of information on this ligand exists, including its structure, transport receptor, and mechanism of action [56]. Gaillard et al. have demonstrated in vitro targeting of CRM197 by using liposomal CRM197 [78]. Wang et al. have successfully delivered microvascular endothelial cells to the brain and reported subsequent up-regulation of caveolin-1 expression, mediated by glucose transporter (GLUT) activity, thereby improving transcytosis mediated by caveolae and promoting CRM197-targeted delivery [79]. HB-EGF is upregulated in many inflammatory conditions caused by various brain diseases and thus may be a useful receptor for site-specific disease targeting [36]. In vivo studies with diphtheria toxin have been successful in guinea pig models, but not rats or mice, owing to an amino acid substitution in the receptor-binding domain [12, 56]. Although transgenic mice expressing human HB-EGF are available for experimental purposes, a potential complication of using CRM197 is the individual antibodies against diphtheria toxin, which may decrease the efficacy of the drug delivery system; therefore, the possibility of long-term application remains unclear [12, 56, 80]. CRM197 remains toxic but has approximately 100-fold less potency than that of the wild-type diphtheria toxin.
2.2. Transporter-mediated transport
Transporter-mediated transport uses three different types of transporters: glutathione transporters, choline transporters, and GLUT transporters ( Table 3 ).
Summary of categories of transporters for transporter-mediated transport across the BBB
| Transporter | Structure | Function | Mechanism | Application | Refs. |
|---|---|---|---|---|---|
| Glutathione transporter |
 | Maintain cellular redox homeostasis and suppress oxidative stress | Transport glutathione across the BBB | Improved drug penetration for Alzheimer’s disease | [81, 82] |
| Choline transporter |
 | Synthesize the neurotransmitter acetylcholine | Transport choline for delivery to the brain | Delivery of over-expression plasmids | [83] |
| GLUT transporters |
 | Transport glucose, the essential fuel for neurons | Transport glucose from the blood to the brain | Peptidomimetic MAb inhibition | [84, 85] |
The abundant endogenous tripeptide glutathione is delivered to the brain via glutathione transporters (Table 3). Glutathione is an important cellular antioxidant in maintaining cellular redox homeostasis and suppressing oxidative stress [81, 82]. Rip et al. have reported a four-fold greater drug brain accumulation for glutathione PEGylated liposomes than unmodified liposomes [86]. Rotman et al. have demonstrated the effectiveness of glutathione with their liposomal delivery system, which provides ten-fold higher uptake into the brain than that of free drug. Rotman et al. have also found significantly improved penetration of a single-domain antibody in an Alzheimer’s disease mouse model by using glutathione PEGylated liposomes rather than free drug [87]. Lindqvist et al. have used glutathione PEGylated liposomes and reported a doubling of drug uptake into the brain without use of a specific brain-targeting ligand [88]. Different degrees of PEGylation result in different amounts of uptake and transport, and may sometimes result in higher uptake and decreased transport, because these two processes are similar but have several distinct differences. Literature searches and experimental determination should be used to identify the ideal amount of PEGylation for GIT and BBB uptake specific to a given drug delivery system. Geldenhuys et al. have developed glutathione-coated nanoparticles that provide significantly higher brain uptake than that of drug solution [89]. Utilization of glutathione as a ligand for brain delivery appears promising, but its best quality is its excellent long-term safety profile and low likelihood of hindering physiological function [90].
An essential neurotransmitter, acetylcholine, requires choline, which is transported to the brain via choline transporters for synthesis of the neurotransmitter acetylcholine [83]. Because choline itself is not suitable to modify into a ligand, Li et al. have developed bis-quaternary ammonium compounds with high affinity for the choline transporter [91, 92]. These bis-quaternary ammonium compounds provide effective delivery across brain capillary endothelial cells, and higher gene expression in delivery of plasmid deoxyribonucleic acid (DNA) both in vitro and in vivo, thus suggesting successful targeted brain delivery [93, 94]. The choline derivative compounds appear promising, although limited research has been performed on improving their brain delivery [83, 91].
Glucose is an essential fuel for the brain that must be transported from the blood to the brain by dedicated transporters known as the GLUT transporters [95], because neurons are unable to synthesize or store glucose. GLUT1 and GLUT3 are the main transporters in the mammalian brain, and are present in approximately equal amounts. GLUT1 facilitates glucose transport from the blood to the extracellular spaces in the brain, whereas GLUT3, the major neuronal GLUT, transports glucose from the extracellular space to neurons [84, 85]. Qin et al. have prepared novel glycosyl derivatives of cholesterol that yield three-fold-greater drug concentration of glycosyl liposomes than liposomes [85]. Jiang et al. have developed promising D-glucose-functionalized nanoparticles that significantly increase the transport of nanoparticles across the BBB both in vitro and in vivo [96]. Gromnicova et al. have developed glucose-coated gold nanoparticles with favorable brain selectivity providing three-fold-higher transport of the drug in human brain endothelium than non-brain endothelium [97]. Dufes et al. have developed glucose-functionalized niosomes and demonstrated 86% drug uptake into the brain after 5 minutes of treatment [98]. Although concerns have been raised regarding pathologies, such as Alzheimer’s disease and hyperglycemia, in which the number of GLUT receptors may be altered, overall, these glucose-targeting ligands may be an effective method for brain delivery [84].
3. CELL-PENETRATING PEPTIDES
CPPs are a broad and diverse group of peptides that usually comprise small (˜30 amino acid) peptide chains that ubiquitously cross cellular membranes without a need for receptors [36, 99, 100]. The three main CPPs used are shown in Table 4 . In 1988, the first evidence of CPPs was provided by Green et al., who demonstrated that the viral protein TAT can rapidly translocate across the cellular membrane into the cytoplasm [100, 101]. Although the mechanism of action is not completely known, two pathways with markedly different internalization efficiency have been suggested: direct penetration (energy independent) and endocytosis (energy dependent) pathways. The direct-penetration pathway uses the electrostatic interactions between the CPP and hydrogen bonds on the cell membranes for direct transduction of small molecules across the lipid bilayer. The primary endocytic pathway is pinocytosis, in which large molecules and nanoparticulates cross the cell. Depending on the CPP or the CPP-cargo conjugate, cell entry can occur through single or even multiple endocytic methods, thus leading to different deposition rates in cells [36, 99, 100].
Summary of three typical cell-penetrating peptides applied for brain delivery
| Cell-penetrating peptides | Structure | Function | Mechanism | Application | Refs. |
|---|---|---|---|---|---|
| Transactivator of transcription derived peptides |
 | Transport of large proteins (˜480 kDa) across the BBB | Pore formation with adsorptive-related endocytosis | Transfect siRNA and prevent ischemic damage | [102, 103] |
| Penetratin |
 | Internalization across epithelial cells | Binding via electrostatic interactions, and translocation via tryptophan-induced destabilization | Insulin and doxorubicin delivery | [104] |
| Poly-L-arginine |
 | Enhanced penetration | across the phospholipid membrane through electrostatic interactions and polarization forces | Gene delivery | [93, 105] |
3.1. TAT peptide
TAT-derived peptides ( Table 4 ) originate from the human immunodeficiency virus transcriptional activator and facilitate the transport of large proteins (˜480 kDa) across the BBB [36, 91]. The mechanism of entry has been suggested to partly involve adsorptive-related endocytosis, direct penetration, and pore formation [99]. Cao et al. have constructed a fusion protein of the death-suppressor Bcl-xl with TAT, which significantly decreases neuronal cell death after ischemic damage [102, 103]. Malhotra et al. have developed TAT PEG chitosan nanoparticles and used them to successfully transfect siRNA in an in vitro model after 48 hours [106]. Gregori et al. have demonstrated a three-fold increase in the uptake of their nanoliposomes across human brain capillary endothelial cells after TAT peptide attachment, with no difference in the cell viability [107]. TAT has also been conjugated with many drug delivery systems and yielded mostly positive results, although its lack of specificity may be a concern [102, 103, 108, 109].
3.2. Penetratin
Derived from the Drosophila species, penetratin ( Table 4 ) facilitates internalization across epithelial cells in a two-step mechanism [110, 111]. Penetratin binds cell-surface lipids through electrostatic interactions and is translocated via tryptophan-induced destabilization [110, 111]. Kamei et al. have demonstrated an increase in insulin bioavailability from 35% to 50% after co-administration with penetratin [112]. A comparison study on the effects of three CPPs—TAT, penetratin, and mastoparan—on doxorubicin-loaded transferrin liposomes has been performed by Sharma et al. [104]. Matoparan has higher cytotoxicity and hemolytic problems, whereas penetratin and TAT have high biocompatibility. The transferrin penetratin liposomes have shown more than 90% cellular internalization of the drug with maximum drug translocation across brain endothelial cells [104]. Tseng et al. have concluded that attaching either penetratin or TAT to liposomes increases the intracellular accumulation of the liposomes but does significantly improve drug activity, owing to its low unloading efficiency [113]. Because mixed conclusions have been reported regarding the final outcomes of penetration, this method should be considered with caution.
3.3. Poly-L-arginine
Poly-L-arginine, a synthetic peptide consisting of eight or more arginine residues, is the most readily available and widely used peptide in various drug delivery systems [93, 105]. Nonamers of arginine have better penetration than their truncated analogs and lysine counterparts [93, 114]. Cationic amino acids, such as arginine, adsorb onto the phospholipid membrane of the cell not only by electrostatic interaction but also by polarization forces causing surface rearrangements [93]. An early study by Westergen and Johansson examining three compounds—poly-L-arginine, poly-L-lysine, and protamine—has indicated that poly-L-arginine has the greatest permeability across the BBB [115, 116]. Poly-L-arginine has also been used in gene delivery and has been largely successful in improving transfection, but only in certain cell types [117, 118]. Sharma et al. have developed a bi-ligand liposomal system and investigated the distribution of the liposomes among organs in rats. The addition of poly-L-arginine significantly increases uptake into highly perfused organs and the brain [119]. Poly-L-arginine appears to be a promising CPP; although it is not specific to the BBB, it does improve penetration across.
4. DRUG DELIVERY SYSTEMS
Encapsulating neuropeptides into a drug delivery system, such as niosomes, can overcome protein binding and improve stability and permeability across the BBB, thus resulting in desirable effects on the brain at lower dosages. Drug delivery systems must be considered with respect to applications in oral brain delivery. These systems include polymeric nanoparticles, solid-lipid nanoparticles, liposomes, niosomes, and nanogels. All drug delivery systems have advantages and limitations, but should be able to be modified for effective brain delivery. The ideal characteristics for brain delivery are a particle size < 100 nm, good stability in the GIT and blood, BBB-targeting ability, nonimmunogenicity, nonthrombogenicity, scalability, cost-effectiveness, and ability to incorporate a wide variety of molecules [120]. Nevertheless, drug delivery systems show large differences in their encapsulation, uptake, and transport, for a variety of reasons, including the size, hardness, and elasticity of the formulations. Drug delivery systems are variable in their parameters and behave differently in various drug carriers. If required, these properties can be further modified within a drug delivery system for further optimization and utilization.
4.1. Polymeric nanoparticles
Polymeric nanoparticles ( Figure 3 ) consist of 10–1,000 nm biocompatible and biodegradable polymers that are loaded with drugs [121–123]. Depending on the preparation method, the drug is dissolved, entrapped, encapsulated, or attached to the nanoparticle matrix [124, 125]. Polymeric nanoparticles are advantageous because of their low-cost and easy preparation methods, high drug stability, high flexibility in drug-release parameters, favorable efficacy, and low drug toxicity [121, 125]. The mononuclear phagocyte system rapidly clears polymeric nanoparticles from the blood, thus limiting their distribution in the body; however, this effect can be alleviated by size manipulation and surface modification. The most common natural polymers used are chitosan, gelatin, sodium alginate, and albumin [121, 125]. Synthetic polymers such as poly-lactide co-glycolides, poly-vinyl alcohol, and poly-ethylene glycol have been used [124, 125]. All polymers suggested for fabrication are non-toxic, biodegradable, and biocompatible. Some common methods of preparation include nanoprecipitation, solvent evaporation, salting out, emulsion diffusion, and emulsion evaporation [124, 125]. Calvo et al. have demonstrated the effectiveness of surface modification on polymeric nanoparticles to increase the half-life and therefore penetration of the drug delivery system into the brain [126]. Cheng et al. have significantly improved the brain transport of a centrally active peptide by using polylactic acid nanoparticles [127]. Polymeric nanoparticles are easily tuned, and provide many options regarding materials and methods, thus resulting in effective scalable drug delivery systems.
4.2. Solid lipid nanoparticles
SLNs ( Figure 3 ) are a colloidal drug delivery system with a spherical solid lipid matrix between 10 and 1,000 nm, which are dispersed in aqueous solution [128, 129]. The main advantages of SLNs are the increased scope of drug targeting, long shelf-life, ability to incorporate both hydrophilic and hydrophobic drugs, and controlled release of the incorporated drug for periods as long as several weeks [128–130]. For SLNs, similarly to polymeric nanoparticles, concerns have been raised regarding their uptake into the mononuclear phagocyte system, which might lead to therapeutic failure [131]. The two predominant pathways of SLN uptake are the transcellular lipophilic pathway and the receptor-mediated transcytosis pathway. Ligands can be attached to SLNs to modulate absorption through the different pathways. The nanoparticles can be formulated to have a small particle size and surface modifications to avoid the mononuclear phagocyte system and improve absorption across both the GIT and BBB [132–134]. Some challenges in using SLNs include their possibility of particle growth, unpredictable gelation tendencies, and unexpected dynamics of polymeric transitions [132, 135]. Various preparation methods include high-pressure homogenization, ultra-sonication/high-speed homogenization, solvent evaporation, solvent emulsification-diffusion, the supercritical fluid method, microemulsion, spray-drying, emulsion, precipitation, and film ultrasound dispersion [132–134]. Dhawan et al. have used Tween 80 in quercetin-loaded SLNs to demonstrate strong neuroprotective effects in a mouse Alzheimer’s disease model [136, 137]. Mulik et al. have conjugated transferrin to SLNs and achieved effective in vitro delivery of curcumin [138, 139]. Ramalingam and Ko have shown that trimethyl chitosan coated SLNs improve the oral bioavailability, brain distribution, and stability of curcumin [140]. SLNs are suitable for brain delivery because of their generally lower cytotoxicity, and higher loading capacity and production scalability than most other drug delivery systems [132–134].
4.3. Liposomes and niosomes
Liposomes and niosomes ( Figure 3 ) are vesicular drug delivery systems made up of amphiphilic molecules and cholesterol. Structurally, liposomes and niosomes have a sealed spherical structure ranging in size from nanometers to micrometers [125, 141]. The amphiphilic molecules enclosing the surrounding solution create a bilayer, owing to the water-soluble head and hydrophobic tail [142, 143]. Liposomes and niosomes are similar, but liposomes use phospholipids (e.g., phosphatidylcholine and dipalmitoylphosphatidylcholine) as the amphiphilic molecule, whereas niosomes typically use non-ionic surfactants (e.g., Tween, Span, or Brij) [144]. The common methods of preparation for both liposome and niosomes are thin film hydration, ether injection, reverse phase evaporation, heating, homogenization, freeze-thaw, sonication, and extrusion [125, 144]. Liposomes and niosomes have modifiable characteristics, such as particle size, zeta potential, and stability, that enable their optimization [125, 142, 145]. Vesicular delivery systems less than 100 nm in diameter are preferred for brain delivery, and a zeta potential beyond ±30 mV has been found to improve overall formulation physical stability. Liposomes and niosomes have favorable stability characteristics, and they can protect encapsulated drugs from harsh environments. Further optimization with cholesterol; antioxidants; or coatings such as chitosan, pectin, or Eudragit can improve the stability of delivery systems in vivo [142, 146, 147]. Qin et al. have formulated liposomes with TAT-modified cholesterol and demonstrated promising uptake into brain gliomas in animals [142, 146, 147]. Using in vivo mouse models, Chen et al. have found two-fold-greater brain penetration for lactoferrin-conjugated procationic liposomes than conventional liposomes [52, 148]. Dufes et al. have proposed that glucose-targeted niosomes are a promising carrier for any neuroactive peptides unable to cross the BBB by themselves [98]. Bragagni et al. have reported a significantly higher antinociceptive effect after intravenous delivery of dynorphin-B in N-palmitoyl-glucosamine-functionalized niosomes than dynorphin-B in saline [149]. Many successful instances of brain delivery using both niosomes and liposomes have been reported. The advantages of this system are its biocompatibility, its ability to encapsulate any drug, and the possibility to incorporate various ligands and/or peptides at the surface. Niosomes are generally preferred over liposomes because of their longer half-life, less expensive fabrication, and higher stability [142, 146].
4.4. Microemulsions
Microemulsions ( Figure 3 ) are thermodynamically stable systems of water and oil, which are typically transparent and isotropic, have low viscosity, and are stabilized with a surfactant and a cosurfactant [150]. The stability of microemulsions usually arises from the ultra-low interfacial tension between the water and oil phases. Beyond improving oral bioavailability and enhancing absorption, the main advantages of microemulsions are the spontaneous formation of stable systems with long shelf life, and the ability to load large quantities of lipophilic and/or hydrophilic drugs [150]. Both oil-in-water and water-in-oil microemulsions can be used to enhance the oral bioavailability of drugs, although oil-in-water emulsions are favored for the treatment of brain disorders [150]. The main limitation of microemulsion applications in the pharmaceutical field is the compatibility of each individual component with the high concentrations of surfactants and cosurfactants [151]. Microemulsions have a dispersed domain diameter of approximately 1 to 200 nm, most commonly between 10 and 50 nm [150, 152, 153]. The conventional high-energy emulsification methods for emulsions cannot be used to create microemulsions, because they do not readily produce the required microdroplet sizes [152, 154]. Low-energy methods including phase inversion temperature, phase inversion composition, and emulsion inversion point can form microemulsions [154, 155]. Another method for producing microemulsions is the emulsion titration or dilution method, which uses both high-energy and low-energy techniques. Typically, low-energy methods are limited to non-ionic small-molecule surfactants in high concentrations as their emulsifiers, and they require non-triglyceride oils with very low viscosity [156, 157]. To maximize drug delivery to the brain, microemulsions can be functionalized with various excipients that typically improve circulation time or brain uptake [150]. Valduga et al. have found that a formulation of etoposide oleate with a cholesterol-rich microemulsion not only decreases toxicity but also may enhance brain uptake because of cholesterol’s strong affinity toward LDL-Rs [150, 158]. Kang et al. have developed a microemulsion containing PLGA that provides controlled drug-release characteristics with improved therapeutic efficacy [150, 159]. Many examples of functionalized microemulsions have included different ligands, polymers, lipids, or other components, to further improve drug delivery to the brain. Microemulsions for brain delivery are typically designed for intranasal delivery as a non-invasive and rapid-acting treatment [160, 161]. However, oral drug delivery of microemulsions for brain targeting is an underexplored area in which functionalized microemulsions may have potential.
4.5. Nanogels
Nanogels ( Figure 3 ) are nanoparticles composed of physically or chemically cross-linked hydrophilic polymers in the nanoscale size range [162]. Nanogels are fabricated with natural polymers, synthetic polymers, or a combination thereof [162, 163]. Similarly to hydrogels, nanogels encapsulate small molecules or macromolecules, and exhibit tunable swelling, degradation, and chemical functionality [162]. Nanogel preparations can be made from polymeric precursors through methods such as disulfide cross-linking, amine-based cross-linking, clock-chemistry-based cross-linking, photo-induced cross-linking, and physical cross-linking. Alternatively, nanogels can be prepared from heterogeneous polymerization of monomers via emulsion or inverse emulsion polymerization. Azadi et al. have prepared surface-treated methotrexate-loaded chitosan nanogels that provided considerably greater brain concentrations than those of free drug alone [163]. Blackburn et al. have successfully produced a maleimide-functionalized nanogel that improves siRNA activity through effective protection during endosomal uptake and escape in cells [164, 165]. Soni et al. have shown that nanogels’ substantial penetration into the brain can be further improved with a polysorbate 80 ligand [166, 167]. Nanogels are favorable delivery systems because of their high drug loading, which can reach 50 wt%, and their flexibility in incorporating various ligands into the structure [168].
5. CONCLUSION
Peptide drugs generally cannot be orally delivered to the brain because of their stability and/or the barriers present in vivo. Oral delivery to the brain poses many challenges, in which not only the GIT but also the BBB must be considered. Various ligands with limited comparative data were explored to determine the ligands with the best penetration, applicability to in vivo models, and BBB selectivity. The RI7 ligand has the greatest potential; although it may not have the highest penetration, it is favorable because of its high BBB selectivity. Many researchers have found success in combining more than one ligand, usually an active-targeting ligand and a CPP. Poly-L-arginine is a useful CPP that can serve as a second ligand in drug delivery systems and significantly improve penetration. Niosomes appear to be the most promising formulation type for brain delivery, owing to their surface modifiability, stability, biocompatibility, and flexibility to incorporate any drug molecule. Therefore, ligand-conjugated niosomes are a favorable method for oral delivery of peptides to the brain.