Insulin can be degraded in the presence of protease in the gastrointestinal tract; therefore, preserving the structural integrity of insulin before systemic circulation is a prerequisite for achieving insulin’s therapeutic effect. Thus, insulin requires to be encapsulated in proper vehicles or engineered to resist protease-mediated degradation [12]. Recent advancements in material science have led to the design of various material types, such as synthetic polymers, inorganic materials, and metal-organic frameworks (MOFs), aiming at preparing carriers to improve the stability of insulin within the gastrointestinal environment [11]. The design methodology and properties of each type of material will be introduced below.

The extreme acidic and enzyme-rich environment of gastrointestinal tract predisposes insulin to denaturation. As a result, synthetic polymers that can minimize premature insulin release or reduce enzyme-insulin interaction can maximally retain insulin’s structure integrity and physiological activity. Regarding the capability of acidic groups to be protonated in an acidic environment and become poorly soluble in an aqueous solution, various acid-group-derived materials have been designed and studied. For example, Fahmy and co-workers have prepared polymerized ursodeoxycholic acid for encapsulating insulin (Fig. 2a). Under acidic conditions, this material protonated and became hydrophobic, resulting in reduced penetration of gastrointestinal fluid into particles, inhibiting insulin release and thus protecting insulin from degradation. In the stomach milieu, this formulation has shown minimal insulin leakage [13]. Similarly, Feng and co-workers have reported on the development of a benzoboroxole-containing multi-armed polyethylene glycol (PEG) amphiphilic dendrimer, which became hydrophobic in strongly acidic conditions. When insulin was loaded in this material-based carrier, less than 1% premature insulin release was observed in acidic conditions within four hours (Fig. 2b) [14]. The hyaluronic acid-modified core-shell structure was also reported to maintain insulin stability and showed less than 10% leakage of insulin within 12 h despite strong acidic conditions [15]. In addition, Sun and co-workers have prepared materials based on hydroxypropyl-β-cyclodextrin that has a relatively hydrophobic central cavity and a hydrophilic outer surface. Its host-guest interactions with insulin serve to hinder the degradation of insulin by enzymes [16].

Inorganic materials, including gold and silicon-based materials, can resist the acidic condition and enzymatic degradation in the gastrointestinal tract, protecting the encapsulated insulin [17]. Silicon-based materials have high physiochemical stability, biocompatibility, and tunability [18]. Brayden and co-workers have prepared silica-coated nanoparticles loaded with insulin, zinc, and L-arginine, enabling a slow release of insulin in the acidic environment [19]. In another research, lignin-encapsulated porous silicon nanoparticles were engineered to deliver insulin (Fig. 2c) [20]. The nanoparticles with or without lignin encapsulation both showed negligible insulin release at pH 1.2 and 6.5, demonstrating excellent insulin-protective capability [20].

MOFs are a class of hybrid crystalline materials consisting of organic ligands and metal ions based on reticular chemistry [21]. MOFs can be designed to have porous structures for loading and protecting insulin from harsh gastrointestinal environments. Farha and co-workers have loaded insulin within NU-1000, an acid-stable Zr6- based porous MOF. The NU-1000 was designed with a pore size of approximately 3 nm and was suitable for loading insulin (~ 2 nm in monomeric form) and excluding pepsin (~ 6 nm) (Fig. 2d) [22]. Hence, the insulin-loaded UN-1000 demonstrated minimal release of insulin and strong insulin protection in the simulated stomach acid solution [22]. Similarly, Tian and co-workers have reported an insulin delivery system based on UiO-68-NH₂ with a diameter of ~ 2.6 nm [23]. Insulin-loaded UiO-68-NH₂ released more than 90% of insulin in PBS (pH 7.4) and less than 20% in simulated gastric and intestinal fluids, demonstrating an excellent protective effect of MOFs for insulin [23]. Besides, the release of insulin under varied pH conditions can also be controlled. Chen and co-workers have designed a nanocomposite vehicle based on ironbased MOF nanoparticles (MIL-100) with a 2.6 nm pore size (Fig. 2e) [24]. At pH 7.4, the particle size of insulin tetramer was 2.5 ± 0.3 nm [25]. However, the size of insulin aggregates increased rapidly as the pH approached the isoelectric point. Consequently, the release of insulin was almost inhibited in intestinal fluid (pH 5.4) and accelerated dramatically in the blood (pH 7.4) [24].

In addition to materials encapsulation, engineering insulin with amino acid alterations has also shown a potential to protect insulin from proteolytic degradation [26]. The deletion of B27 Thr and substitutions of amino acid at A14 Tyr and B25 Phe were proven to increase the stability of human insulin in pepsin by 38.5-fold. The alterations of amino acid and lipidation also reduced the insulin receptor affinity, contributing to a longer pharmacokinetic profile. Remarkably, with the above strategies, the insulin analogue OI338 has been tested in Phase 2a clinical trial with a similar hypoglycemic effect as insulin glargine [27].

In general, acid-resistant synthetic polymers and inorganic materials can be used for the oral delivery of insulin to maintain the integrity of the carrier and minimize the premature release of insulin in the stomach. The strong interaction between insulin and these materials can prevent the premature release of insulin and protect it from degradation caused by protease-mediated hydrolysis, thereby preserving the activity of insulin. Moreover, strategic modifications of insulin at specific sites can significantly enhance its resistance to protease degradation. Ensuring the passage of insulin through the stomach with the help of carriers is a crucial requirement for efficient absorption.

The limited permeation through the mucus and intestinal epithelium layers is another challenge for oral insulin delivery. The mucus is a viscous liquid mainly composed of glycosylated and lipid-modified mucin, with negative charges and hydrophobic properties [7]. Therefore, cationic and hydrophobic materials can be easily trapped by the mucus, while neutrally charged and hydrophilic materials, such as zwitterionic and polyethylene glycol-modified materials, can efficiently penetrate mucus [28]. In addition, the size of particles is also a key factor affecting mucus penetration. The large size always impedes mucus permeation due to the mucus’ mesh-like structure with pore sizes in the range of 10 to 500 nm [29]. After getting through the mucus layer, carriers that are cationic and lipophilic can be easily internalized and transported by the intestinal epithelium, despite the contrasting properties required for successful mucus penetration [30]. Hence, designing materials with proper surface properties is crucial to penetrate the mucus layer, pass through the epithelial layer, and deliver insulin into blood circulation. As such, zwitterionic materials, cell-penetrating peptides, and intestinal receptor-specific ligands are incorporated into oral delivery systems to work together to overcome these barriers. In this section, we will focus on the design of materials used to enhance insulin permeability in recent years.

Zwitterionic materials contain abundant pairs of oppositely charged groups, thus obtaining a neutral charge and robust hydration effect [31]. The zwitterionic materials are reported to penetrate mucus easily due to negligible repulsion or attraction by the negatively-charged mucin [32]. Cao and co-workers have prepared an amphiphilic material composed of 1,2-distearoyl-sn-glycero-3-phosphoethanolamine (DSPE) and zwitterionic polycarboxybetaine (PCB) (Fig. 3a) [33]. This material was capable of forming insulin-loaded zwitterionic micelles of about 25 nm diameter in the aqueous phase and had a very low critical micelle concentration, below 10- 6 mM. The zwitterionic micelles displayed excellent mucus permeability, about 12 times larger than PEG-surfaced particles, and enhanced epithelium uptake mediated by the proton-assisted amino acid transporter 1, achieving 42.6% bioavailability of insulin in diabetic rats [33]. In another research, PCB was used to load insulin directly by electrostatic interaction. The PCB/insulin nanoparticles enabled rapid mucus penetration and enhanced epithelium transport, contributing to a high bioavailability of about 27.0% (Fig. 3b, c) [34]. In addition to PCB, Huang and co-workers have reported an insulin delivery system based on dilauroylphosphatidylcholine with a hydrophilic zwitterionic phosphorylcholine headgroup, the carrier based on which was validated for good mucus penetrating ability and enhanced cellular uptake (4.5-fold) compared to PEGylated nanoparticles [35]. In alternative to the direct modification of zwitterionic materials, researchers also modified the surface of nanoparticles with equal amounts of positively and negatively charged fractions to enhance oral insulin delivery. Huang and co-workers have prepared zwitterionic poly (lactic-co-glycolic acid) (PLGA) nanoparticles by coating the cationic octa-arginine (R8) peptide and specific anionic phosphoserine. The virus-like surface achieved enhanced mucus penetration. Subsequently, the anionic phosphoserine was hydrolyzed as catalyzed by phosphatase, making its surface positively charged and thus more easily transported by epithelial cells (Fig. 3d) [36].

In addition, modifying materials with cell-penetrating peptides (CPPs) can enhance the internalization by epithelial cells and the consequent transportation efficiency through the intestinal epithelium. CPPs are amphiphilic and positively charged short peptides that can transport cargo into cells without requiring receptors and energy [37]. Arginine, a semi-essential amino acid, is a major component of many CPPs. Zhang and co-workers modified octaarginine (R8) on the carboxymethyl-β-cyclodextrin for enhanced permeability. A three-fold enhanced transport efficiency on Caco-2 cells was observed compared with that of the carrier without R8 modification [38]. Alonso and co-workers conjugated R8 with cholesterol to form a complex with insulin, and then the complex was enveloped by the protecting polymer poly (glutamic acid)-poly (ethylene glycol) [39]. This system showed a good cellular uptake of insulin (47.59 ± 5.79%) but little transportation (2%). In addition to R8, penetratin is another promising CPPs [40]. Researchers engineered nanoparticles with core-shell structures. The core complex was composed of insulin and penetratin. The exterior was reversibly coated with a hydrophilic N-(2-hydroxypropyl) methacrylamide copolymer (pHPMA) derivatives (Fig. 4a) [41]. This nanoparticle exhibited good mucus penetration owing to the pHPMA coating and a high cellular uptake mediated by the penetratin. As the nanoparticle passed through the mucus, the pHPMA shells gradually dissociated, contributing a 20-fold higher absorption on the mucus-secreting cellular model than free insulin [41].

Moreover, ligand-mediated insulin transport is also a conspicuous strategy for enhancing carriers’ cellular uptake by intestinal epithelium [42]. Among the varied used ligands, fragment crystallizable (Fc) modification is one of the most used approaches in recent years. The Fc receptor could bind and transport immunoglobulin G in a pH-dependent manner, mediating the translocation across the intestinal mucosal barrier [43]. Based on this, Farokhzad and co-workers modified Fc on the poly (lactic acid)-b-poly (ethylene glycol), and this insulin delivery system showed an over 10-times absorption efficiency than nanoparticles without Fc in mice (Fig. 4b) [44]. Subsequently, many researchers modified Fc on the surface of materials to enhance intestinal uptake [15,17,45]. In addition to Fc, albumin has also been utilized to improve binding to the Fc receptor. Sarmento and co-workers conjugated albumin and PLGA-PEG to form insulin-loaded nanoparticles, mediating the transport of insulin across the epithelium (Fig. 4c) [46]. Besides, transferrin (Tf) is also a widely-used ligand for oral insulin delivery [24,47]. For instance, Tf was coated on polymeric nanoparticles by self-assembly to enhance cellular uptake and transepithelial transportation (Figs. 4d, e) [47]. Despite proteins, other ligands such as folate [48], biotin [49], Vitamin B12 [50] and peptides [4,51] could also be used to modify materials and enhance permeability.

Overall, the utilization of hydrophilic material-coated carriers with sizes smaller than 200 nm have shown promising results by minimizing interaction with the mucus mesh and facilitating smooth penetration through the mucus layer. Additionally, the incorporation of cell-penetrating peptides and ligands that target epithelial cells enhances the uptake of nanocarriers by the epithelium, thereby promoting efficient transepithelial transportation. Among the various materials investigated, one standout candidate is PCB. PCB possesses the desired characteristics of hydrophilicity, neutrality in charge, and a strong affinity for epithelial cells. The simultaneous fulfillment of these criteria makes PCB highly desirable for achieving optimal absorption efficiency.

Glucose-responsive release of insulin is an efficient way for patients to overcome the risk of hypoglycemia and achieve desirable diabetes management [52⇓⇓-55]. Therefore, several insulin delivery systems based on glucose-binding molecules, glucose oxidase (GOx), and phenylboronic acid (PBA) have been developed to sense the variation in glucose concentration [56⇓⇓⇓-60]. Hence, researchers have also designed glucose-responsive materials for oral insulin delivery to better control blood glucose levels. For instance, Gu and co-workers have engineered a liposome with a core-shell structure [15]. The PBA conjugated hyaluronic acid shell could reversely bind with the catechol and detach rapidly when the postprandial intestinal glucose increased (Fig. 5a). Subsequently, the exposed Fc could facilitate insulin uptake and reduce the postprandial glucose fluctuations in diabetic mice [15]. Similarly, researchers have prepared insulin loaded nanoparticles based on DSPE-PEG-Mal and a glucoseresponsive polymer poly (L-glutamic acid-co-L-glutamyl phenylboronic acid pinacol ester). The nanoparticles were further modified with Fc (Fig. 5b). The phenylboronic acid pinacol ester could bind with glucose, leading to increased hydrophilicity and negative charge density and realizing the glucose-responsive release of insulin (Fig. 5c) [45]. In addition, GOx and insulin could be loaded into the H₂O₂ sensitive polymer methoxypolyethylene glycol-polymethionine. Upon polymer oxidation under hyperglycemic conditions, insulin was released and regulated high blood glucose levels. Also, the blood glucose level could be maintained even after an oral glucose administration [4]. Trabolsi and co-workers have reported a glucose-responsive material based on an imine-linked-covalent organic framework (nCOF). The nCOF consisted of stacked porous nanosheets with a height of 7 nm and a pores size of 1.7 nm. Insulin (~ 2.5 nm) can only be loaded between the nanosheets, while glucose (~ 0.8 nm) can be loaded inside the nCOF pores. At normoglycemic concentrations, the nCOF pores are filled. Under hyperglycemia, glucose is forced to diffuse through the pores and replace the space between nanosheets, therefore triggering the release of insulin (Fig. 5d) [61]. The plasma insulin concentration of diabetic rats also showed a continuously high level after oral administration. The combination of glucose-responsive materials and oral insulin delivery contributed to a safe and prolonged hypoglycemia effect.

The development of glucose-responsive oral insulin delivery has been limited due to the challenges involved in integrating glucose-responsive elements with oral delivery vehicles. Nevertheless, certain studies have made notable progress by incorporating glucose-responsive moieties with specifically designed materials for oral insulin delivery. These approaches have shown a promising paradigm for enhancing the regulation of blood glucose levels through oral insulin delivery. However, further research is needed to overcome the remaining obstacles and fully realize the potential of oral glucose-responsive insulin delivery.