Self-injectable formulation that enables extend-drug release such as depot formulation has been increasing attention because it can improve patients’ quality of life and reduces the risk of recurrence or deterioration of symptoms. In addition, it would be helpful to reduce the burden on healthcare professionals (Lim et al., 2022). Polymer-based depot formulations such as poly (lactic-co-glycolic acid), PLGA, has been already used, and various PLGA-based long-acting drug products have been approved by the U.S. Food and Drug Administration (FDA) for therapeutic delivery. However, PLGA-based formulations are difficult due to their complex manufacturing process, and most of these formulations exhibit initial burst release followed by slow and incomplete release (Lim et al., 2022).

Lipid formulations, such as liposomes and lipid nanoparticles, have gained increasing attention as vehicles for controlled drug delivery (Seo et al., 2023). Recently, non-lamellar liquid crystal (NLLC) technology has emerged as a novel formulation with sustained release properties (van ’t Hag et al., 2017). NLLCs composed of amphiphilic lipids can be classified into discontinuous cubic phases, such as primitive (Im3m), diamond (Pn3m), and gyroid phases (Ia3d), inversed hexagonal phase (H2), and discontinuous micellar cubic phase (Fd3m) based on their assembly shape (Muszynski et al., 2018). The critical packing parameter (CPP) is used to explain the shape of the amphiphiles, which determines the type of their construction form. The CPP is calculated by V/al, where V is the hydrophobic chain volume, a is the cross-sectional area of the hydrophilic head group, and l is the cross-sectional area of the hydrophobic chain length in the molecule. The drug release rate of the loaded drug is affected by the constructed structure of the NLLC. In particular, Fd3m, which has a closely packed inverse micelles structure, displays slower drug release compared with other NLLC structures. In situ systems that form NLLC structures by contacting water in body fluids may be a preferable formulation for drug delivery with microneedles and needle-free injectors.

Amphiphilic lipids form an NLLC structure, with glycerol monooleate and glycerol dioleate, mono- and di-oleic ester compounds with glycerol as a hydrophilic group having been used (Nielsen et al., 1998). Because these NLLC forming lipids have CPP >1, higher negative curvature of the membrane is induced by the addition, and many strategies have been reported for engineering structures that are a compositional mix of various lipids with different CPPs (Martiel et al., 2014). Changes in the constructed structure and diameter of aqueous channels in the structure are methods to control drug release.

In our previous study, a novel NLLC-forming lipid, mono-O-(5, 9, 13-trimethyl-4-tetradecenyl) glycerol ester, MGE, was used on contact with body fluids (Okada et al., 2021). The formulation consisting of MGE, 1, 2-dioleoyl-sn-glycero-3-phosphoglycerol sodium salt, DOPG, showed that the blood concentration of LA was maintained for 21 days or more after administration. However, a higher injection force was required for injection. As the injection force is a very important parameter for self-injectable formulations (Okada et al., 2021), the injection force should be considered when developing a self-injectable formulation.

Recently, non-ionic surfactants of sorbitan fatty acid esters, such as sorbitan monooleate, have been used as alternative lipids to fabricate NLLC structures (Baez-Santos et al., 2016). Sorbian fatty acid esters are cheap nonionic surfactants, commonly used as food emulsifiers, in cosmetic and pharmaceutical products. In particular, sorbitan trioleate (Span 85) has a large lipophilic moiety and is used as an excellent emulsifier. Liposomes composed of a phospholipid bilayer (CPP = 1) are traditionally used as drug delivery systems, and recently liposomal formulations have been utilized in a depot-injectable formulations (Ki et al., 2014; Rahnfeld and Luciani, 2020).

Controlled drug release by adding Span 85 to phospholipid-based formulations to form NLLCs by inducing higher negative curvature may be an effective procedure. Therefore, in the present study, phospholipid-based gel formulations that turn into an NLLC structures by contacting water in body fluids, including soya phosphatidyl choline (SPC), 1,2-dioleoyl-sn-glycero-3-phosphoglycerol sodium salt (DOPG), and Span 85 were prepared, and leuprolide acetate (LA) was entrapped as a model drug with a middle-sized molecular weight. In addition, injectability force was also investigated to show the usefulness of a self-injectable formulation.

Injectable depot formulations are of increasing interest because they offer a reduction in the number of required injections and maintain therapeutic efficacy by stabilizing blood concentrations. These advantages help in reducing unwanted side effects. In addition, depot formulations have been used for many different medical controls (antipsychotic, substance abuse, and hormonal therapy drugs) (Chaudhary et al., 2019). Polymers and lipids are the major excipients used in depot formulations. PLGA, a commonly used polymer in depot formulations, has different molecular weights and copolymer compositions, providing a wide range of drug release control. However, PLGA formulations are difficult due to their complex manufacturing process, and lipid-based formulations have gained attention instead of PLGA formulation for delivering drugs for long-term-period, in addition to PLGA formulation. Oil-based (Wilkinson et al., 2022) and nanodispersion-based (Altamura et al., 2003) formulations, such as liposomes and emulsions, have been used as lipid-based depot formulations (Lim et al., 2022). Recently, in situ lipid-based crystal-forming system have been assessed (Mei et al., 2018; Li et al., 2019; Pineda-Hernandez et al., 2020). NLLC structures can be formed spontaneously by contacting aqueous fluid, and it may be possible to control the drug release of entrapped drugs from the constructed structure. Drug release controlled by a constructed non-lamellar structure and its interplanar distance have been reported (Clogston and Caffrey, 2005). In addition, entrapped drug stability in the constructed non-lamellar structure has also been reported (Boge et al., 2016). However, the normal viscosity of the formulations is increased by forming an NLLC. When microneedle systems and non-needle-free injection systems as well as conventional needles were used as administration devices into the skin, formulations with a lower viscosity are preferable. Thus, in the present study, in situ crystal-forming systems with a simple composition were investigated with amphiphilic components to provide sustained release.

Many researchers have reported drug release from NLLC nanoparticles, and NLLCs can slowly release entrapped drugs. Drugs can diffuse in the lamella phase along a planner bilayer, whereas in a cubic phase, they can diffuse three-dimensionally. On the other hand, a hexagonal phase has one-dimensional water parallel with the water cylinder, and hydrophilic drugs exist in the polar core of the inversed micellar cube. Therefore, in general, the release rate of drugs exhibit the following order; diffusion in the lamellar structure (D _lamellar) > D _{cubic phase} > D _{hexagonal phase} > D _{micellar cubic} (Muszynski et al., 2018). In the present study, higher Q _7d values seen with F_70:5:25 and F_55:5:40, which exhibited H ₂ structures, were confirmed compared with formulations with an Fd3m structure (F_60:15:25, F_45:15:40, F_65:25:10, F_50:25:25, F_35:25:40, and F_55:35:10). Although this result corresponded with other reports, almost the same Q _7d value was confirmed in even H ₂ constructing formulations such as F_75:15:10 compared with Fd3m constructing formulations. Electrostatic interactions between the charged drug and the polar group in the lipid component may also achieve drug release control (Lynch et al., 2003; Negrini et al., 2014). Therefore, in the present study, formulations containing higher DOPG were listed (Table 1), but no experiments were performed with those with more than 50% DOPG because of their high viscosity.

The F_75:15:10 formulation had a higher proportion of DOPG, which has a positive charge at pH 7.4, compared with F_70:5:25 and F_55:5:40. Thus, negatively charged LA was thought to have a slower release rate. In addition, the F_85:5:10 formulation showed a relatively lower release rate compared with the other formulations. The release rate depended on the constructed structure in addition to the size of the aqueous channel (Muszynski et al., 2018; Zhai et al., 2022) because, in the case of hydrophilic drugs, they must diffuse through an aqueous channel and cross the lipid membrane. Therefore, a smaller aqueous channel in F_85:5:10 may be the reason for the slower LA release rate.

Generally, when the rate-limiting step for drug absorption with aqueous solubility is the rate of drug release; the release rate can be considered to be related to the blood concentration–time profile. However, no blood concentration differences were confirmed between F_55:35:10 and F_65:25:10, although these have different release profiles of LA. On the other hand, the blood concentration of LA after s.c. administration of F_55:5:40, which had a faster LA release rate, displayed the lower limit of detection within 8 h. Figure 5 shows the observation results of the remaining formulation in the body 21 days after administration. The observation was performed by surgically incising around the administration site. Phospholipids and Span 85 were degraded by lipases and non-enzymatic hydrolytic processes (He et al., 2023). Biodegradation of the formulation causes erosion, and subsequent formulation reduction until sufficient excretion is reached. Therefore, the biodegradation process may also affect the drug release rate from the formulation, especially in formulations with sustained drug release. In the present study, retention of the formulation was confirmed 7 days and 28 days after administration. When the F_55:5:40 formulation was administered, no retained formulation was observed (Figure 6A). On the other hand, the F_55:35:10 and F_65:25:10 formulations were observed 7 days after administration (Figures 6B, C, respectively), but these formulations were not clearly observed 28 days after administration (Figures 6D, E, respectively). Since blood concentration of LA was not detected 21 days after administration, only formulation excipients would remain at the injection site. No signs of irritation, inflammation, or any other apparent toxicity were observed grossly at the injection site. In addition, body weight loss was not observed in the present study 28 days after administration compared with just before administration.

Controlled drug release may be attributed to the time required for water penetration into the core of the administrated formulation. In in vitro release conditions, drug release can be achieved without enzymatic degradation. In addition, in in vivo conditions, the transition time to construct an NLLC structure might be delayed in the body because of limited interstitial fluid volume. When the F_55:5:40 formulation was administered, the absence of retention was confirmed 7 days after administration. On the other hand, the F_55:35:10 and F_65:25:10 formulations, which showed almost the same Q _7d value, had a similar blood concentration–time profile. Therefore, it might be possible that the spread of components of the NLLC formulation at the administration site, which affects the construction of the NLLC structure, may be involved as a factor. This might be related to the lack of significant differences in blood concentration–time profiles between the F_55:35:10 and F_65:25:10 formulations.

When F_55:35:10 and F_65:25:10 formulations were administered, AUC _21days values were 2.2 ± 0.42 μg•h/mL and 1.47 ± 0.44 μg•h/mL, respectively. Compared with the AUC _21days after s.c. administration of LA solution, increased AUC _21days was confirmed in both formulations. Ki et al. (2014) have reported AUC value after s.c. administration of a commercial product of PLGA formulation. The calculated AUC was 3.7 μg•h/mL when the dose of 12.5 mg/kg was administered to rats (approximately 300 g body weight). As these AUC values were corrected for administrated LA dosage, the commercial product of PLGA formulation showed about 3 times higher AUC/dose value compared with that of F_55:35:10. In the present study, Span 85 mixed with phospholipids was used to provide sustained release of LA by constructing NLLC formulations. Furthermore, the average maximum force requirement of these formulations was less than 10 N when 29G needle was used. Patki et al. (2021) reported on self-injection formulations with sustained release and concluded that formulations with an average maximum force of 10 N or less are very easy to inject. F_55:35:10 and F_65:25:10 showed that around 7 N when the formulation was injected with 29G needle. In our previous report with NLLC forming lipid of MGE, which has CPP>1 (Okada et al., 2021), a maintained the blood concentration of LA over 21 days or more after administration. However, there was a problem for injectability because they needed a higher expulsion force (higher than 10 N, when the formulation was tested according to the experiment method in 2.7). On the other hand, in the present study, phospholipid formulation showed a maintained concentration of LA until 21 days with a good injectability. Molecular interaction is generally related to viscosity (Pichot et al., 2013), so the difference in injectability properties of the prepared formulation would be related to additives and components in the formulation.

Examination of tail chain length on the effect of the constructed NLLC structure revealed that an increase in CPP value induced phase transition from lamellar or cubic to reverse hexagonal structures (van ’t Hag et al., 2017; Ruela et al., 2016). Because Span 85 has a large lipophilic moiety similar to medium- and long-chain triacylglycerol, thus phase transition to increasing negative interfacial curvature may occur by adding it to a formulation. However, AUC/dose value in our prepared formulation was lower than that in the commercial product of PLGA formulation, suggesting that further formulation development to improve bioavailability should be necessary.

In the present study, in situ NLLC constructed formulations by contacting water were developed using a ternary mixture of SPC, DOPG, and Span 85. In particular, the negative curvature was increased with an increase in the amount of Span 85 in the formulation, and an Fd3m structure was obtained with a sustained release of LA. A maintained blood concentration of LA over 21 day was confirmed by s.c. administration of the formulation. However, the elimination of components of the formulation by enzymatic degradation might need to be considered together with in vitro release experiments for the evaluation of depot formulations with a sustained release. In the present study, evaluation of systemic and local toxicity after s.c. administration was not performed.

This lipid-based formulation can be prepared as a depot formulation with sustained release of the encapsulated drug more simply than polymer-based formulations. Further investigations, such as the detailed mechanism of self-assembly of the NLLC structure by analyzing lipid-lipid and lipid-drug interactions, as well as rheological studies to enable injection, will be useful in the selection of lipid compositions capable of forming the NLLC structures in situ. Although bioavailability equivalent to commercial PLGA products would be required to indicate the usefulness of the phospholipid and Span 85-based formulation, these investigations will be helpful for developing injectable depot formulations with controllable release of drugs such as peptides and proteins.