Brucea javanica-Derived Natural Lipid Droplets: Selective Oral Lymph Targeting and Endocytic Transport Mechanisms

Abstract

Background: Brucea javanica oil (BJO) suffers from poor oral bioavailability due to oxidative degradation and hepatic first-pass effect. Methods: Here, we report a one-step, solvent-free isolation of endogenous Brucea javanica lipid droplets (BJLDs) that function as a “drug-in-carrier” delivery platform. Results: BJLDs exhibited a uniform size distribution and superior oxidative stability. In vitro digestion showed 80% long-chain fatty acids released within 4 h following first-order kinetics. Caco-2 transport studies revealed caveolin-dependent endocytosis as the dominant uptake route and a 2.3-fold increase in rhodamine 123 accumulation versus free drug, indicating potent P-gp inhibition. A cycloheximide-blocked rat model quantified the intestinal lymphatic transport rate at 89.73%. Plasma t_1/2 and MRT of linoleic acid were 8.44 ± 3.16 h and 11.45 ± 2.72 h, respectively. LC-MS/MS confirmed retention of brusatol and bruceine inside BJLDs. Conclusions: This study provides direct evidence that micron-sized lipid droplets derived from plants can achieve >80% lymphatic targeting after oral administration, offering a green and scalable alternative to conventional BJO formulations.

1. Introduction

Brucea javanica oil (BJO), derived from the dried mature fruits of Brucea javanica (L.) Merr., contains high levels of unsaturated fatty acids, including oleic and linoleic acids, and demonstrates well-documented lipid-lowering and broad-spectrum antitumor properties [1,2,3]. Nevertheless, its chemical structure also confers significant limitations: susceptibility to oxidative degradation, poor aqueous dispersibility, and consequently, low oral bioavailability. Although formulation strategies such as emulsions and liposomes have been developed to improve their delivery [4], these conventional methods rely on a two-step process—extraction followed by formulation—that introduces risks of solvent residues, oxidative damage during processing, and potential excipient-related toxicity [5,6]. Thus, an integrated delivery strategy that preserves the stability of BJO and enhances its oral absorption while minimizing additive use remains an urgent need.

Lipid droplets (LDs), natural organelles in eukaryotic cells, have emerged as promising delivery vehicles. Their characteristic structure, which comprises a neutral lipid core enclosed by a phospholipid monolayer decorated with proteins, offers inherent pharmaceutical benefits: the monolayer acts as a natural emulsifier and oxidation barrier, while the core encapsulates lipophilic compounds [7,8,9,10]. Plant-derived LDs, in particular, represent biocompatible carriers for lipophilic active pharmaceutical ingredients [11,12]. Despite these merits, current research on plant LDs suffers from several notable limitations (

Table 1. Research on plant-derived lipid droplets and their delivery systems.

Source	Content	Conclusions	References
Brassica napus	To prepare astaxanthin-loaded lipid droplets and assess their antioxidant activity and cellular activity via in vitro experiments.	To markedly enhance the stability and antioxidant activity of astaxanthin	[11]
Soybean	A pH-driven method was employed to prepare curcumin-loaded oil bodies, and their preparation process and in vitro stability were systematically investigated.	To clarify the stability characteristics of curcumin-loaded oil bodies.	[12]
Almond	To extract oil bodies and explore their changes in the in vitro digestive tract environment.	To elucidate the changes in the physicochemical properties of oil bodies during in vitro digestion.	[15]
Safflower	To prepare oleosin-hEGF-linked oil bodies and evaluate their physicochemical properties, cellular activity, wound healing efficacy, and mechanism of action.	To clarify the effects and molecular mechanisms of drug-loaded oil bodies in accelerating wound healing.	[14]
Brucea javanica	To extract and evaluate its physicochemical properties, in vitro digestive properties, cellular uptake, as well as oral absorption and transport mechanisms.	To clarify the oral absorption and transport mechanisms of BJLDs.	This paper

). First, the plant sources of LDs investigated to date are mostly limited to low-physiological-activity species such as soybean and rapeseed, with relevant studies predominantly confined to in vitro models [13]. Second, existing research on drug-loaded plant LDs has focused primarily on transdermal delivery, with scarce exploration into their oral absorption mechanisms [14]. Third, no plant-derived LD delivery platform has yet achieved a lymphatic transport efficiency exceeding 80%. Finally, the interplay between the LDs surface proteome and key intestinal transporters, such as P-glycoprotein (P-gp), remains largely unexplored.

To address this, we isolated endogenous lipid droplets from Brucea javanica seeds, which naturally contain antitumor unsaturated fatty acids and brusatol. A green, solvent-free density-gradient protocol diverging from conventional “extract-then-encapsulate” methods was employed, which preserved the native oleosin/caleosin corona and its oxidation barrier. The resulting BJLDs were systematically evaluated for their size, stability, digestibility, cellular uptake, and P-gp inhibition. Crucially, their oral absorption and lymphatic transport were quantified in a cycloheximide-blocked rat model, revealing a pathway that enhances the bioavailability of lipophilic compounds by bypassing hepatic first-pass metabolism. This work establishes an innovative oral delivery system for BJO and underscores the potential of plant-derived LDs as a versatile pharmaceutical platform.

2. Materials and Methods

2.1. Materials

Brucea javanica was purchased from Guangzhou Zisun Pharmaceutical Co., Ltd. (Guangzhou, China). Other reagents and solvents used were of analytical grade (Supplementary Material Section S1). Human colorectal cancer Caco-2 cells were purchased from ATCC and maintained in our laboratory.

Male SD rats and male KM mice were obtained from the Animal Experiment Center of Guangzhou University of Chinese Medicine (Permit No. SYXK (Yue) 2024-0202). All animals were housed in an SPF animal facility at Guangzhou University of Chinese Medicine. All experimental procedures complied with the Guangzhou University of Chinese Medicine Experimental Ethics Review Regulations.

2.2. Separation of BJLDs

Reference extraction methods were modified as follows [16,17]. The Brucea javanica was crushed and passed through a 30 mesh sieve. The powder with buffer (1 g of powder per 8 mL) was placed at 4 °C for 12 h, then it was sonicated (390 W, JY92-IIN, ultrasonic processor, Ningbo Scientz Biotechnology Co., Ningbo, China). And then crude Brucea javanica lipid droplets (C-BJLDs) and pure Brucea javanica lipid droplets (P-BJLDs) were obtained by sucrose density gradient centrifugation. The specific steps are shown in the Supplementary Materials (Supplementary Material Section S2).

2.3. Physicochemical Properties and Stability of BJLDs

BILDs were diluted with PBS to a concentration of 8 mg lipid/mL. The particle size and zeta potential were determined using a Malvern Zetasizer Nano ZS (Malvern Panalytical, Great Malvern, UK). For fluorescence imaging, the samples were stained by mixing BJLDs with Nile red and Fast Green FCF solutions (volume ratio of 400:1:1) and incubating for 15 min. Subsequently, the stained samples were embedded in low-melting-point agarose (0.5%, w/v), sectioned, and observed under a laser confocal microscope (LSM 800 with Airyscan, Carl Zeiss, Obercohen Germany) [18,19].

The physical stability of BJLDs was assessed during storage at 4 °C. Changes in particle size and zeta potential were monitored at predetermined time points (0, 6, 12, 24, and 36 h) as key indicators of colloidal stability. The oxidative stability was evaluated under accelerated conditions at 60 °C. BJLDs were diluted to 50 mg lipid/mL with PBS for testing. For comparison, a Brucea javanica oil emulsion (BJE) at the same lipid concentration was prepared. Briefly, Brucea javanica oil (BJO) and phospholipids (at a phospholipid-to-lipid ratio of 7.41%) were dissolved in ether, dried under nitrogen, and then dispersed in PBS by sonication. The oxidative stability of BJLDs, BJE, and pure BJO was determined by measuring malondialdehyde (MDA) levels using a commercial assay kit (Beyotime Biotechnology, Shanghai, China) on days 0, 1, 3, 5, and 7.

2.4. Composition of BJLDs

The contents of neutral lipids, proteins, and phospholipids in lyophilized BJLDs were quantitatively determined. Neutral lipids were extracted using cold acetone and quantified gravimetrically. The remaining fractions were digested, and the protein and phospholipid contents were analyzed using an IC100 chemical analyzer (Innyao Technology Co., Ltd., Hangzhou, China).

Neutral lipids were subjected to saponification and methylation with 0.5 M KOH and 14% BF₃ in methanol. The resulting fatty acid methyl esters were analyzed using an Agilent 7890 GC-MS (Agilent Technologies, Santa Clara, CA, USA). Besides, the protein component of BJLDs was comprehensively characterized. The molecular weight was preliminarily assessed by SDS-PAGE, while the secondary structure was examined using a Nicolet iS5 FT-IR spectrometer (Thermo Fisher, Waltham, MA, USA) [19]. Further identification was performed via LC-MS/MS on an UltiMate 3000 RSLCnano coupled to a Q-Exactive Plus system (Thermo Fisher, Waltham, MA, USA) [20,21].

Given that Brucea javanica contains bioactive constituents such as brusatol and brucein, BJLDs were also analyzed using a TripleTOF 5600 LC-MS/MS system (Thermo Fisher, Waltham, MA, USA) for the detection and identification of these compounds. Detailed methodological procedures are provided in the Supplementary Materials (Supplementary Material Section S3).

2.5. In Vitro Gastrointestinal Digestion and Characterization of BJLDs

2.5.1. Simulated Gastric and Intestinal Phase

The in vitro digestion model was adapted from established methods [13,22] to simulate the gastrointestinal absorption of BJLDs. BJLDs were diluted with PBS to a concentration of 10 mg lipid/mL. Simulated gastric fluid (SGF) was prepared containing 6.9 mM KCl, 0.9 mM KH₂PO₄, 72.2 mM NaCl, 0.1 mM MgCl₂·6(H₂O), and 0.5 mM (NH₄)₂CO₃, with the pH adjusted to 1.6 using 0.1 M HCl. Prior to use, 3.2 g of pepsin was added per 100 mL of SGF. The preheated BJLDs sample was mixed with SGF at a 1:1 volume ratio, and the pH was readjusted to 1.6. The mixture was then incubated at 37 °C with continuous stirring at 100 rpm for 120 min. Aliquots were collected at predetermined time points (5, 15, 30, 60, and 120 min), and the enzymatic reaction was terminated by heating in an 80 °C water bath for 3 min.

The digested sample from the gastric phase was adjusted to pH 7.0 using 2 M NaOH. Simulated intestinal fluid (SIF) was prepared containing 6.8 mM KCl, 0.8 mM KH₂PO₄, 123.4 mM NaCl, and 0.33 mM MgCl₂·6(H₂O), with the pH adjusted to 7.0. Subsequently, bile salts (20 mg/mL) and trypsin (6.4 mg/mL) were added, and the pH was readjusted to 7.0 before use. The gastric-phase digest was mixed with SIF at a 1:1 volume ratio, and the pH was maintained at 7.0. The mixture was incubated at 37 °C with stirring at 100 rpm for 120 min. Samples were taken at specified intervals (5, 15, 30, 60, 120, 180, and 240 min), and the reaction was terminated as described above.

2.5.2. Post-Digestion Characterization

The digested samples from the gastric and intestinal phases were subjected to comprehensive characterization: particle size and zeta potential were measured after 100-fold dilution with 20 mM citrate buffer (pH 1.60) and 20 mM phosphate buffer (pH 7.00), respectively [15]; for fluorescence microscopy observation, samples were stained with Nile red solution (1 mg/mL in DMSO) at a volume ratio of 200:1, incubated in the dark for 15 min, and then examined under a fluorescence microscope (Motic AE31E, MOTIC CHINA GROUP CO., LTD., Xiamen, China) [18]; additionally, the extent of free fatty acid release was quantified by adding 20 mL of an ether/ethanol solution (1:1, v/v) and 40 µL of phenolphthalein indicator to 1 mL of the sample, followed by titration with 0.01 M KOH solution until the endpoint was reached [23]. The percentage of free fatty acid release was calculated according to the formula below:

$$\mathrm{F}\mathrm{F}\mathrm{A}\mathrm{\%}={\displaystyle \frac{\mathrm{C}\times \left({\mathrm{V}}_1-{\mathrm{V}}_2\right)\times \mathrm{M}}{3\times \mathrm{m}}}\times 100$$

(1)

Note: C(M): concentration of KOH titrant; V₁(L): volume of titrant consumed in titrating the sample; V₂(L): volume of titrant consumed in titrating the blank sample; M(g/mol): relative molecular weight of the lipid; m(g): mass of the lipid in 1 mL of the sample.

2.6. Cellular Uptake

The cellular uptake and transport efficiency of BJLDs (labeled with coumarin-6, 15.06 μg/mL) in Caco-2 cells were systematically evaluated under various conditions, including concentration, temperature, and incubation time. First, the cytotoxicity of BJLDs on Caco-2 cells was assessed using the CCK-8 assay after treatment with a range of concentrations (0.125 to 4 mg lipid/mL). Based on the cytotoxicity results, a concentration of 2 mg lipid/mL was selected for subsequent uptake experiments.

For fluorescence imaging, Caco-2 cells were seeded in 24-well plates at a density of 1 × 10⁵ cells per well and cultured for 24 h. The culture medium was then replaced with fresh medium containing BJLDs (2 mg lipid/mL) and incubated for 0.5, 1.0, 2.0, and 4.0 h. After incubation, the medium was removed, and the cells were washed twice with PBS, fixed with 4% paraformaldehyde for 15 min, and washed again with PBS. The cells were then mounted with an anti-fade sealing medium containing DAPI and observed under a fluorescence microscope (OLYMPUS BX53F, Olympus Corporation, Tokyo, Japan). The fluorescence intensity was semi-quantified using ImageJ 2.0 software. The same methodology was applied to investigate the effects of temperature (4 °C and 37 °C) and various endocytosis inhibitors (10 μg/mL chlorpromazine, 2 mg/mL methyl-β-cyclodextrin, 40 μg/mL quercetin, and 10 μg/mL amiloride) on the cellular uptake of BJLDs.

2.7. Cellular Efflux

This study investigated the potential of BJLDs as a drug delivery system to inhibit P-glycoprotein-mediated drug efflux in Caco-2 cells, using Rhodamine 123 (Rho123) as a model substrate. The cytotoxicity of the formulations was first evaluated using the CCK-8 assay. The intracellular accumulation of Rho123 was assessed in four groups: Blank Control, Free-Rho123, Free-Rho123 + Verapamil (50 µM), and Rho123-loaded BJLDs. Caco-2 cells were treated with the corresponding medium for 1 h. The cells were then fixed, and the subcellular distribution of fluorescence was visualized using a fluorescence microscope. The fluorescence intensity and co-localization were analyzed with ImageJ software.

Furthermore, the efflux inhibition was quantified by flow cytometry. Caco-2 cells were seeded in 24-well plates and cultured for 14 days to form confluent monolayers. On day 15, the cells were treated with the drug-containing medium for 1.5 h, harvested, and analyzed for Rho123 fluorescence intensity using a DxP Athena^® flow cytometer (Cytek Biosciences, Fremont, CA, USA).

2.8. Intestinal Distribution

The intestinal distribution of BJLDs was assessed using a chylomicron flow-blocking mouse model combined with near-infrared imaging. Thirty-six mice were fasted for 12 h with free access to water prior to the experiment and randomly allocated into three groups: the model group, the experimental group, and the control group.

The model group received an intraperitoneal injection of cycloheximide (4.2 mg/kg, 0.42 mg/mL) to establish the chylomicron blockade, while the other two groups were administered an equivalent volume of saline. After one hour, both the model and experimental groups were given BJLDs (50 mg lipid/mL, containing 10.6 μg/mL DiR) by oral gavage at a dose of 0.14 mL/10 g body weight. The control group received an equal volume of free DiR solution.

At 1, 2, 4, and 6 h after administration, the mice were euthanized by cervical dislocation. The entire small intestine was then excised, rinsed thoroughly to remove intestinal contents, and immediately subjected to fluorescence imaging using an IVIS system (PerkinElmer Inc., Waltham, MA, USA) to evaluate the spatial and temporal distribution of BJLDs.

2.9. Lymphatic Transport of BJLDs

To investigate the absorption and transport of BJLDs within intestinal lymphoid tissues, a chylomicron flow-blocking rat model was employed in conjunction with laser confocal microscopy. Nine SD rats were fasted for 12 h and randomly assigned to three groups. The model group received an intraperitoneal injection of cycloheximide (3 mg/kg, 0.6 mg/mL) to inhibit chylomicron flow, while the other groups received saline. After one hour, both the model and experimental groups were administered BJLDs (50 mg lipid/mL, containing 15.06 μg/mL coumarin-6) by oral gavage at a dose of 1 mL/100 g body weight. The control group received an equivalent volume of free coumarin-6 solution.

The rats were euthanized after four hours, and segments of the duodenum, jejunum, ileum, and mesenteric lymph nodes were collected for frozen sectioning. The tissue sections were fixed with 4% paraformaldehyde, washed with PBS, and mounted with an anti-fade sealing medium containing DAPI for nuclear counterstaining. Fluorescence imaging was performed using a laser confocal microscope (LSM 800 with Airyscan, Carl Zeiss AG, Obercohen, Germany).

2.10. Pharmacokinetics

Twelve SD rats were fasted for 12 h with free access to water and randomly divided into a model group and a normal group. One week prior to drug administration, baseline blood samples were collected from all rats via the orbital venous plexus at 0.083, 0.25, 0.5, 1, 2, 4, 6, 8, 12, and 24 h. The samples were placed in sodium heparin-coated tubes, centrifuged at 3000× g for 15 min at 4 °C, and the resulting plasma was stored at −20 °C to serve as the self-control for each animal.

On the day of the experiment, the model group received an intraperitoneal injection of cycloheximide (3 mg/kg, 0.6 mg/mL), while the normal group received an equal volume of saline. After one hour, all rats were administered BJLDs (50 mg lipid/mL) by oral gavage at a dose of 1 mL/100 g body weight. Blood samples were subsequently collected at the same predetermined time points as the baseline phase, and plasma was separated and stored at −20 °C until analysis.

The concentration of linoleic acid in plasma was quantified using an Agilent 7890 GC-MS (Agilent Technologies, Santa Clara, CA, USA). The net blood concentration of methyl linoleate was calculated by subtracting its pre-dose baseline level from the post-administration value at each corresponding time point. Pharmacokinetic parameters were derived using PKSolver 2.0 pharmacokinetic software, and the lymphatic transport rate of BJLDs was calculated with the following equation:

$$\mathrm{L}\mathrm{y}\mathrm{m}\mathrm{p}\mathrm{h}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{c} \mathrm{t}\mathrm{r}\mathrm{a}\mathrm{n}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{r}\mathrm{t} \mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}={\displaystyle \frac{{\mathrm{A}\mathrm{U}\mathrm{C}}_{0-\mathrm{\infty }}\left(\mathrm{u}\mathrm{n}\mathrm{b}\mathrm{l}\mathrm{o}\mathrm{c}\mathrm{k}\mathrm{e}\mathrm{d}\right)-{\mathrm{A}\mathrm{U}\mathrm{C}}_{0-\mathrm{\infty }}\left(\mathrm{b}\mathrm{l}\mathrm{o}\mathrm{c}\mathrm{k}\mathrm{e}\mathrm{d}\right)}{{\mathrm{A}\mathrm{U}\mathrm{C}}_{0-\mathrm{\infty }}\left(\mathrm{u}\mathrm{n}\mathrm{b}\mathrm{l}\mathrm{o}\mathrm{c}\mathrm{k}\mathrm{e}\mathrm{d}\right)}}$$

(2)

2.11. Statistical Analysis

Unless otherwise stated, data are reported as mean ± standard deviation. Statistical significance was determined by one-way ANOVA (IBM SPSS Statistics 26), with p < 0.05 considered statistically significant.

3. Results and Discussion

3.1. Isolation, Characterization, and Physicochemical Evaluations of BJLDs

As depicted in Figure 1A,B, BJLDs appeared milky white in color and exhibited a negative surface charge, with a particle size below 2 μm. The size distribution ranged from 0.5 to 2.5 μm, and the negatively charged surface was consistent with the typical charge behavior of LDs surface proteins in neutral systems [17]. To improve sample purity and dispersity, impurities were removed using a 9 M urea solution. All in all, both the particle size and PDI of purified BJLDs (P-BJLDs, Size: 1645.67 ± 13.01 nm, PDI: 0.249 ± 0.032, Zeta: −13.33 ± 0.32 mV) were significantly lower than those of crude BJLDs (C-BJLDs, Size: 1965.67 ± 104.51 nm, PDI: 0.338 ± 0.028, Zeta: −13.70 ± 0.07 mV), supporting the effectiveness of urea in removing impurities. In contrast, the zeta potential remained largely unchanged, indicating that urea treatment did not alter the protein charge profile.

The internal structure of BJLDs was visualized by laser confocal microscopy after dual staining with Nile Red and Fast Green FCF (Figure 1C). The neutral lipid core appeared in red, surface proteins in green, and overlapping regions in yellow, confirming a core–shell architecture. Morphologically, BJLDs were predominantly spherical or oval, with a uniform size distribution between 1 and 2 μm.

The physical stability of BJLDs was assessed during storage at 4 °C for 36 h. As shown in Figure 1D, neither the average particle size nor the zeta potential changed significantly, indicating good colloidal stability. Oxidative stability was further evaluated at 60 °C over seven days by measuring malondialdehyde (MDA) levels (Figure 1E). Throughout the first three days, MDA in BJLDs remained undetectable, reflecting strong oxidative resistance. In contrast, both Brucea javanica oil (BJO) and its emulsion (BJE) showed a marked increase in MDA. This divergence from trends reported for Echium plantagineum LDs [24] may be attributed to differences in preparation processes. Overall, BJLDs demonstrated superior oxidative stability compared to both emulsion and pure oil.

Compositional analysis revealed that neutral lipids, proteins, and phospholipids together accounted for 95.96% of purified BJLDs, indicating high purity. The phospholipid content (6.92–7.41%) fell within the 0.18–17% range typical of monolayer membrane structures [25]. Fatty acid profiling identified seven long-chain species in BJLDs (Figure 2A), predominantly oleic acid (53%) and linoleic acid (21%), followed by palmitic acid (9%) and stearic acid (8%); other fatty acids each constituted about 2% (Supplementary Material Section S4). SDS-PAGE indicated that BJLDs proteins ranged from 8 to 30 kDa, with a major band near 15 kDa (Figure 2B). This aligns with reports that oleosin and caleosin, key structural proteins in plant LDs, typically fall within 15–30 kDa and contribute to colloidal stability [26].

FT-IR analysis (Figure 2C) further characterized the protein structure. Both purified and crude BJLDs showed characteristic amide I, II, and III bands, along with peaks corresponding to N-H, C-H, and C=O stretches, consistent with nitrogen-containing biomolecules. The secondary structure in Figure 2D was rich in α-helix (36.4%) and β-turn (33.8%), with moderate β-sheet (29.8%) and no random coil, suggesting a tightly folded, stable conformation [27]. The minimal structural change after urea washing confirms that the lipid droplet integrity was preserved, corroborating prior findings [28].

Due to the absence of BJLD protein entries in the NCBI database, direct sequence matching was not feasible. However, plant LD proteins often share conserved structural domains and high homology [29,30]. By comparing with well-studied species such as soybean and Arabidopsis, four LD proteins showed high peptide similarity to BJLD proteins (

Table 2. Protein identification in BJLDs by LC-MS/MS.

Accession ID	Description [Species]	MW [kDa]	Calc.pI	Score Sequest	Unique Peptides
Q9LII2	Oleosin family protein [Arabidopsis thaliana]	18.1	9.35	11.91	2
I1MUH0	Oleosin [Glycine max]	16.7	9.17	19.16	5
I1L364	Caleosin [Glycine max]	19.2	6.62	9.77	1
A0A0R0I7K9	Caleosin [Glycine max]	22.8	8.13	3.18	1

), primarily identifying as oleosin and caleosin in the 16–22 kDa range. Thus, we preliminarily conclude that BJLD proteins are mainly composed of oleosin and caleosin.

Additionally, LC-MS/MS analysis confirmed the presence of seven active components in BJLDs, including bioactive brusatol and bruceine (Supplementary Material Section S5) [31].

3.2. Elucidating the Gastrointestinal Absorption and Transport of BJLDs In Vitro

3.2.1. Behaviors of BJLDs in Simulated Gastrointestinal Fluids (SGFs)

In SGF, the zeta potential of BJLDs shifted dynamically from negative to positive. This transition is attributed to the protonation of membrane proteins under acidic conditions (pH < isoelectric point). As pepsin hydrolysis progressed, the zeta potential gradually decreased, reducing electrostatic repulsion between droplets and leading to aggregation, as reflected by the increase in particle size over time (Supplementary Material Section S6).

Upon transition to simulated intestinal fluid (SIF, pH 7.0), the surface charge of BJLDs reverted to negative and its absolute magnitude increased progressively, accompanied by a continuous decrease in particle size. This can be explained by the adsorption of bile salts, which imparts a negative charge and enhances colloidal stability. As enzymatic hydrolysis proceeded, the reduction in droplet size increased the specific surface area, further promoting bile salt adsorption and elevating the zeta potential [32].

Fluorescence microscopy corroborated these trends. In SGF (Figure 3A), BJLDs aggregated due to diminished inter-particle repulsion, though the core–shell structure remained intact, as phospholipids and long-chain fatty acids are not substrates of pepsin. In contrast, no aggregation was observed in SIF; instead, numerous small round structures-likely micelles, vesicles, or fragmented lipid droplets-were visible (Figure 3B).

In vitro release studies indicated that over 80% of free fatty acids were released from BJLDs (Supplementary Material Section S7). The release profile followed first-order kinetics (Regression equation: y= −0.0044x − 0.8544; R² = 0.9227). The absorption rate constant of BJLDs (K_a: 0.262 ± 0.047 h⁻¹) did not differ significantly from that of the emulsion (K_a: 0.226 ± 0.048 h⁻¹), suggesting similar lipid digestion kinetics between the two formulations.

3.2.2. Cellular Uptake and Transport Mechanisms in Caco-2 Cells

Caco-2 cells, which resemble human intestinal epithelia, were used to evaluate the absorption and transport behavior of BJLDs [33]. Cytotoxicity assays confirmed that C6-BJLDs were non-toxic to Caco-2 cells at concentrations up to 2 mg/mL. Cellular uptake of C6-BJLDs was time-dependent and significantly greater than that of free C6 (Figure 4A), indicating that BJLDs enhance the absorption of highly lipophilic compounds (log P > 3). Uptake was markedly suppressed at 4 °C (Figure 4B), suggesting an energy-dependent process such as active transport or endocytosis. Furthermore, methyl-β-cyclodextrin, a caveolae-mediated endocytosis inhibitor, significantly reduced intracellular C6-BJLDs levels, whereas other inhibitors had no effect (Figure 4C).

These results suggest that the cellular uptake of BJLDs is most likely mediated via a caveolin-dependent pathway. Nanoparticles are typically internalized by cells mainly through endocytosis, whereas microparticles (>1 μm) are primarily taken up via phagocytosis [34]; however, the endocytic pathway is not strictly determined by particle size. For instance, previous studies have demonstrated that HeLa cells can endocytose 1–3 μm particles through an actin-dependent mechanism involving clathrin- or caveolin-mediated uptake [35]. Notably, caveolin-dependent internalization is known to facilitate the transport of lipoproteins and lipids [36], which also implies a potential entry pathway for 1.6 μm BJLDs into cells. Methyl-β-cyclodextrin disrupts lipid raft structures by depleting cholesterol from the cell membrane, thereby impairing caveolin function [37]; this finding further supports the aforementioned hypothesis. However, given that β-cyclodextrin may cause extensive interference with cholesterol-dependent physiological processes, the involvement of other endocytic pathways cannot be completely excluded, and the underlying transport mechanisms of BJLDs remain to be further elucidated.

3.2.3. Inhibition of P-gp Mediated Efflux

Caco-2 cells express P-glycoprotein (P-gp), an ATP-dependent efflux transporter. Verapamil, a P-gp inhibitor, and Rho123, a P-gp substrate, were used to assess whether BJLDs can counteract drug efflux [38]. Rho123-BJLDs showed no cytotoxicity below 2 mg/mL. Confocal imaging revealed that both the inhibitor (verapamil) and BJLD-based formulations facilitated Rho123 accumulation near the nucleus, whereas free Rho123 was poorly retained due to P-gp efflux (Figure 4D). Flow cytometry confirmed that BJLDs significantly enhanced Rho123 delivery in differentiated Caco-2 cells with high P-gp expression (Figure 4E), supporting their role as efflux-inhibiting carriers.

Combined with published literature and our experimental findings [39,40,41], BJLDs may reverse P-gp-mediated multidrug resistance via two distinct pathways: (1) The abundant unsaturated fatty acids in BJLDs competitively bind to the functional sites of P-gp, directly abrogating its drug efflux activity; (2) The native monolayer phospholipid membrane of BJLDs fuses with the cytoplasmic membrane, which modulates membrane fluidity and induces conformational changes in P-gp, thereby impairing its substrate binding capacity or ATP hydrolysis process. Collectively, BJLDs may exert a P-gp inhibitory effect through dual pathways. Whether the structural features of BJLDs or their endogenous components dominate this mechanism of action remains to be further investigated. Notably, the monolayer phospholipid membrane and neutral lipid components that underpin these mechanisms are inherent structural features of lipid droplets, which provide a critical theoretical and experimental reference for the rational design and development of other plant-derived lipid droplet delivery carriers.

3.3. Investigating the Intestinal Distribution and Lymphatic Transport of BJLDs

The intestinal distribution and lymphatic transport pathways of BJLDs were investigated using two lipophilic fluorescent tracers: coumarin-6 (green emission) and DiR iodide (near-infrared emission). Cycloheximide (CHX), a known inhibitor of lymphatic transport that does not affect other pathways, was employed to specifically block chylomicron-mediated uptake [42]. Near-infrared imaging of the small intestine revealed that the fluorescence signal peaked at 4 h post-administration, with the ileum segment exhibiting the most intense signal (Figure 5A). Throughout the observation period, the signal intensity in the unblocked group was consistently higher than that in the CHX-blocked group (Figure 5B). This spatial and temporal distribution was corroborated by laser confocal microscopy, which showed stronger fluorescence from BJLDs in the ileum and mesenteric lymph nodes of the unblocked group compared to the blocked group (Figure 6).

The concentration of BJLDs in the ileum suggests this segment is the primary site of absorption, likely due to the abundance of Peyer’s patches, which facilitate particle uptake via M cells [43]. The markedly stronger signal in the mesenteric lymph nodes of the unblocked group confirms that BJLDs predominantly enter systemic circulation via the lymphatic pathway. This is consistent with the known fate of long-chain fatty acids, which are incorporated into chylomicrons and transported through the lymph, unlike short- and medium-chain fatty acids that enter the portal blood directly [44]. As BJLDs are rich in long-chain fatty acids, their predominant lymphatic transport aligns with established lipid absorption physiology.

Cycloheximide is a commonly used lymphatic blocker, widely applied in studies related to drug lymphatic transport. As a global protein inhibitor [45], this reagent may exert certain off-target effects, such as impairing the function of intestinal epithelial cells; these effects could further interfere with the drug lymphatic transport process and potentially compromise the evaluation of lymphatic transport efficiency. However, previous studies have confirmed that cycloheximide does not interfere with other absorption pathways and exhibits no obvious adverse reactions [43]. Thus, its application will not affect the judgment of the core conclusion regarding “whether the drug is transported via the lymphatic pathway”.

3.4. Intestinal Lymphatic Pathway as the Primary Route for Oral Absorption of BJLDs

To evaluate the pharmacokinetic profile and quantify the lymphatic transport of BJLDs, plasma drug concentrations were determined by GC-MS in a chylomicron flow-blocking rat model. Linoleic acid, a major and analytically stable component of BJLDs, was selected as the marker for plasma analysis. The assay demonstrated excellent linearity over the concentration range of 10–320 µg/mL, with a regression equation of y = 0.0037x − 0.044 (R² = 0.9971).

Cycloheximide was used to selectively inhibit the lymphatic transport of drugs by blocking chylomicron formation and secretion, without interfering with other absorption pathways. As shown in the plasma concentration-time curve (Figure 7), a significant difference was observed in the pharmacokinetic behavior of BJLDs between the unblocked and blocked groups. The plasma concentration in the unblocked group was consistently higher at all-time points. Pharmacokinetic analysis revealed that the area under the curve (AUC) was 251.88 ± 25.92 µg/mL·h for the unblocked group, compared to 132.76 ± 23.55 µg/mL·h for the blocked group. Based on the AUC difference, it was calculated that 89.73% of the orally administered BJLDs entered the systemic circulation via the intestinal lymphatic pathway (Supplementary Material Section S8).

The above-mentioned results demonstrated that the intestinal lymphatic system serves as the predominant route for the oral absorption of BJLDs.

4. Conclusions

This study successfully establishes a novel oral delivery strategy based on naturally derived BJLDs, with key innovations manifested in three aspects. First, a green, solvent-free extraction protocol was developed that preserves the native core–shell structure of BJLDs, endowing them with exceptional colloidal stability and superior oxidative resistance, thereby fundamentally addressing the instability issues associated with conventional BJO formulations. Second, we systematically elucidated the oral absorption pathway of BJLDs, which involves caveolin-mediated endocytosis into intestinal epithelial cells, inhibition of P-glycoprotein efflux, and efficient incorporation into chylomicrons, ultimately achieving targeted lymphatic delivery with a remarkably high transport rate of 89.73% and significantly enhanced bioavailability.

Most notably, BJLDs exemplify an advanced “drug-in-carrier” design, functioning dually as both natural biomimetic carriers and intrinsic reservoirs of bioactive constituents (e.g., brusatol and bruceine), eliminating the need for synthetic excipients and aligning with the principles of green pharmacy. The strategy and methodology established in this work demonstrate broad applicability and can be extended to other oil-rich plant seeds, laying a theoretical and technical foundation for a new class of plant-derived natural drug delivery systems.