Molecular targeting of FATP4 transporter for oral delivery of therapeutic peptide

See allHide authors and affiliations

Science Advances  01 Apr 2020:
Vol. 6, no. 14, eaba0145
DOI: 10.1126/sciadv.aba0145


Low oral bioavailability of peptide drugs has limited their application to parenteral administration, which suffers from poor patient compliance. Here, we show that molecular targeting of the FATP4 transporter is an effective approach to specifically transport long-chain fatty acid (LCFA)–conjugated peptides across the enterocytic membrane and, thus, enables oral delivery of drug peptides. We packaged LCFA-conjugated exendin-4 (LCFA-Ex4) into liposomes and coated with chitosan nanoparticles to form an orally deliverable Ex4 (OraEx4). OraEx4 protected LCFA-Ex4 from damage by the gastric fluid and released LCFA-Ex4 in the intestinal cavity, where LCFA-Ex4 was transported across the enterocyte membrane by the FAPT4 transporter. OraEx4 had a high bioavailability of 24.8% with respect to subcutaneous injection and exhibited a substantial hypoglycemic effect in murine models of diabetes mellitus. Thus, molecular targeting of the FATP4 transporter enhances oral absorption of therapeutic peptides and provides a platform for oral peptide drug development.


Oral delivery is the preferred route of drug administration; yet, it is not compatible with a large number of biological drugs due to low oral availability (1). Consequently, parenteral injection is the only route of administration for most protein and peptide drugs in clinic, although it suffers from poor patient compliance and high treatment costs (2). Any breakthrough in this area may lead to development of new therapeutic agents. Oral semaglutide (Rybelsus) was recently approved by the U.S. Food and Drug Administration as the first oral peptide drug, but improvement in oral bioavailability (1 to 2.5% in dogs and barely detectable in human) is still needed (3). The key factor that limits development of oral delivery of protein and peptide drugs is the existence of multiple physical and biological barriers in the gastrointestinal (GI) tract, including the harsh acidic environment in the gastric fluid, multiple digestive enzymes, the thick mucosal layer, and the tightly aligned intestinal epithelial cell (4). Although attempts have been made in the last decade to address the problem of oral delivery of peptide drugs, most focused on only one or a few barriers. For example, plenty of efforts have been devoted to generate instantaneous disruption on the membrane of GI epithelium to enhance passive absorption of peptide drugs (57). Yet, this approach may cause nonspecific absorption of toxins and pathogens into the bloodstream as a result of instantaneous disruption of GI epithelium membrane, and thus, carries risk in long-term treatment (8). Another strategy is to incorporate peptide drugs into receptor or transporter targeting nanoparticles to enhance active transportation of drug peptide across the intestinal epithelium (9, 10). However, the big-size nanoparticles (usually in the 50- to 200-nm range) can be blocked by the thick mucosal layer before they get into contact with epithelial cells (11, 12). In addition, the big size also compromises their efficient transportation by transporters [pore sizes of transporters, 4 to 6 nm (13)] or receptor protein molecules (14, 15). Furthermore, neither current receptors nor transporters are ideal targets for oral delivery of peptide drugs due to their low transport efficiency and limited expression levels in the small intestine (10). Moreover, receptor-mediated endocytosis of nanoparticles normally experiences lysosomal degradation (16), which further reduces bioavailability of peptide drugs. Therefore, there is an urgent need for new strategies to overcome the sequential barriers to enhance oral bioavailability of peptide drugs.

Triglycerides are a major part of fat intake in our diet. They are digested into fatty acids and glycerol in the GI track (17). Fatty acid transport protein 4 (FATP4; also known as SLC27A4) is a unique transporter in enterocytes that mediates long-chain fatty acid (LCFA; chain length, >C12) absorption into the intestinal epithelium (1820). Once inside the cell, LCFAs are trafficked into the endoplasmic reticulum (ER) and excluded from the basolateral of the enterocyte. They enter blood circulation through the lymphatic blood vessel system (21). High efficiency of this route of transport is demonstrated by the over 95% absorption rate of LCFA in our daily diet (22). The expression level of FATP4 is also the highest among intestinal transporters (23). On the basis of these merits, the FATP4 transporter is an ideal and rational target to explore active peptide absorption across the intestinal epithelium.

Herein, we propose a strategy that exploits molecular targeting of the FATP4 transporter for oral absorption of peptide drugs. Molecular targeting of FATP4 is achieved through direct conjugation of LCFA onto the peptide drug. Conjugation of LCFA to peptide is expected to protect the peptide from degradation by the digestive enzymes, thanks to the long hydrophobic carbon chains of LCFA (24). To further protect the LCFA-peptide conjugates from acidic gastric fluid and digestive enzymes, they are encapsulated into liposomes that are coated with chitosan nanoparticles (ChiNPs). LCFA-peptides are released in the pH-neutral small intestine as a result of the sponge effect from chitosan. Since the size of a peptide is normally in the 1- to 2-nm range (25), it permits diffusion across the thick mucosal layer and easy reach by the FAPT4 transporter. In addition, transport of LCFA-peptide conjugates as molecules, but not as nanoparticles, across the intestinal epithelium greatly increases efficiency and avoids degradation by enzymes in the lysosomal pathway. Furthermore, absorption of LCFA-peptide conjugates in molecules can also eliminate toxicity concerns associated with materials that form the building blocks of a nanoparticle (26, 27).

Exendin-4 (Ex4) is a 39–amino acid peptide that was originally identified in the saliva of Gila monster. It functions as an agonist of the glucagon-like peptide 1 (GLP-1) receptor, which is one of the most popular targets for peptide drugs on type 2 diabetes (28). Unlike human GLP-1 that has a very short circulating half-life (t1/2 = 2 min), Ex4 is more resistant to peptidase and has a circulating half-life of 2.4 hours (29). However, oral administration of Ex4 remains a challenging task due to its high polarity and large molecular weight. In this study, we applied Ex4 as a model peptide drug to test the hypothesis that molecular targeting of FATP4 transporter is a feasible approach for oral delivery of therapeutic peptide drugs.


Site-specific synthesis of LCFA-Ex4 conjugates

Site-specific conjugation of LCFA to Ex4 was carried out in two steps (Fig. 1A and fig. S1). Briefly, diethylpyrocarbonate was conjugated to the histidine residue in Ex4 to form an ethyl ester as an intermediate product (30). The histidine residue was chosen as it was not involved in GLP-1 receptor binding. Since it is a unique amino acid at the C terminus in Ex4, its modification was not expected to compromise Ex4 activity (31). In the second step of reaction, LCFA was conjugated to the intermediate product through transesterification catalyzed by tetranuclear zinc cluster (32). LCFA-Ex4 conjugates were characterized with matrix-assisted laser desorption/ionization (MALDI) and reverse-phase high-performance liquid chromatography (HPLC; figs. S2 to S4). Its stability was also assessed after incubation in murine plasma. LCFA-Ex4 showed a twofold increase in half-life compared with free Ex4 (fig. S5), a result attributed to LCFA protection of peptides from enzymatic degradation by noncovalent binding to albumin. This phenomenon is also confirmed by other peptides modified with fatty acid for injection (28). To evaluate the biological activity of LCFA-Ex4, a glucose tolerance test was performed in C57BL/6 mice. LCFA-Ex4 exhibited comparable glucose-lowering activity as the free Ex4 after subcutaneous injection (Fig. 1B), and both LCFA-Ex4 and the free Ex4 demonstrated 59% decrease in blood glucose compared with the phosphate-buffered saline (PBS)–treated group in 2 hours (Fig. 1C).

Fig. 1 FATP4 transporter mediates uptake of LCFA-Ex4 in epithelial cells.

(A) Schematic view of chemical structure of LCFA-Ex4 (upper panel) and FATP4-mediated transport of LCFA-Ex4 across intestinal epithelial membrane (bottom panel). (B) Glucose tolerance test in C57BL/6 mice to evaluate bioequivalence between free Ex4 and LCFA-Ex4 after subcutaneous injection (n = 6). (C) AUC (area under the curve) of blood glucose levels in glucose tolerance test. (D) Flow cytometry analysis on FATP4 expression in Caco-2 and HCC827 cells. (E) Uptake of Alexa Fluor 594–labeled LCFA-Ex4 (Alexa 594–LCFA-Ex4) in the FATP4-expressing Caco-2 cells and non–FATP4-expressing HCC827 cells. (F) Microscopic analysis on uptake of Alexa 594–LCFA-Ex4 conjugates [short-chain fatty acid (SCFA), C6; medium-chain fatty acid (MCFA), C10; long-chain fatty acid (LCFA), C16] in Caco-2 cells. (G) Competitive binding to the FATP4 transporter by Alexa 594–LCFA-Ex4 in the presence of increasing concentrations (500, 5000, or 10,000 times) of palmitic acid (C16). (H) Uptake of LCFA-Ex4 in FATP4 knockdown Caco-2 cells. FATP4-specific small interfering RNA (siRNA) oligos were used to knock down FATP4 expression, and a scrambled siRNA served as a control. Inset: Western blot analysis on FATP4 expression. Black bars: normalized FATP4 expression levels; white bars: normalized fluorescence intensity (n = 3). (I) Microscopic analysis on intracellular trafficking of Alexa 594–LCFA-Ex4. Nucleus is in blue, the ER track is in green, and Alexa 594–LCFA-Ex4 is in red. Data are presented as means ± SEM. *P < 0.05, **P < 0.01.

In vitro analysis of FATP4 transporter–mediated LCFA-Ex4 uptake in epithelial cells

Flow cytometry analysis was performed to determine FATP4 protein levels in immortalized human cell lines. High FATP4 expression was detected in Caco-2 human colorectal cancer cells, but not in HCC827 human lung cancer cells (Fig. 1D). Correlatively, LCFA-Ex4 conjugates were effectively taken up by the FATP4-postive Caco-2 cells, but not by the FATP4-negative HCC827 cell (Fig. 1E). To test specificity of FATP4 on transport of LCFAs, we conjugated Ex4 with short-chain fatty acid (SCFA; C6), medium-chain fatty acid (MCFA; C10), or LCFA (C16). Confocal microscopic analysis revealed that only LCFA-Ex4 could be efficiently taken up by Caco-2 cells (Fig. 1F).

To confirm dependence of active LCFA-Ex4 transport on the FATP4 transporter, we pretreated Caco-2 cells with phloretin, an inhibitor of the transporter (33). Phloretin effectively blocked cellular uptake of LCFA-Ex4 even after LCFA-Ex4 concentration was increased by eightfold (fig. S6). In addition, in vitro assay showed that uptake of LCFA-Ex4 was competitively inhibited in a dose-dependent manner by palmitic acid (Fig. 1G). Furthermore, knockdown of FATP4 in Caco-2 cells with FATP4-specific small interfering RNA (siRNA) resulted in a dose-dependent reduction in protein expression with concomitant inhibition of LCFA-Ex4 absorption (Fig. 1H). Together, these experiments established a direct correlation between FATP4 expression and LCFA-Ex4 uptake in Caco-2 cells.

To investigate route of intracellular transport of LCFA-Ex4 after FATP4 transporter–mediated uptake, we stained subcellular organelles to track intracellular location of fluorescently labeled LCFA-Ex4. One hour after Caco-2 cells were treated with LCFA-Ex4, colocalization between LCFA-Ex4 and early endosome, late endosome, or lysosome was assessed using the analysis of Mander’s colocalization coefficients. The data showed that a high level of LCFA-Ex4 was colocalized with the ER. In comparison, minimal levels of LCFA-Ex4 were colocalized with the early endosomes, late endosomes, and lysosomes (Fig. 1I and fig. S7). The results indicate that LCFA-Ex4 was not guided to the lysosomal pathways after transmembrane uptake. This process provides a great benefit to peptide drugs by avoiding lysosomal degradation, thus keeping the activity intact during the transcellular transport process.

Preparation and characterization of OraEx4

Well-dispersed ChiNPs were prepared by following an ionic crosslinking method (Fig. 2A) (34). The positively charged ChiNPs (~20 nm, +35 mV) were electrostatically adsorbed onto the negatively charged DOPA (1,2-dioleoyl-sn-glycero-3-phosphate) liposome (~200 nm, −17 mV) (Fig. 2B) by simply mixing the two together to form ChiNP-coated liposomes (ChiNP-Lipo). Transmission electron microscopy (TEM) imaging showed that the liposomes had well-defined structures, and ChiNPs evenly dispersed on the surface of liposomes. Coating density of ChiNPs on the liposome could be adjusted by changing the mass ratio between the two components (Fig. 2C). Fourier-transform infrared (FT-IR) spectrum demonstrated that the characteristic peak of ─NH3+ in chitosan at 1630 cm−1 was shifted to 1628 cm−1 in ChiNP-Lipo at pH 6.0 (Fig. 2D), which could be attributed to electrostatic interaction between the two components. This characteristic peak disappeared at pH 7.0, indicating detachment of ChiNP from the liposome as a result of deprotonation of chitosan (Fig. 2D). To evaluate pH-responsiveness of ChiNP-Lipo, we monitored changes in hydrodynamic size under different pH conditions. ChiNP-Lipos with 2:10 and 4:10 mixing ratios displayed sharp pH-responsive changes in size, with a threefold size increase when pH was changed from 6 to 7 (Fig. 2E). Liposomal swelling was accompanied with a change in surface charge from positive to near neutral (fig. S8), leading to detachment of ChiNP and disruption of the liposome.

Fig. 2 Characterization of OraEx4 on zeta potential, size, and in vitro release.

(A) Schematic view of OraEx4 synthesis, including loading of LCFA-Ex4 into DOPA liposome (Lipo) and sequential coating with chitosan nanoparticle (ChiNP). Chito, chitosan; TPP, sodium triphosphate; DOPA, 1,2-dioleoyl-sn-glycero-3-phosphate; Chole, cholesterol; LCFA,16-hydroxyhexadecanoic acid; Ex4, exendin-4. (B) Zeta potential and size of ChiNP, Lipo, and ChiNP-Lipo. (C) Transmission electron microscopy (TEM) images of ChiNP-Lipo with increasing ChiNP/Lipo ratio. (D) Fourier-transform infrared (FT-IR) spectra of ChiNP, Lipo, and ChiNP-Lipo under different pH environments. (E) Changes in hydrodynamic radius of ChiNP-Lipo under different pH conditions. (F) Release profiles of LCFA-Ex4 from OraEx4 and conventional liposome (T-Lipo). (G) TEM images of OraEx4 at different pH. Data are presented as means ± SEM.

LCFA-Ex4–loaded and ChiNP-stabilized liposomes (OraEx4) were prepared by encapsulating LCFA-Ex4 into liposomes followed by ChiNP decoration (Fig. 2A). ChiNP/Lipo ratio of 2:10 was chosen for OraEx4 preparation to achieve an optimal trade-off between drug loading capacity (table S1) and swelling property of ChiNP-Lipo (Fig. 2E). It was expected that the cargo would be released in the pH-neutral small intestinal environment (fig. S9). In vitro release study indicated that 90% LCFA-Ex4 was released from OraEx4 at pH 7, which mimicked the small intestinal pH condition (Fig. 2F). In comparison, release of LCFA-Ex4 molecules packaged in conventional liposomes (without ChiNP coating) did not respond to pH change well, indicating that half of the payload would have been released before reaching the intestine and, thus, would not be absorbed in the small intestine. Swollen deformed liposomes were observed at pH 7.0 and pH 7.4 (Fig. 2G), and detached ChiNPs were visible next to the deformed liposomes (fig. S10). LCFA-Ex4 released from OraEx4 at pH 7.0 was collected and used to test resistance to intestinal enzymes in vitro. LCFA-Ex4 presented much higher resistance to degradation by intestinal enzymes than free Ex4 (fig. S11), which is a significant advantage for enhanced oral stability. The mechanism of LCFA conjugation–increased resistance of Ex4 to intestinal enzymes can be contributed to the coupling of hydrophobic fatty acid moiety to the hydrophobic binding site in the vicinity of the active center of the enzyme, which greatly decreased the catalytic activity of this enzyme (35). Thus, LCFA conjugation increased the stability of Ex4 under intestinal enzyme treatment.

Biodistribution and lymphatic transportation of OraEx4

To elucidate biodistribution of LCFA-Ex4 after oral administration, Alexa Fluor 594–labeled LCFA-Ex4 (Alexa 594–LCFA-Ex4) was used to prepare OraEx4, and the nanoformulation was applied to treat C57BL/6 mice orally. Mice were euthanized at different time points in the next 12 hours, and major organs were collected to measure fluorescence intensity ex vivo. There was almost no signal in the stomach at all the time points, indicating that OraEx4 remained stable, and there was no leak of LCFA-Ex4 in the gastric tract (Fig. 3B and fig. S12). In contrast, high levels of uptake of LCFA-Ex4 were observed in the small intestine, beginning as early as 1 hour after administration and reaching peak levels at the 2- and 4-hour time points (Fig. 3B). Fluorescence intensity within the small intestine correlated with FATP4 expression, displaying high levels of FATP4 expression and fluorescence intensity in the jejunum and ileum (Fig. 3A). In addition, LCFA-Ex4 was also detected in the liver, lung, and mesenteric lymph nodes (MLNs), indicating that LCFA-Ex4 entered systemic circulation after oral administration. High levels in the kidney indicated renal excretion of the peptide. Immunofluorescence staining revealed overexpression of FATP4 in the villi of the jejunum and ileum and colocalization of LCFA-Ex4 and FATP4 on epithelial cells within these sections (Fig. 3C). Oral absorption only applied to OraEx4, as oral administration with free LCFA-Ex4 did not result in sufficient drug retention in the small intestine or any other organs (fig. S13). Together, these results suggest that LCFA-Ex4 was locally released in the intestinal tract and successfully transported into enterocytes by the FATP4 transporter before they entered the blood circulation.

Fig. 3 Biodistribution and lymphatic transport of OraEx4.

(A) Western blot analysis of FATP4 expression in jejunum, ileum, duodenum, and colon and correlation with uptake of LCFA-Ex4 (n = 3). (B) Quantification of fluorescence in major organs at different time points after oral administration of Alexa Fluor 594–labeled OraEx4. All data are normalized to tissue weight (n = 3). (C) LCSM images of Alexa 594–LCFA-Ex4 in different intestinal sections after oral treatment. Representative images are shown. (D) IVIS (In vivo imaging system) images of LCFA-Ex4 in mesenteric lymph node (MLN) from mice treated with OraEx4. (E) Quantitative analysis on fluorescence intensity of LCFA-Ex4 in MLN from (D). (F) Intravital microscopic images of MLN from mice after treatment with OraEx4. (G) Hematoxylin and eosin (H&E) staining of MLN from mice treated with nanogold-conjugated OraEx4 with or without phloretin. Nanogold was stained with silver enhancer. Data are presented as means ± SEM, **P < 0.01.

Detection of LCFA-Ex4 in MLNs 2 hours after oral administration and its persistence over 8 hours (fig. S12) suggested that LCFA-Ex4 entered blood circulation through lymphatic transport after absorption in the small intestine. Simultaneous gavage of LCFA-Ex4 with phloretin substantially reduced the accumulation of LCFA-Ex4 in MLNs (Fig. 3, D and E). Intravital microscopy detected LCFA-Ex4 accumulation in MLNs of live mice 2 hours after oral gavage (Fig. 3F). To further confirm LCFA-Ex4 accumulation in MLNs, C57BL/6 mice were orally administered with OraEx4 prepared with nanogold-labeled LCFA-Ex4. Histological analysis revealed nanogold particles in MLNs from mice treated with OraEx4, but not in mice treated with OraEx4 in combination with phloretin (Fig. 3G). These results support the notion that LCFA-Ex4 enters blood circulation through lymphatic transportation after FATP4-mediated uptake from the small intestine.

Blood glucose regulation of OraEx4

To evaluate efficacy from OraEx4, intraperitoneal glucose tolerance test (IPGTT) was performed in fasted C57BL/6 mice treated with free Ex4 (4 mg/kg), free LCFA-Ex4 (4 mg/kg), OraEx4 (4 mg/kg) alone, or in combination with phloretin (50 mg/kg). OraEx4 raised plasma insulin level by fourfold within 30 min (fig. S14) and significantly inhibited surge in blood glucose level (BGL; ~2-fold decrease) (Fig. 4A). Specifically, area under the curve (AUC) of blood glucose was reduced by 41.6% in the OraEx4 treatment group compared with the PBS control group. OraEx4 activity was completely blocked by phloretin (Fig. 4A). Neither free Ex4 nor free LCFA-Ex4 had any significant impact on insulin secretion or BGL (Fig. 4A and figs. S14 and S15). Efficacy from OraEx4 was further evaluated in diabetic db/db mice, a murine model that recapitulates human type 2 diabetes mellitus (36). BGL was monitored over 12 hours after a single treatment of OraEx4 at LCFA-Ex4 (4 or 8 mg/kg). Dose-dependent reduction in BGL was observed. BGL control lasted for 4 hours with the low-dose group and for 12 hours with the high-dose group (Fig. 4C). Correlatively, the high-dose treatment maintained a plasma LCFA-Ex4 concentration for over 12 hours above the 50 pg/ml threshold (Fig. 4D), a level that was proven to be the minimum effective concentration for Ex4 (37). Relative bioavailability of OraEx4 was 24.8% compared with subcutaneously injected LCFA-Ex4 (fig. S16). Long-term glucose regulation in db/db mice was evaluated by daily gavage of OraEx4 (8 mg/kg) for 12 days. BGL in OraEx4-treated mice was maintained below 300 mg/dl throughout the treatment period, while BGL in the PBS control group remained between 400 and 450 mg/dl (Fig. 4E). Inclusion of phloretin blocked OraEx4 activity. No significant difference on body weight change was observed in the treatment groups (Fig. 4F). After 12 days of OraEx4 treatment, the db/db mice showed greatly improved glucose tolerance with concomitant threefold increase in plasma insulin level compared with mice in the control group during IPGTT (Fig. 4, G and H). OraEx4 efficacy in db/db mice was also inhibited by phloretin (Fig. 4, G and H and fig. S17). We compared the size of pancreatic islets and insulin secretion in the islets in wild-type C57BL/6 mice with posttreatment db/db mice. Immunofluorescence staining showed substantially increased insulin content in pancreatic islets in the OraEx4 treatment group (Fig. 4I). Hematoxylin and eosin (H&E) staining showed that islets in control db/db mice treated with PBS were substantially larger than those in healthy C57BL/6 mice (Fig. 4J). This is consistent with the adaptive islet hyperplasia of insulin resistance in type 2 diabetes mellitus (38). However, immunofluorescence staining demonstrated that the islet of control db/db mice exhibited noticeably lower staining intensity for insulin than those in healthy C57BL/6 mice (Fig. 4I), which is most likely caused by islet β cell degranulation in type 2 diabetes mellitus (39). It has been previously shown that, as a GLP-1 receptor agonist, Ex4 is able to promote proliferation of pancreatic islets and enhance insulin production (40). After treatment for 12 days, increased islet size and restored insulin production were observed in OraEx4-treated mice (Fig. 4, I and J), demonstrating the effectiveness of oral delivery of Ex4. Such effects from OraEx4 were markedly abrogated when the FATP4 transporter inhibitor phloretin was included in the treatment. However, it should also be noted that OraEx4 treatment did not have a big impact on other pathological features in db/db mice such as hepatic steatosis and edema in renal tubular epithelia (fig. S18). These results provide solid support for oral administration of peptide drugs as an effective intervention for management of type 2 diabetes.

Fig. 4 In vivo efficacy evaluation of OraEx4.

(A) IPGTT after C57BL/6 mice were treated with oral administration of Ex4 solution (4 mg/kg), OraEx4 (4 mg/kg), or OraEx4 (4 mg/kg) together with phloretin (50 mg/kg; n = 6). (B) AUC of blood glucose levels in (A). (C) Blood glucose levels after a single treatment of OraEx4 (4 or 8 mg/kg) in db/db mice (n = 5). (D) Pharmacokinetic profiles of Oral-Ex4 (4 and 8 mg/kg), compared with subcutaneous (SC) injection of LCFA-Ex4 (n = 5). (E) Long-term blood glucose regulation by daily gavage of OraEx4 (8 mg/kg; n = 5). (F) Body weight changes in db/db mice during treatment with OraEx4 for 12 days (n = 5). (G) IPGTT after treatment with a single dose of OraEx4 (8 mg/kg) in db/db mice (n = 5). (H) Insulin levels before and after OraEx4 treatment in db/db mice (n = 5). (I) Immunofluorescence images on insulin secretion from islets in C57BL/6 and db/db mice. (J) H&E staining of pancreatic islets from posttreatment db/db mice. Images show representative islets from five mice per group. Data are presented as means ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001.

Desirable safety profile from OraEx4

Last, the safety profile from OraEx4 was assessed in wild-type C57BL/6 mice by oral gavage of OraEx4 (40 mg/kg), which was five times of the therapeutic dosage. Blood samples were collected 24 hours later, and biochemical parameters were measured to assess its safety profile on functions of major organs. Activities of aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALKP) were in the normal range, indicating no liver damage. Parameters that are reflective of renal function including blood urea nitrogen (BUN) and creatinine were also in the normal range. Other parameters including amylase (Amyl) and lactate dehydrogenase (LDH) were within the range as well (Fig. 5, A to C). Hematological analysis was also carried out to assess any abnormality in mice treated with OraEx4, and no significant difference in hematological values was identified between the OraEx4 and PBS treatment groups (Fig. 5, D and E). In addition, histopathological analysis of major organs revealed no apparent morphological changes after treatment with the high-dose OraEx4 (Fig. 5F). Together, the results indicate a satisfactory safety profile from OraEx4.

Fig. 5 Analysis on safety profiles after C57BL/6 mice were treated with OraEx4.

Serum samples were collected from mice (n = 5) after daily treatment with indicated agents for 10 days, and levels of enzymes and other biomarkers indicating (A) hepatic function, (B) renal function, and (C) other organs were analyzed. Changes in (D) white blood cells and (E) red blood cell and platelets were also measured. (F) H&E staining of tissue blocks from major organs. Images are representative of five mice per group. Data are presented as means ± SEM.


In this work, we presented an effective approach to enhance oral absorption of peptide drugs by molecular targeting of the FATP4 transporter. FATP is an intestinal transporter with the highest expression level and transport efficiency and is responsible for active transport of LCFA (>C12) across the intestinal epithelium (22, 23). Other fatty acids including SCFAs (<C7) and MCFAs (C7 to C12) are absorbed through passive diffusion, thanks to their low molecular weight (41). After conjugation of SCFA and MCFA to Ex4, the big molecular weight of Ex4 made conjugates difficult to be absorbed through passive diffusion in Caco-2 cells (Fig. 1F). LCFA-Ex4 was trafficked to the ER after FATP4-mediated transport across the intestinal epithelium, avoiding entrapment of the peptide drug in lysosomes (Fig. 1I). The result suggests that the transmembrane absorption process of LCFA-Ex4 was through the pores of the FATP4 transporters rather than through receptor-mediated endocytosis and, thus, prevented damage of the peptide drug by lysosomal enzymes.

Liposomes have obvious advantages over other nanocarriers, standard drug tablets, or capsule formulations, since they are able to enhance drug absorption in the intestinal tract owing to their excellent entrapment capacity, biocompatibility, and safety (42). However, their application in oral delivery has been hampered by their low stability in the gastric tract (43). In OraEx4, liposomes are stabilized by binding of protonated ChiNP at gastric pH. ChiNP is deprotonated and detaches from the liposome when pH turns neutral in the intestinal lumen, resulting in swelling of the liposome and release of the drug cargo. LCFA-Ex4 had a similar transport pathway as free LCFA, exiting the basolateral membrane of enterocytes to enter the lymphatic system transport (Fig. 3, D to F). Lymphatic absorption of peptide drug after subcutaneous injection has been demonstrated as a safe route of drug transport (44). While biodistribution analysis, pharmacokinetic and efficacy studies, and in vitro release studies all suggested effective release of LCFA-Ex4 in the intestine, it should be noted that current data do not fully resolve the release process in vivo, which can be further investigated in future work through separate labeling of the liposome and the drug cargo. Last, OraEx4 showed excellent control of BGLs in murine type 2 diabetes model, a high level of oral bioavailability compared with subcutaneous injection, and no acute toxicity even at an extremely high dose.

In conclusion, this work demonstrates that molecular targeting of the FATP4 transporter can safely and efficiently enhance oral delivery of therapeutic peptides. To the best of our knowledge, this is the first study to report a strategy of taking advantage of the FATP4 transporter to enhance oral absorption of peptide drugs directly in the molecular form. This technology is not limited to oral delivery of antidiabetic peptide. It can be extended to enhance oral delivery of a variety of peptide drugs for many different types of diseases.



Ex4 was purchased from GL Biochem Ltd. (Shanghai, China). RPMI 1640, fetal bovine serum, rabbit anti-murine and human FATP4 polyclonal antibody (PA5–42446), Alexa Fluor 594–NHS Ester, LysoTracker Green DND-26, CellLight Early Endosomes–GFP, CellLight Late Endosomes–GFP, and ER-Tracker Green were obtained from Thermo Fisher (Waltham, MA, USA). Rabbit anti-murine insulin antibody was from Abcam (ab63820, Cambridge, MA, USA). Chitosan (molecular weight, 50,000 to 190,000 Da), pentasodium triphosphate, zinc trifluoroacetate hydrate, 16-hydroxyhexadecanoic acid, silver enhancer kit, and diisopropyl ether were ordered from Sigma-Aldrich (St. Louis, MO, USA). Mono-Sulfo-NHS-Nanogold (1.4 nm) was purchased from Nanoprobe, Inc. (Yaphank, NY, USA), and 1,2-dioleoyl-sn-glycero-3-phosphate (sodium salt) and cholesterol were purchased from Avanti Polar Lipids Inc. (Alabaster, AL, USA).

Site-specific conjugation of peptide with fatty acid

Site-specific conjugation of fatty acid to Ex4 peptide was carried out in a two-step process. Briefly, diethylpyrocarbonate reacted with the histidine amino acid of Ex4 peptide to form an intermediate ethyl ester product. Subsequently, the ethyl ester was replaced by fatty acid through transesterification catalyzed by tetranuclear zinc cluster. Detailed procedure is described below. Diethylpyrocarbonate (0.25 mM) in ethanol was added dropwise into an equal volume of peptide (4 mg/ml) in potassium phosphate buffer (pH 6), and the mixture was incubated for 1 hour in a 4°C ice bath. Excess diethylpyrocarbonate was hydrolyzed to carbon dioxide and ethyl alcohol (30). The intermediate ethyl ester was then lyophilized. For transesterification, 0.038 mmol tetranuclear zinc cluster, 3 mmol intermediate ethyl ester, 3.6 mmol 16-hydroxyhexadecanoic acid (LCFA), and 5 ml of diisopropyl ether were refluxed for 30 min under an argon atmosphere. The resulting mixture was concentrated and purified with silica gel column chromatography. Conjugation of SCFA (C6) and MCFA (C12) was performed following the same procedure. The fatty acid conjugates were characterized with MALDI and reverse-phase HPLC.

Synthesis of fluorescently labeled LCFA-Ex4

Alexa Fluor 594–NHS ester (5 mg/ml) in dimethyl sulfoxide was slowly added into an equal volume LCFA-Ex4 (10 mg/ml) in 0.1 M sodium bicarbonate buffer under vortex. The reaction was incubated for 1 hour at room temperature with continuous stirring. Half volume of freshly prepared 1.5 M hydroxylamine (pH 8.5) was added, and the mixture was incubated for 1 hour at room temperature to stop the reaction. The product was purified with Sephadex G-50 gel column.

Preparation of ChiNPs

ChiNPs were prepared following an ion–cross-link method with slight modification (34). Briefly, chitosan (molecular weight, 50,000 to 190,000 Da) was dissolved in 0.1 M HCl to a final concentration of 0.02 mg/ml, and the mixture was stirred for 24 hours. pH of the chitosan solution was adjusted to 4.5 with 0.01 M NaOH. Pentasodium triphosphate (TPP; 0.01 mg/ml) in deionized water was added dropwise to the chitosan solution to a 3:1 (chitosan/TPP) final volume ratio under vigorous vortex at 20°C. ChiNPs were collected by centrifugation at 10,000 rpm for 10 min. The nanoparticles were washed three times with deionized water and reconstituted with deionized water to a final concentration of 0.015 mg/ml.

Preparation of ChiNP-decorated liposome

Liposomes were prepared following a thin film method with minor modifications (45). Briefly, 20 mg of lipid mixture consisting of cholesterol and 1,2-dioleoyl-sn-glycero-3-phosphate in a 1:1 molar ratio was dissolved in 1 ml of chloroform. Chloroform was removed by rotary evaporation to form a thin lipid film. To prepare empty liposomes, the lipid film was rehydrated with 2 ml of deionized water under vortex, followed by sonication for 5 min in a bath sonicator. The solution then was extruded through 400- and 100-nm filters 10 times to form unilamellar vesicles with a narrow range of size distribution. To prepare LCFA-Ex4–loaded liposome, the lipid film was rehydrated with 2 ml of deionized water containing 300 μg of peptide and then mixed at 50°C in a water bath to bring a homogenous solution. The solution went through repetitive freeze-thaw cycles by following this order: freeze solution in dry ice, thaw at room temperature, incubate at 40°C in a water bath for 5 min, and then vortex for 30 s. The entire process was repeated for five times. Last, peptide-loaded liposomes were ultracentrifuged and washed with deionized water. Liposome pellet was resuspended to a desirable concentration with deionized water. To prepare ChiNP-decorated liposomes, pH in both ChiNP and liposome suspensions was adjusted to 6, and the suspensions were mixed together at the desired mass ratio, followed by vortex for 2 hours. The nanoparticle-decorated liposomes were then centrifuged at 30,000 rpm for 30 min.

Flow cytometric analysis

In brief, 100 nM Alexa 594–LCFA-Ex4 was incubated with Caco-2 cells and HCC827 cells for 1 hour to test uptake of Alexa 594–LCFA-Ex4. Alexa 594r–LCFA-Ex4 mixed with different concentrations of unlabeled LCFA were coincubated with Caco-2 cells at 4°C for 1 hour to test competitive binding to FATP4 transporters. Cells (1 × 106) were incubated with an anti-mouse FATP4 antibody followed by a phycoerythrin (PE)–labeled secondary antibody, and then analyzed with flow cytometry (BD Biosciences). FATP4 knockdown Caco-2 cells were seeded into 24-well plates and incubated with 100 nM Alexa 594–LCFA-Ex4, and uptake efficiency was measured with flow cytometry. In addition, Caco-2 cells were pretreated with phloretin at different concentrations followed by incubation with Alexa 594–LCFA-Ex4. Flow cytometry was applied to determine uptake efficacy.

Confocal laser scanning microscopy

Caco-2 cells were seeded in an eight-well chamber slide and incubated with 100 nM Alexa 594–LCFA-Ex4 for 1 hour. LysoTracker Green DND-26, CellLight Early Endosomes–GFP, CellLight Late Endosomes–GFP, or ER-Tracker Green was added following manufacturer-suggested protocols. Cells were fixed with 4% paraformaldehyde and then stained with 4′,6-diamidino-2-phenylindole (DAPI). Slides were mounted, and confocal images were taken using a FluoView 3000 confocal microscope from Olympus.

Measurement of encapsulation efficiency

Peptide-loaded ChiNP-Lipo (150 μl) was centrifuged at 10,000g for 10 min, and supernatant was removed. One-hundred fifty microliters of 40% ethanol containing 0.1 M NaOH was added to break up the ChiNP-Lipo under vigorous stirring in a 50°C water bath for 30 min. The resulting mixture was centrifuged at 10,000g for 10 min. A 50-μl aliquot of the supernatant was applied to measure the amount of peptide in ChiNP-Lipo with HPLC. Chromatography was performed with an Agilent C18 column that was maintained at 30°C. Gradient elution was applied with 0.2% acetic acid in PBS as solvent A and acetonitrile as solvent B. Flow rate was adjusted to 0.8 ml/min without a split. Total amount of peptide and encapsulation efficiency in ChiNP-Lipo were calculated on the basis of the HPLC peak areas.

Evaluation of pH responsiveness and in vitro drug release

Changes in size, zeta potential, and morphology were used as parameters for measuring pH responsiveness of ChiNP-Lipo. To measure zeta potential and hydrodynamic size, samples were dispersed homogeneously in deionized water at 1 mg/ml concentration with ultrasonication, and zetasizer was applied for measurement. To study pH-dependent drug release in vitro, OraEx4 was resuspended in different pH conditions that mimic the gastric fluid (pH 1.2) and intestinal fluids (pH 6.0, 7.0, and 7.4) for 2 hours. Samples were collected at scheduled time intervals and centrifuged at 10,000g for 10 min. The supernatants were then analyzed with HPLC as described above. All experiments were performed in triplicate, and the mean value of three measurements is shown.

Sample preparation for TEM

A negative staining procedure was used to examine ChiNPs on the surface of liposomes and swelling of ChiNP-Lipo with TEM. Briefly, ChiNP-Lipo (0.2 mg/ml) samples were prepared, and one drop of the sample was placed on a carbon-coated copper grid. After air drying for 10 min on the grid, particles were stained with 2% phosphotungstic acid for 15 min. The samples were then imaged with a Hitachi 3H-7000FA transmission electron microscope.

Investigation on stability of LCFA-Ex4 in intestinal enzymes and plasma

Trypsin, chymotrypsin, and aminopeptidase N were dissolved in 2 ml of prewarmed (37°C) simulated intestinal fluid (PBS, pH 7.0) at a final concentration of 0.2 U/ml. Murine plasma (2 ml) was prepared and prewarmed. LCFA-Ex4 and Ex4 were added into the enzyme solution or murine plasma at a final concentration of 100 μg/ml. At different time points, a 100-μl aliquot was taken from the solution, and an equal volume of ice-cold 1 M perchloric acid was added to stop enzyme reaction. LCFA-Ex4 and Ex4 concentrations in the samples were determined with HPLC following the same procedure as described above. Each experiment was repeated three times.

Biodistribution study

All animal studies were carried out in accordance with guidelines determined by the Animal Welfare Act and the Guide for the Care and Use of Laboratory Animals and complied with protocols approved by the Institutional Animal Care and Use Committee at the Houston Methodist Research Institute. OraEx4 or Alexa 594–LCFA-Ex4 (1 mg/kg) was administered to C57BL/6 mice (male, 5 to 6 weeks old) alone or in combination with phloretin (50 mg/kg) via oral gavage. Mice were euthanized, and major organs were harvested at 1, 2, 4, 8, and 12 hours after oral gavage for fluorescence imaging using an IVIS-200 in vivo imaging system (PerkinElmer) (46). Quantitative analysis of fluorescence in organs was performed using Living Image Software 4.2 (Caliper Life Sciences, Hopkinton USA). Live imaging of OraEx4 accumulation in MLN was performed on a Nikon A1R laser scanning confocal microscope adapted for intravital microscopy. Duodenum, jejunum, ileum, and colon from treated mice were harvested after 2 hours. Tissues were washed with filtered saline five times. Each intestinal section (2 cm in length) was separated and homogenized in a lysis buffer for protein extraction. Extracted protein from each intestinal section was used to measure FATP4 expression with Western blot. The remaining portions of each intestinal section were embedded in optimal cutting temperature (OCT) compound and cryo-sectioned for confocal microscopy. Immunofluorescence staining was performed using anti-FATP4 antibody and fluorescein isothiocyanate–labeled secondary antibodies. DAPI was used to stain nuclei of the cells. Uptake of LCFA-Ex4 was observed under a confocal microscope.

Nanogold label for tracking accumulation in MLNs

Mono-Sulfo-NHS-Nanogold was conjugated to LCFA-Ex4 to facilitate microscopic analysis. Briefly, LCFA-Ex4 was dissolved in 0.02 M PBS with 0.15 M sodium chloride at pH 7.4, and Mono-Sulfo-NHS-Nanogold was added. Peptide concentration was adjusted to 6 mM to achieve optimal labeling. The solution was incubated for 1 hour at 20°C. Unbound gold nanoparticles were separated from the peptide conjugates with Superdex-75 gel filtration chromatography, and the gold-peptide conjugates were concentrated by centrifugation using Centricon-30 columns (Millipore). Nanogold-labeled LCFA-Ex4 was encapsulated into ChiNP-Lipo following the same procedure as described above. After oral gavage with 2 mg of nanogold (Au)–labeled LCFA-Ex4/ChiNP-Lipo (Au-LCFA-Ex4) alone or together with phloretin (50 mg/kg) for 2 hours, MLNs were isolated for H&E staining and silver enhancer staining. To perform silver enhancer staining, solution A and solution B in the silver enhancer kit (Sigma-Aldrich-SE100) were mixed 1:1 immediately before use and applied to the slide with tissue section. After 5 to 10 min, the slide was rinsed with distilled water, fixed for 2 to 3 min in a sodium thiosulfate solution, and rinsed again. The tissue section was mounted before microscopic analysis.

In vivo efficacy evaluation

IPGTT was performed in wild-type C57BL/6 mice. Briefly, 5- to 6-week-old male C57BL/6 mice were fasted overnight and then received subcutaneous administration of PBS, LCFA-Ex4 (300 μg/kg), or free Ex4 (300 μg/kg), or oral administration of Ex4 solution (4 mg/kg), LCFA-Ex4 (4 mg/kg), or OraEx4 (4 mg/kg LCFA-Ex4) alone or together with phloretin (50 mg/kg). One hour after drug administration, all mice were intraperitoneally injected with glucose (2 g/kg). Blood samples were collected to monitor BGL. Blood samples were also collected and applied to measure insulin levels with a mouse insulin enzyme-linked immunosorbent assay (ELISA) kit (Phoenix Biotech, San Antonio, USA). AUC of blood glucose was calculated to evaluate hypoglycemic response in each treatment group.

To determine glucose regulation after a single dose treatment, 6- to 7-week-old male db/db mice were randomly divided into three groups (five mice per group), and each mouse received OraEx4 (4 mg/kg or 8 mg/kg; LCFA-Ex4-equivalent) as a single dose oral treatment. BGL was monitored and plasma LCFA-Ex4 concentrations were measured with an exnatide-4 mouse ELISA kit (Phoenix Biotech, USA). For long-term efficacy evaluation, db/db mice were treated with daily gavage of OraEx4 (8 mg/kg) for 12 days. Separate cohorts of db/db mice were treated with PBS control or with OraEx4 plus phloretin. BGL was monitored 1 hour after gavage. IPGTT in mice after OraEX4 (8 mg/kg) treatment was evaluated as described above (five mice per group). Pancreas and other major organs from the treatment groups were collected. One part of each organ was fixed with 4% paraformaldehyde, paraffin embedded, and sectioned for H&E staining, and the other part was frozen in OCT compound and sectioned for immunofluorescence staining to monitor insulin and nuclei.

Safety evaluation

C57BL/6 mice (6- to 8-week-old, female) were randomly assigned into groups and maintained in a pathogen-free facility under a day-night cycle. Mice were orally treated with OraEx4 at a low dose of 8 mg/kg or a high dose of 40 mg/kg, and mice in the control group were treated with an equal volume of vehicle control by gavage. Blood samples were collected 24 hours after the last treatment. The samples were analyzed for total number and percentage of white blood cell, neutrophils, eosinophils, basophils, lymphocytes, monocytes, and red blood cell, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, red cell distribution width, and platelet. Serum samples were analyzed to measure hepatic enzymes including AST, ALT, and ALKP, kidney function biomarkers including BUN and creatinine, and other biomarkers including amylase and LDH.

Statistical analysis

Two-tailed Student’s t test was used to examine statistically significant difference for two-group analysis. One-way analysis of variance (ANOVA) with Tukey’s post hoc multiple comparison test was used for multigroup comparison analysis. All data are presented as means ± SEM. GraphPad Prism 5 software was used to perform all statistical analyses. *P < 0.05, **P < 0.01, ***P < 0.001 as indicated.


Supplementary material for this article is available at

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: Funding: This work was partially supported by NIH grants R01CA193880, R01CA222959, and U54CA210181, U.S. Department of Defense grant W81XWH-17-1-0389, and an internal grant of Houston Methodist Research Institute. Author contributions: Z.H. performed the materials synthesis and characterization, in vitro and in vivo experiments, data analysis, statistical analysis, and manuscript preparation. H.S. designed the experiments, analyzed the data, and wrote the manuscript. S.N. and S.G. analyzed the data and critically reviewed and edited the manuscript. X.W. performed animal experiments. C.L. analyzed the FT-IR data. L.E.H. edited the manuscript. M.F reviewed the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article