• Open Access

A fusion protein derived from plants holds promising potential as a new oral therapy for type 2 diabetes

Authors


Summary

The incretin hormone glucagon-like peptide-1 (GLP-1) is recognized as a promising candidate for the treatment of type 2 diabetes (T2D), with one of its mimetics, exenatide (synthetic exendin-4) having already been licensed for clinical use. We seek to further improve the therapeutic efficacy of exendin-4 (Ex-4) using innovative fusion protein technology. Here, we report the production in plants a fusion protein containing Ex-4 coupled with human transferrin (Ex-4-Tf) and its characterization. We demonstrated that plant-made Ex-4-Tf retained the activity of both proteins. In particular, the fusion protein stimulated insulin release from pancreatic β-cells, promoted β-cell proliferation, stimulated differentiation of pancreatic precursor cells into insulin-producing cells, retained the ability to internalize into human intestinal cells and resisted stomach acid and proteolytic enzymes. Importantly, oral administration of partially purified Ex-4-Tf significantly improved glucose tolerance, whereas commercial Ex-4 administered by the same oral route failed to show any significant improvement in glucose tolerance in mice. Furthermore, intraperitoneal (IP) injection of Ex-4-Tf showed a beneficial effect in mice similar to IP-injected Ex-4. We also showed that plants provide a robust system for the expression of Ex-4-Tf, producing up to 37 μg prEx-4-Tf/g fresh leaf weight in transgenic tobacco and 137 μg prEx-4-Tf/g freshweight in transiently transformed leaves of N. benthamiana. These results indicate that Ex-4-Tf holds substantial promise as a new oral therapy for type 2 diabetes. The production of prEx-4-Tf in plants may offer a convenient and cost-effective method to deliver the antidiabetic medicine in partially processed plant food products.

Introduction

Diabetes is a global epidemic that is expected to affect approximately 350 million people by 2030 (Collins et al., 2011). Type 2 is the most common form of diabetes, accounting for about 90%–95% of all cases of the disease. Type 2 diabetes is a costly disease, associated with substantial morbidity and mortality, particularly from its cardiovascular complications (Potenza et al., 2011; Zhang P et al., 2010). Type 2 diabetes is characterized by insulin resistance, impaired glucose-induced insulin secretion and inappropriately regulated glucagon secretion which in combination eventually leads to the development of hyperglycaemia (Alonso-Magdalena et al.,2011). Although several classes of antidiabetic drugs are available, achieving and maintaining long-term glycaemic control is often challenging, and many current agents have treatment-limiting side effects (Molitch, 2013). Therefore, the development of new drugs to treat diabetes and to prevent associated complications is necessary and urgent.

Incretin-based therapies represent an important and major step forward in the development of new treatments for type 2 diabetes (Koliaki and Doupis, 2011; Schwartz and Defronzo, 2013). Incretins are hormones released from the gut in response to nutrient ingestion that potentiate glucose-stimulated insulin secretion. The predominant incretin hormone is glucagon-like peptide-1 (GLP-1). In addition to stimulating insulin secretion, GLP-1 suppresses glucagon release, slows gastric emptying, improves insulin sensitivity and reduces food intake (Mudaliar and Henry, 2012). However, native GLP-1 has a very short physiological half-life due to degradation by dipeptidyl peptidase-4 (DPP-4). GLP-1 survives <90 s following intravenous infusion and 1 h after subcutaneous injection (Holst and Gromada, 2004). Therefore, to be clinically effective as a diabetes treatment, native GLP-1 would have to be given as a continuous infusion. For this reason, the clinical focus for potential treatments for T2D has shifted towards long-acting GLP-1 receptor agonists and DPP-4 inhibitors.

Exendin-4 (Ex-4), a 39 amino acid peptide isolated from salivary secretions of Gila monster, displays 52% identity to mammalian GLP-1 and is a potent and long-acting agonist of GLP-1 receptor (Lovshin and Drucker, 2009). Ex-4 is resistant to degradation by DPP-IV and consequently has a longer half-life compared with GLP-1 (Giorgino et al., 2007). Ex-4 increased insulin production by β-cells in response to glucose, promoted β-cell regeneration and protected against apoptosis (Hadjiyanni et al., 2008; Lovshin and Drucker, 2009). The synthetic version of Ex-4, exenatide, was approved by the Food and Drug Administration (FDA) in 2005 for the treatment of T2D and was the first marketed GLP-1 receptor agonist (Kyriacou and Ahmed, 2010; Robles and Singh-Franco, 2009). While exenatide has demonstrated its value in treating patients with T2D, its therapeutic utility is limited due to the frequent injections required (twice daily), thus making it an inconvenient and expensive treatment (Garber, 2011). If an effective oral incretin mimetic is developed, it could lead to numerous benefits including convenience, greater patient satisfaction and compliance, and fewer side effects compared with the same drug that is administered systemically.

We are keenly interested in developing transferrin fusion protein technology as a novel strategy to improve the therapeutic efficacy of protein and peptide drugs and as well to achieve the preferred noninvasive oral delivery of biopharmaceuticals. Transferrin (Tf) is an abundant, naturally occurring serum protein with the capacity to bind and transport iron to cells through Tf receptor-mediated endocytosis. Tf receptor (TfR) is present on the surface of most proliferating higher eukaryotic cells. TfRs are also highly expressed in human GI epithelial cells. Furthermore, Tf is stable and resistant to proteolytic enzymes, which leads to a tremendously long plasma half-life (14–17 days) (Brandsma et al., 2011; Li and Qian, 2002; Melanie, 2005). These properties of human Tf make it a very valuable tool in developing Tf-based novel fusion protein technology to enhance protein expression, extend the serum half-lives of protein/peptide drugs and to achieve active targeted (smart) drug delivery. Being an endogenous protein, Tf as a drug delivery system is nonimmunogenic and nontoxic in humans (Jiang et al., 2007).

In recent years, plants have emerged as an attractive alternative expression platform for the production of recombinant pharmaceutical proteins and offer several advantages. Given that plant cells are eukaryotic cells, the cells are able to perform many of the post-translational protein modifications (glycosylation, disulphide bond formation and protein processing and folding) necessary for the biological functions of many mammalian proteins. Moreover, plant cells do not harbour human or zoonotic pathogens, making them a safe host for the production of biopharmaceuticals. Also, plant production of foreign proteins can be targeted into edible parts of the plant, which allows for the possibility of oral administration of therapeutic proteins in partially processed plant food products, reduces the cost of manufacturing and therefore decreases the cost of therapy for the patient (Ma et al., 2004; Ma and Wang, 2012; Tremblay et al., 2010). Recently, Kwon et al. (2012) reported the production of Ex-4 as a fusion protein with cholera toxin B subunit (CTB) in tobacco chloroplasts and demonstrated that oral delivery of plant-made Ex-4-CTB lowers blood glucose levels in mice. While CTB has proven to be an effective carrier system for transmucosal delivery of biopharmaceuticals, CTB is in itself a potent immunogen and adjuvant that elicit strong systemic and mucosal antibody responses (Kim et al., 1998). Accordingly, CTB is mostly employed as a carrier system for the delivery of oral vaccine antigens, exploiting the intrinsic immunogenic and adjuvant properties of CTB to increase immunogenicity of a linked vaccine. However, the intrinsic adjuvanticity and immunogenicity of CTB could become detrimental in cases where CTB is being used to deliver an oral protein/peptide drug to which its therapeutic effect is not related to or dependent on the induction of a specific immune response. The development of unwanted immune response against a therapeutic protein can have serious clinical consequences (Chirino et al., 2004).

In this study, we report the production of Ex-4-Tf in plants using stable and transient systems and the characterization of plant-derived Ex-4-Tf (prEx-4-Tf). We show that prEx-4-Tf retain functions of both fusion partners. In particular, the fusion protein stimulated insulin secretion from pancreatic β-cells, promoted proliferation of β-cells, stimulated differentiation of pancreatic precursor cells into insulin-producing β-cells, retained the ability to internalize into human intestinal epithelial cells and resisted stomach acid and proteolytic enzymes. Importantly, oral administration of partially purified prEx-4-Tf significantly improved glucose tolerance, whereas administration of commercial Ex-4 via the same route failed to show any significant improvement in glucose tolerance in mice. Furthermore, intraperitoneal (IP) injection of prEx-4-Tf showed a beneficial effect in mice similar to IP-injected Ex-4. We also demonstrated that plants provide a robust system for the expression of Ex-4-Tf, producing up to 37 μg prEx-4-Tf/g fresh leaf weight in T0 transgenic tobacco plants and 137 μg prEx-4-Tf/g freshweight in transiently transformed leaves of N. benthamiana. These results indicate that Ex-4-Tf holds substantial promise as a new oral therapy for T2D. The production of Ex-4-Tf in plants may offer a convenient and cost-effective method to deliver the antidiabetic medicine in partially processed plant food products.

Results

Vector construction

The plant transformation vector pBI101.1-Ex-4-Tf was constructed as described under Experimental Procedures. As shown in Figure 1, Ex-4 was fused in-frame to the N-terminal end of the mature form of human Tf through a flexible linker to minimize the steric hindrance between the two fusion partners. To ensure correct N-terminal processing of the fusion protein in plant cells to obtain a final chimeric product containing histidine at the N-terminus, a barley alpha amylase signal peptide (SP) was added to the fusion protein. The tobacco etch virus (TEV) RNA 5' untranslated leader (UTL) sequence as well as a 3'endoplasmic reticulum (ER) retention signal motif KDEL was incorporated onto pBI101.1-Ex-4-Tf to maximize fusion protein accumulation (Brandsma et al., 2010; Wang et al., 2008). A 6xHis tag was included to facilitate fusion protein purification.

Figure 1.

Diagrammatic representation of the T-DNA region of vector pBI101.1-Ex-4-Tf. RB, right border; LB, left boarder; PNOS, nopaline synthase promoter; npt II, neomycin phosphotransferase gene; TNOS, nopaline synthase terminator; e35S, enhanced double CaMV 35S promoter; TEV, tobacco etch virus 5'leader sequence; SP, barley α-amylase signal peptide; Ex-4, a synthetic gene encoding 39 amino acids (aa); GGGGSx3, linker sequence; Tf, Tf cDNA encoding a mature protein of 679 aa; 6xHis, six consecutive histidine residues; KDEL, ER-retention signal.

Stable and transient expression of Ex-4-hTf fusion protein in plants

Low-alkaloid tobacco (cultivar 81V9) was stably transformed using Agrobacterium-mediated transformation. Over 25 kanamycin-resistant primary transgenic lines (T0 generation) were generated. The expression of Ex-4-Tf fusion protein in putative T0 transgenic tobacco plants was analysed by Western blot using a commercial anti-human Tf antibody. As shown in Figure 2a, the anti-Tf antibody detected a protein band of 76 kDa, the expected size of the fusion protein. No protein of similar size was detected in extracts from wild-type plants. The yields of prEx-4-Tf in the T0 stable transgenic tobacco reached up to 37 μg/g fresh leaf weight.

Figure 2.

Western blot analysis of prEx-4-Tf expression in transgenic tobacco (a) and in infiltrated N. benthamiana leaves (b). Protein samples (40 μg per well) were separated by 10% SDS-PAGE, transferred onto PVDF membrane and probed with anti-hTf antibodies. T2 to T14, representatives of the individual T0 transgenic lines; Tf, commercial Tf (13 ng/well); dpi, days of postinfiltration. The position of prEx-4-Tf is indicated by the arrow.

Transient expression of Ex-4-Tf in N. benthamiana was achieved with the co-infiltration of Agrobacterium harbouring the gene for p19, a viral protein that inhibits RNA silencing and is known to boost the yield of transiently expressed proteins (Lakatos et al.,2004; Tremblay et al.,2011). As shown in Figure 2b, the expression of Ex-4-Tf in infiltrated leaves reached the highest level at day 6 postinfiltration (dpi), with yields upto 137 μg prEx-4-Tf/g freshweight.

Plant-derived Ex-4-Tf-stimulated glucose-dependent insulin secretion from pancreatic beta cells

Ex-4 is known to stimulate the release of insulin from pancreatic beta cells in a glucose-dependent fashion (Holst and Gromada, 2004). We therefore investigated whether prEx-4-Tf retained the ability of Ex-4 to stimulate glucose-dependent insulin secretion in the mouse β-cell line MIN6. To test its effect, the fusion protein was partially purified from leaves of stably transformed tobacco plants using a HiTrap Chelating HP column (around 50% recovery) and the purity confirmed by SDS-PAGE gel/coomassie blue staining (Figure 3). Chen et al. (2012) demonstrated that Ex-4 at 10–15 nm stimulated significant insulin secretion from MIN6 cells. Thus, we tested the effect of prEx-4-Tf on insulin secretion from MIN6 cells under different concentrations (15, 5 and 1 nm). As shown in Figure 4, in the presence of 10 mm glucose, prEx-4-Tf stimulated insulin secretion from MIN6 cells in a dose-dependent manner. Commercial Ex-4 was used as a positive control. As expected, incubation of MIN6 cells with Ex-4 resulted in a dose-dependent increase in insulin secretion. Interestingly, Ex-4-Tf appears to stimulate insulin secretion from pancreatic MIN6 cells significantly more than the commercial Ex-4 of the same concentration (P < 0.05). No stimulatory effect on insulin secretion was observed in MIN6 cells cultured with plant-derived Tf alone. Both prEx-4-Tf and commercial Ex-4 had no effect on insulin secretion in the absence of glucose in MIN6 cells (data not shown), suggesting their glucose-dependent action.

Figure 3.

Coomassie-stained 10% SDS-PAGE gel showing the partial purification of prEx-4-Tf. The fusion protein was partially purified by affinity chromatography. M, protein markers; TSP, total soluble protein (100 μg loaded); E, eluate from HiTrap™ Chelating HP Column (750 ng loaded).

Figure 4.

Effect of prEx-4-Tf on insulin secretion from MIN6 cells. MIN6 cells were treated with prEx-4-Tf or commercial Ex-4 at different concentrations in the presence of 10 mm glucose. Supernatants from treated cells were collected and assayed for insulin concentration using insulin ELISA Kits. The assay was performed in triplicates and repeated. Data were expressed as means ± SEM. *, statistically significant difference (P < 0.05) in insulin release from MIN 6 cells induced by prEx-4-Tf vs. commercial Ex-4.

Plant-derived Ex-4-Tf-stimulated proliferation of pancreatic beta cells

We further examined whether prEx-4-Tf retained the ability to stimulate pancreatic β-cells to proliferate using INS-1 cells. INS-1 cells were chosen as a β-cell line model because unlike slow-growing glucose-responsive β-cell lines such as MIN6, INS-1 cells grow at a rate that allow sensitive measurement of rate of DNA synthesis using a [3H]thymidine incorporation assay (Asfari et al., 1992; Gahr et al., 2002; Hügl et al., 1998). We tested the effect of prEx-4-Tf on glucose-dependent INS-1 cell proliferation at several concentrations (10, 1, 0.1, 0.01, 0.001 and 0.0001 nm). A previous study indicated that the thereshold concentration for GLP-1 to induce INS-1 cell proliferation was in the range of 10−12–10−11m (Buteau et al., 1999). As shown in Figure 5a, glucose by itself was able to stimulate INS-1 cell proliferation in a dose-dependent manner, and the addition of prEx-4-Tf further increased this glucose-induced INS-1 cell proliferation. For example, glucose at 8 mm increased INS-1 cell proliferation by ~9-fold compared with the control (i.e. medium without glucose). Upon addition of 0.001 nm or 0.0001 nm prEx-4-Tf, INS-1 cell proliferation increased ~8- and 5-fold, respectively, above the proliferation observed at 8 mm glucose alone, suggesting a dose-dependent effect of prEx-4-Tf. The addition of higher concentrations of partially purified prEx-4-Tf (≥0.1 nm) was found to be toxic causing considerable cell death due to the high sensitivity of INS-1 cells towards the impurities present in partially purified prEx-4-Tf (Figure 3). Interestingly, prEx-4-Tf displayed a stimulatory effect on INS-1 cell proliferation even in the absence of glucose. prEx-4-Tf at 0.001 nm, without glucose, induced ~2.5-fold increase in INS-1 cell proliferation. There was little effect of plant-derived Tf (prTf) alone on INS-1 cell proliferation in the presence or absence of glucose (data not shown), suggesting that Tf itself has no effect on β-cell proliferation. We also tested the effect of commercial Ex-4 on INS-1 cell proliferation using the same concentrations. As shown in Figure 5b, Ex-4 enhanced glucose-induced proliferation of INS-1 cells in a dose-dependent manner as expected. However, prEx-4-Tf appears to be a more potent stimulator of INS-1 cell proliferation than Ex-4, as only prEx-4-Tf demonstrated considerable stimulatory effect on INS-1 cell proliferation at concentrations below 0.1 nm (Figure 5a,b).

Figure 5.

Effect of prEx-4-Tf on pancreatic β-cell proliferation. INS-1 cells were incubated with different concentrations of prEx-4-Tf (a) or commercial Ex-4 (b) in the presence of different glucose concentrations. Cell proliferation was determined by [3H]thymidine incorporation. Data were expressed as means ± SEM. The assay was performed in triplicate and repeated. M, medium only.

Plant-derived Ex-4-Tf induced pancreatic precursor cells to differentiate into insulin-producing beta-like cells

We next examined whether prEx-4-Tf retained the ability to promote differentiation of pancreatic precursor cells into insulin-secreting cells using rat pancreatic ductal cells (ARIP) as a model. Previous work has shown that GLP-1 treatment can induce the differentiation of ARIP cells into insulin-synthesizing β-cells (Hui et al. (2001). ARIP cells were cultured with different concentrations of prEx-4-Tf (10, 1, 0.1, 0.01, 0.001 and 0.0001 nm) for 48 h in the presence of 12 mm glucose. The supernatant was assayed for insulin secretion by ELISA. As shown in Figure 6, insulin levels in the supernatants of ARIP cells treated with 0.01, 0.001 nm or 0.0001 nm Ex-4-Tf were all significantly elevated compared with the basal insulin level measured in supernatants of ARIP cells grown in 12 mm glucose without prEx-4-Tf (P < 0.05). Partially purified prEx-4-Tf at higher concentrations (≥0.1 nm) was toxic to ARIP cells. The treatment of ARIP cells with Ex-4 of the same concentrations were used as controls. As expected, Ex-4 induced insulin secretion in a dose-dependent manner (Figure 6). As prEx-4-Tf, but not Ex-4, at concentrations below 0.01 nm were still capable of inducing insulin secretion, prEx-4-Tf appears to be more potent than Ex-4 in promoting the differentiation of ARIP cells into insulin-secreting cells.

Figure 6.

Effect of prEx-4-Tf on differentiation of pancreatic precursor cells. ARIP cells were treated with different concentrations of prEx-4-Tf or Ex-4 for 48 h in the presence of 12 mm glucose (detailed in the section of Experimental Procedures). Supernatants from treated cells were collected and assayed for insulin concentrations using a commercial insulin ELISA Kit. M, medium containing 12 mm glucose only. *, statistically significant difference (P < 0.05) in levels of insulin in supernatants from cells treated with prEx-4-Tf or Ex-4 compared with basal insulin levels measured in the supernatant from untreated control cells.

Plant-derived Ex-4-Tf retained the ability to be internalized into human intestinal epithelial cells

Internalization into mammalian cells is an important property of Tf (Li and Qian, 2002). Therefore, we investigated whether prEx-4-Tf would carry the ability of Tf to get internalized using the human intestinal epithelial cell line Caco-2, an in vitro cell model widely used to study cellular uptake, transport, metabolism processes or oral bioavailability of drug candidates (Angelis and Turco, 2011). Caco-2 cells were incubated with 1.0 μg or 0.2 μg/mL of prEx-4-Tf, prTf or Tf standard for 60 min. Cells were then treated with acid buffer to remove membrane-bound foreign proteins. Following cell lysis, the supernatant was collected after centrifugation and analysed for the presence of target proteins by ELISA. The ELISA results shown in Figure 7a, clearly demonstrate the presence of prEx-4-Tf, prTf or Tf in the lysate supernatant from the treated cells. A correlation between the amount of the protein added to the cell culture and the amount of the protein internalized into cells was also observed. No significant difference was observed between prEx-4-Tf and prTf in their ability to get internalized. However, compared with Tf standard, both prEx-4-Tf and prTf had a significant reduction in their ability to get internalized (P < 0.05). The presence of prEX-4Tf, prTf or Tf standard in the supernatant was additionally visualized by Western blot (Figure 7b). No detectable levels of endogenous Tf were found in lysate supernatants from control Caco-2 cells grown in the absence of prEx-4-Tf or prTf.

Figure 7.

Analysis of cellular uptake of prEx-4-Tf. Caco-2 cells were incubated with 1.0 μg/mL (1) or 0.2 μg/mL (0.2) of prEx-4-Tf, prTf or Tf. Following incubation, the cell membrane-bound foreign protein was removed. Cells were lysed, and supernatants were collected after centrifugation and assayed for the concentrations of internalized target proteins by ELISA (a) and Western blot (b). The ELISA was performed in triplicate and repeated. Data were expressed as the mean with standard deviation. For Western blot, 6 μg of total protein was separated on a 10% SDS-PAGE. +, Tf standard (16 ng); *, statistically significant difference (P < 0.05) in concentrations of internalized Tf standard in lysate supernatants from Tf treated cells compared with concentrations of internalized prEx-4-Tf or prTf in lysates from prEx-4-Tf or prTf-treated cells. The position of internalized proteins is indicated by the arrow.

Plant-derived Ex-4-Tf resisted acidic conditions of simulated stomach environment

We also investigated the stability of prEx-4-Tf in a simulated acidic stomach environment, as the development of an oral version of Ex-4 requires the protein to be stable and remain intact in the acidic environment of the stomach (pH 1.5–3.5). This was tested by subjecting prEx-4-Tf to a synthetic gastric fluid solution containing pepsin, an enzyme in the stomach that breaks down proteins, as reported by Shaji and Patole (2008) and Tremblay et al. (2010). As revealed by Westen blot, prEx-4-Tf was found to be stable in the synthetic gastric fluid solution (Figure 8). Plant-derived Tf was used as a control. There was no difference observed in the stability against digestion by the synthetic gastric fluid between the two proteins.

Figure 8.

Digestion of partially purified prEx-4-Tf in synthetic gastric fluid (pH 2.5). Partially purified prEx-4-Tf (a) or prTf (b) was digested in the synthetic gastric fluid for different lengths of time. Digestion stopped by adding neutralization buffer. The digested samples were separated by 10% SDS-PAGE, transfered to membrane and probed with anti-Tf antibodies. Each lane contains the starting equivalent of approximately 100 ng purified protein.

Oral and intraperitoneal routes of administration of partially purified Ex-4-Tf improved glucose tolerance in mice

We first investigated whether oral delivery of prEx-4-Tf could enhance glucose metabolism, as one of our major goals was to develop novel Ex-4 derivatives that can be used as safe and effective oral antidiabetic medicines. To determine its effect, fasting plasma glucose levels of C57BL/6J male mice before and after receiving oral prEx-4-Tf were measured and compared with those in the control and Ex-4 group. In the Ex-4-Tf group, the plasma glucose level 1 h after oral treatment tended to be lower when compared with basal levels before treatment (P = 0.07, Figure 9a), indicating that oral administration of prEx-4-Tf at 1.5 μg/g did not significantly affect basal plasma glucose levels. This data suggest that unlike many conventional antidiabetic medicines, no side effect risks such as hypoglycaemia are associated with the use of prEx-4-Tf in rodent. No differences were observed in fasting plasma glucose levels in mice from both control and Ex-4 groups before and after oral treatment. While IPGTT results revealed that all experimental groups exhibited a plasma glucose peak at 30 min after glucose loading followed by a gradual return to basal levels by 120 min (Figure 9b), the prEx-4-Tf-treated mice displayed significantly improved glucose tolerance, as demonstrated by decreased area under the curve (AUC) compared with the control (P < 0.01, Figure 9c) or Ex-4 group (P < 0.05, Figure 9c). No significant difference in glucose clearance was observed between the control group and the Ex-4 group.

Figure 9.

Effect of oral administration of partially purified prEx-4-Tf on plasma glucose homoeostasis in mice. (a) fasting plasma glucose and (b) IPGTT of mice treated with sterile saline (Ctrl), partially purified prEx-4-Tf (Ex-4-Tf) or commercial Ex-4 (Ex-4) via an oral route. (c) Glucose responsiveness of the corresponding experimental groups is shown as a measurement of area under the curve (AUC) of the IPGTT graphs with units of mm/min. Data are expressed as means ± SEM (n = 5–6). ** P < 0.01 (One-way anova) and ## P < 0.05 (unpaired Student's t-test) versus the ctrl group.

We also studied the effect of the intraperitoneally delivered prEx-4-Tf on glucose metabolism in mice. No changes in fasting plasma glucose levels was observed among the experimental groups, although both Ex-4-Tf and Ex-4 groups tended to have lower fasting plasma glucose levels when measured at 30 min post-treatment (Figure 10a). However, IPGTT results showed a significant enhancement in glucose clearance, occuring at 15 min post-glucose loading, in both Ex-4-Tf and Ex-4 groups, as evidenced by lower AUC compared with control group (Ex-4-Tf group vs. control, P < 0.01; Ex-4 group vs. control, P < 0.001) (Figure 10b,c). No significant difference in IPGTT was observed between the prEx-4-Tf group and the Ex-4 group.

Figure 10.

Effect of intraperitoneal injection of purified Ex-4-Tf on plasma glucose homoeostasis in mice. (a) fasting blood glucose and (b) IPGTT of mice received an i.p. injection of sterile saline (Ctrl), partially purified prEx-4-Tf (Ex-4-Tf), and commercial Ex-4 (Ex-4). (c) Glucose responsiveness of the corresponding experimental groups is shown as a measurement of area under the curve (AUC) of the IPGTT graphs with units of mm/min. Data are expressed as means ± SEM (n = 5–6). ** P < 0.05, *** P < 0.001 (One-way anova) versus the ctrl group.

Discussion

In this work, we produced in plants Ex-4 as a fusion protein with Tf as a strategy for achieving novel effective oral delivery of prEx-4-Tf for the treatment of T2D. Currently, exenatide (synthetic exendin-4) is only available in injectable form and moreover, twice daily injections are required to achieve near-normal glucose control in T2D patients. This makes the incretin-based treatment to be inconvenient, expensive and stressful. An oral version of exenatide would be more advantageous, enhancing ease in administration, reducing side effects and improving patient acceptance and compliance (Brandsma et al., 2011). We demonatrated that prEx-4-Tf exhibited the activity of both EX-4 and Tf. We also showed that the antidiabetic activity of Ex-4 was also enhanced by fusion to Tf. More importantly, we demonstrated that oral administration of partially purified prEx-4-Tf significantly improved glucose tolerance, whereas commercial Ex-4 administered by the same oral route failed to show any significant improvement in glucose tolerance in mice. These findings have important implications. One implication is that prEx-4-Tf has good potential to be developed into a new effective oral antidiabetic medicine. Another implication is that the production of Ex-4-Tf in plants may offer a cost-effective method to deliver oral prEx-4-Tf in partially processed food products. Indeed, preliminary results from our newest animal feeding trial to test whether oral consumption of unprocessed prEx-4-Tf tobacco leaves can have an effect indicated that mice fed raw leaf material for short time showed some improvement in glucose tolerance compared with mice fed wild-type leaf material or normal feed, although the difference did not reach statistical significance (data not shown), supporting the feasibility of using plants as a vehicle to deliver oral prEx-4-Tf. It should be mentioned that in this new trial, the amount of prEx-4-Tf delivered per mouse daily was lower compared with the amount of prEx-4-Tf delivered as partially purified protein by oral gavage, due to the limitation on the plant material that can be added into animal feed as a dietary supplement.

Our in vitro cellular assay data showed that the antidiabetic activity of prEx-4-Tf was improved compared with commercial Ex-4. This is highly likely to have resulted from increased stability of Ex-4 acquired through its fusion to Tf. Human Tf is very stable, possessing a t1/2 in excess of 14–17 days (Brandsma et al., 2011; Li and Qian, 2002). Therefore, if a fusion molecule between Ex-4 or any peptide and Tf can be created, the stabilizing properties of Tf will likely be imbued to the peptide. The results of our in vitro stability study showed that prEx-4-hTf or prTf was stable and resistant to synthetic digestive fluids (Figure 8), further supporting this assertion. Correct N-terminal processing of prEx-4-Tf may also play a major role in assuring the maximal antidiabetic activity of the fusion protein. The N-terminal histidine (His) of GLP-1 or Ex-4 has been shown to be critical for pancreatic receptor binding and insulinotropic activity. Removal of this N-terminal His from GLP-1or Ex-4 or its replacement with other amino acids was shown to result in drastic loss of its receptor binding and insulinotropic activity (Xiao et al., 2001; Kieffer and Habener, 1999). Previously, we expressed GLP-1as a large multimeric protein (GLP-1 × 10) in plants (Brandsma et al., 2009). As the N-terminus of GLP-1 × 10 contains two additional amino acids (Met and Gly) serving as part of a translational start site, such an N-terminal extension had a significantly negative impact on GLP-1 activity. To ensure that plant-derived Ex-4-Tf has His at its N-terminus, Ex-4-hTf was fused with barley α-amylase signal peptide for effective processing. We have previously demonstrated that barley α-amylase signal peptide is capable of directing efficient and accurate processing of human interleukin-4 when expressed in plants (Ma et al., 2005).

The ability of prEx-4-Tf to get internalized was demonstrated using human intestinal Caco-2 cells. The level of internalized prEx-4-Tf in Caco-2 cell lysates was quantified by ELISA and confirmed by Western blotting (Figure 7). There was good correlation between the amount of the protein added to the cell culture and the amount of the protein internalized into cells. No significant difference was observed between prEx-4-Tf and prTf; however, both prEx-4-Tf and prTf had significant reduction in their ability to get internalized compared with Tf standard (P < 0.05). Although further studies are needed, the difference between Tf standard and prEx-4-Tf or prTf may be due to difference in protein quality. Tf standard used in this study was in highly purified form, whereas prEX-4-Tf or prTf was only in partially purified form. We (Brandsma et al., 2010) and others (Zhang D et al., 2010) have shown previously that plant-derived human Tf is not glycosylated, though native Tf is a glycoprotein. However, the glycosylation on the Tf has no known influence on receptor binding or any other biological function (Huebers and Finch, 1987). With Ex-4, the fusion protein only had a 4-kDa increase in molecular weight. Therefore, the size of the fusion protein is not likely to be the factor affecting its cellular uptake. Chen et al. (2011) investigated the possible effect of different linkers connecting Tf to growth hormone (GH) or granulocyte colony-stimulating factor (G-CSF) on receptor binding. They showed that the linkers can exert positive or negative effects on the binding affinities of the fusion protein to both protein receptors, thereby affecting subsequent endocytosis of Tf. It would be worthwhile to test other linkers than GGGGSx3 to see if the cellular uptake of prEx-4-Tf can be enhanced.

The data from animal feeding trials revealed that oral administration of partially purified Ex-4-Tf (1.5 μg/g body weight) significantly improved glucose tolerance but had little effect on the basal plasma glucose level in mice (Figure 9a–c). These results suggest that Ex-4-Tf as an oral antidiabetic medicine is not only effective but also carries no or low risk of hypoglycaemia that can lead to seizures, coma and even death. The use of conventional antidiabetic medicines such as insulin is often associated with increased risk of developing hypoglycaemia. The results also suggest that Tf fusion protein technology can be used as an effective strategy to achieve oral delivery of Ex-4 and probably many other therapeutic proteins. In recent years, there has been increasing interest in exploiting Tf fusion protein technology for the development of orally effective peptide and protein drugs. Xia et al. (2000) reported that oral administration of insulin-Tf conjugate (In-Tf) to diabetic rats lowered blood sugar levels. Bai et al. (2005) reported that oral administration of G-CSF and Tf fusion protein promoted proliferation of neutrophils in BDF1 mice. Amet et al. (2010) showed that rats receiving oral human growth hormone hGH-Tf fusion protein (GHT) gained body weight. Recently, Bobst et al. (2012) investigated possible mechanisms underlying the successful oral delivery of GHT through analysis of the fusion protein's stability to proteolysis and its binding affinity for Tf receptors, the key factors in overcoming the primary barriers to successful oral delivery. They showed that in addition to the anticipated monomeric form (GHT1), a significant population of GHT exists in an oligomeric form (GHTx), a form proving to be exceptional stable in gut environment. On the other hand, oligomerized GHT did not affect its binding to Tf receptor.

The relative high-level accumulation of prEx-4-Tf in plants (up to 37 μg/g fresh leaf weight in transgenic tobacco and 140 μg/g freshweight in infiltrated leaves) may be attributed to several factors acting in synergy, including the stable nature of Tf itself, the use of a strong CaMV 35S promoter and the targeting of prEx-4-Tf into the lumen of the ER. Moreover, the use of a flexible linker (GGGGS)x3 between Ex-4 and Tf may also contribute to its higher expression. Trinh et al. (2004) reported a 30-fold increase in the yield of a hybrid protein consisting of a single-chain Fv antibody specific for HER2/neu and the CH3 region of human anti-rat transferrin receptor IgG3 heavy chain when they were connected through the (GGGGS)x3 linker and expressed in a mammalian transient expression system. To achieve higher and more stable expression, we have begun to produce transgenic tobacco lines homozygous for the introduced Ex-4-Tf gene.

Recently, Kim et al. (2010) reported the expression of Ex-4 fused to Tf in yeast and demonstrated that injection of yeast-derived Ex-4-Tf lowered blood glucose levels in mice. While yeast may provide a useful system for the expression of Ex-4-Tf or other therapeutic proteins, a major disadvantage of yeast-based expression system is its comparatively low-expression yield. The vacuoles of yeast cells, the homologue to lysosomes of higher cells, are filled with highly active proteases that can cause rapid degradation of recombinant proteins during cell breakage (Holkeri and Makarow, 1998). Moreover, like bacterial and mammalian cells, yeast requires complex cell culture facilities that are not only expensive but also make the scaling up of pharmaceutical protein production difficult. However, these problems can likely be solved with the use of plants as an alternative expression host.

In summary, we have produced a fusion protein consisting of Ex-4 coupled to Tf in plants. The in vitro analysis showed that prEx-4-Tf retained the activity of both proteins. Oral administration of partially purified prEx-4-Tf significantly improved glucose tolerance but had little effect on basal plasma glucose level, whereas administration of Ex-4 by the same oral route had no effect in mice. Furthermore, IP injection of partially purified prEx-4-Tf showed a beneficial effect similar to IP-injected Ex-4 in mice. These results indicate that prEx-4-Tf holds promising potential as a new oral therapy for type 2 diabetes. The expression of Ex-4-Tf in plants may offer a convenient and cost-effective method to deliver the antidiabetic medicine in partially processed plant food products. This study also suggests that Tf fusion protein technology may offer a powerful new method for improving the therapeutic efficacy of peptide and protein drugs.

Experimental procedures

Construction of Ex-4-Tf fusion protein expression vector

Ex-4 is a 39–amino acid peptide. A synthetic gene encoding Ex-4 was assembled based on published cDNA sequence (Pohl and Wank, 1998). A cDNA clone encoding human Tf was obtained from OriGene (Rockville, MD). Standard PCR and recombinant DNA techniques were used to assemble Ex-4 and Tf as a fusion gene. Briefly, the synthetic Ex-4 gene was modified by adding a DNA sequence encoding barley α-amylase signal peptide (SP) (Rogers and Milliman, 1983) to its 5' end, whereas the Tf gene was modified by deleting its native signal peptide sequence (Yang et al., 1984) combined with adding a 6xHis-tag and an ER-retention signal (KDEL) sequence to its 3' end. The modified Ex-4 gene was fused in-frame to the N-terminal end of Tf gene through a (GGGGS)3 linker (Trinh et al., 2004). The resulting Ex-4-Tf chimeric gene was cloned into plasmid pRTL-GUS (Carrington and Freed, 1990) by replacing the GUS gene. The Ex-4-Tf expression cassette, consisting of 35S promoter, SP-Ex-4-Tf -6xHis-KDEL and NOS terminator, was released from pRTL-Ex-4-Tf and cloned into pBI101.1 (Brandsma et al., 2010) to obtain the final construct pBI101.1-Ex-4-Tf.

Plant transformation

Prior to plant transformation, pBI101.1-Ex-4-Tf was transferred into Agrobacterium tumefaciens LBA4404 by tri-parental mating (Ma et al., 2005). For stable transformation, leaf discs of tobacco (cv.81V9) were transformed with Agrobacterium containing pBI101.1-Ex-4-Tf using standard techniques (Horsch et al.,1985). For transient transformation, 6–8-week-old leaves of N. benthamiana were co-infiltrated with two strains of Agrobacterium harbouring pBI101.1-Ex-4-Tf and the p19 silencing suppressor as previously described by Tremblay et al. 2011). Leaf tissues were harvested at 1–7 days post-infection (dpi).

Accumulation of Ex-4-Tf fusion protein in plants

The expression of Ex-4-Tf in plants was analysed by Western blot using commercial anti-human Tf antibodies as described by Brandsma et al. (2010).

The level of prEx-4-Tf accumulation was quantified by ELISA for Tf as described by Brandsma et al. (2010), compared with known quantities of Tf standard (Sigma-Aldrich Canada Co., Oakville, Ontario, Canada).

His-purification of plant-derived Ex-4-Tf fusion protein

Hs-tagged prEx-4-Tf was purified from leaf extracts of stable transgenic tobacco plants using HiTrap™ Chelating HP Columns (GE Healthcare Life Sciences, Baie d'Urfe, Quebec, Canada). Eluted Ex-4-Tf fractions were dialysed extensively against PBS and concentrated using a speed vacuum at 4 °C.

Effect of plant-derived Ex-4-Tf fusion protein on insulin secretion from pancreatic β-cells

The effect of prEx-4-Tf on insulin secretion was evaluated in the mouse beta cells MIN6 as described by Brandsma et al. (2009). In brief, MIN6 cells were cultured in DMEM high glucose with 50 μM 2-mercaptoethanol and 10% (v/v) foetal calf serum (FCS). Once the cells reached about 80% confluence, they were seeded into 96-well (flat-bottomed) microtiter plates at a density of 3x104 cells per well. Following incubation for 3 days, cells were washed twice with Earle's balanced salt solution (Sigma) containing 0.1% BSA. After starvation in EBSS plus 0.1% BSA for 1 h, cells were incubated with prEx-4-Tf, commercial Ex-4 or prTf in the presence or absence of glucose. After incubation for 135 min, cell culture supernatants were collected, and insulin content was measured using mouse insulin ELISA Kits (Crystal Chem Inc., Downers Grove, IL) according to the manufacturer's instructions.

Effect of plant-derived Ex-4-Tf fusion protein on proliferation of pancreatic beta cells

The effect of prEx-4-Tf on beta cell proliferation was evaluated using the beta cell line INS-1 according to Buteau et al. (2001). Briefly, INS-1 cells were grown in RPMI 1640 medium. Two days before the experiment, INS-1 cells were seeded in 96-well plates (8 × 104 cells/well) and grown in RPMI medium. Cells were washed with PBS and preincubated for 24 h in minimal RPMI medium, that is, without serum and glucose but with 0.1% BSA. Cells were then grown in fresh minimal RPMI medium containing prEx-4-Tf, prTf or Ex-4 in the presence of glucose. Proliferation was determined by incorporation of [3H]thymidine during the final 4 h of the 24-h incubation period. Cells were harvested with a PHD cell harvester (Cambridge technology) and the radioactivity retained on the dried glass fibre filters was measured.

Effect of plant-derived Ex-4-Tf fusion protein on the differentiation of pancreatic precursor cells

The differentiating effects of prEx-4-Tf on pancreatic precursor cells into insulin-producing cells was evaluated in the rat pancreatic ductal (ARIP) cell according to the method of Hui et al. (2001). In brief, ARIP cells were grown to 80% confluence in F12K medium (Gibco-BRL) with 10% FCS, washed with serum-free F12K followed by ‘wash-out’ incubation for 6 h with F12K medium. Cells were then incubated with fresh F12K containing prEx-4-Tf or Ex-4 at 12 mm glucose. After incubation for 48 h, the supernatants of the treated cells were collected and analysed for insulin concentration content using mouse insulin ELISA Kits.

Assay of cellular internalization ability of plant-derived Ex-4-Tf fusion protein

Cellular internalization capacity of Ex-4-Tf was assessed using the human intestinal epithelial cells Caco-2. Caco-2 cells were maintained in tissue culture flasks containing RPMI-1640 (Invitrogen) with 10% FCS. To assay cellular uptake, cells were plated to 60 × 15-mm petri plates. Following incubation for 24 h, cells were washed twice with serum-free F12 medium and incubated in fresh serum-free F12 medium for 1 h at 370 °C to starve the cells of serum. Following the addition of prEx-4-Tf, prTf or Tf standard, plates were incubated for a further 60 min at 37 °C. The reaction was stopped by placing plates on ice, followed by washes with cold PBS to remove excess and unbound proteins. The cell membrane–bound foreign protein was removed by treating plates with acid buffer (500 mm NaCl, 200 mm acetic acid, pH 4.5) for 5 min on ice followed by washes with PBS as described by Karin and Mintz (1981). Cells were then lysed with lysis buffer (0.5% (v⁄v) Triton X-100, 10 mm Hepes, 10 mm KCl, 1 mm EDTA, 0.1 mm EGTA, 0.1% NP40, 1 mm DTT, 0.5 mm PMSF) and centrifuged at 10 000 g at 4 °C for 10 min to remove cellular debris. The supernatant (containing cytoplastic and nuclear proteins) was collected and assayed for released prEx-4-Tf by ELISA using anti-human Tf antibodies and visualized on 10% (w/v) SDS–PAGE gels followed by Western blot as desctibed by Brandsma et al. (2010).

Digestion of Plant-derived Ex-4–hTF in synthetic gastric fluid

The stability of prEx-4-Tf in an acidic stomach environment was assessed in synthetic gastric fluid according to Tremblay et al. (2011) with minor modifications. Briefly, the His-purified protein sample was incubated in synthetic gastric fluid (0.2 g NaCl, 0.32 g pepsin, 700 μL HCl, in 100 mL dH2O, pH 2.5) at 37 °C. The digestion was stopped by the addition of neutralization buffer (3.4 g Na2CO3 in 100 mL dH2O) at times 0, 15 s, 30 s, 1 min, 5 min, 15 min, 30 min and 1 h. The neutralized samples were boiled for 10 min and analysed by 10% SDS-PAGE followed by Western blot using antibody against human Tf.

Administration in mice of partially purified Ex-4-Tf to mice by oral and intraperitoneal routes

The C57BL/6J male mice at 8 weeks of age (Charles River, Senneville, Quebec, Canada) were used. All mice had free access to standard diet and water. The animal protocol used was approved by Animal User Subcommittee at the University of Western Ontario in accordance with the guiddlines of the Canadian Council of Animal Care.

Mice with similar body weight underwent overnight fasting (12 h) was randomly divided into three experimental groups: Control, Ex-4-Tf and Ex-4. For the study of oral delivery, the control group received an oral gavage of sterile saline, while the Ex-4-Tf and Ex-4 group received an oral gavage of partially purified prEx-4-Tf and commercial Ex-4 (1.5 μg/g body weight), respectively, 1 h prior to intraperitoneal glucose tolerance test (IPGTT). For the study of intraperitoneal delivery, the control group received sterile saline by injection, while the Ex-4-Tf and Ex-4 group received an injection of partially purified prEx-4-Tf or commercial Ex-4 (10 ng/g body weight) 30 min prior to IPGTT. Plasma glucose level before and after the treatment was measured.

For IPGTT, mice were given an intraperitoneal injection of glucose (2 mg/g; D-(+)-glucose; Sigma), and plasma glucose levels were then measured at 0, 15, 30, 60 and 120 min after injection. Area under curve (AUC) was used to quantify responsiveness (Feng et al., 2012a,b, 2013). Five to six mice were used per experimental group (n = 5–6).

Statistics

Data are expressed as mean±SE (standard error). Statistically significant differences between groups were analysed by using the Student's t-test or an anova followed by the Bonferroni post hoc test. Differences were considered to be statistically significant when P < 0.05.

Acknowlegements

INS-1 cells were received as a gift from Dr. Savita Dhanvantari, Department of Medical Biophysics at the University of Western Ontario. This research was supported in part by NSERC. The authors have no conflicts of interest to declare.

Ancillary