Annals of Biomedical Engineering

, Volume 44, Issue 6, pp 1993–2007

Delivery of Exenatide and Insulin Using Mucoadhesive Intestinal Devices

  • Vivek Gupta
  • Byeong-Hee Hwang
  • Nishit Doshi
  • Amrita Banerjee
  • Aaron C. Anselmo
  • Samir Mitragotri
Emerging Trends in Biomaterials Research

DOI: 10.1007/s10439-016-1558-x

Cite this article as:
Gupta, V., Hwang, BH., Doshi, N. et al. Ann Biomed Eng (2016) 44: 1993. doi:10.1007/s10439-016-1558-x

Abstract

A major disadvantage associated with current diabetes therapy is dependence on injectables for long-term disease management. In addition to insulin, incretin hormone replacement therapies including exenatide have added a new class of drugs for Type-2 diabetes. Although efficacious, patient compliance with current diabetic therapy is poor due to requirement of injections, inability to cross the intestinal epithelium and instability in the gastrointestinal tract. Here, we report the efficacy of a mucoadhesive device in providing therapeutic concentrations of insulin and exenatide via oral administration. Devices were prepared with a blend of FDA-approved polymers, carbopol, pectin and sodium carboxymethylcellulose, and were tested for drug carrying capability, in vitro release, Caco-2 permeability, and in vivo efficacy for insulin and exenatide. Results suggested that mucoadhesive devices successfully provided controlled release of FITC-insulin, released significant amounts of drug, while providing noteworthy enhancement of drug transport across Caco-2 monolayers without compromising monolayer integrity. In-vivo administration of the devices provided significant enhancement of drug absorption with 13- and 80-fold enhancement of relative bioavailability for insulin and exenatide compared to intestinal injections with significant increase in half-lives, thus resulting in prolonged blood glucose reduction. This study validates the efficacy of mucoadhesive devices in promoting oral peptide delivery to improve patient compliance and dose adherence.

Keywords

Type-2 diabetes Oral delivery Peptides Absorption enhancement Exenatide Insulin 

Introduction

Type-2 diabetes mellitus (T2DM), a chronic metabolic disorder, is the most common form of diabetes which impairs body’s ability to metabolize glucose due to inadequate insulin synthesis by pancreatic β-cells and by acquired insulin resistance.25 Diabetes, in particular T2DM, has become an epidemic with eightfold increase in its prevalence in the last 5 decades. While there is no permanent cure for diabetes, there are various small molecule drugs available for management of Type-2 diabetes including thiazolidonediones, glitazides, and sulfonyl urea, among others.3 These agents work by improving insulin secretion and its utilization by the body. However, they fail to work in Type-1 diabetes where insulin is not produced by the body, and also in later stages of Type-2 diabetes where they are no longer capable of providing appropriate glycemic control due to degradation of pancreatic β-cells.

Following the failure of oral antidiabetics, exogenous supply of insulin remains the only available option for optimum control of glucose levels. In addition to insulin, the gold standard for diabetes management, FDA has recently approved several drugs in another class of antidiabetics, glucagon like peptide-1 (GLP-1) receptor agonists (GLP-1 analogs), which work by enhancing glucose-dependent insulin release and reducing glucagon secretion.9 Exenatide (synthetic exendin-4) was the first member of GLP-1 analogs to be approved by the FDA.22 As both insulin and exenatide are macromolecular peptide drugs, their primary mode of administration is by invasive subcutaneous injections. Although effective, this mode of delivery poses a variety of shortcomings including patient non-compliance, pain and irritation at the injection site and non-physiological insulin delivery by injection.27 According to several published studies, the nonadherence/noncompliance rates with antidiabetic therapy are as high as 50%. A recent study by Garcia-Perez et al. determined therapy adherence for insulin in type-2 diabetes patients to be approximately 62%, with average adherence to insulin treatment (measured as the percentage of the number of days per annum of insulin coverage) being 71%. As many as 50% patients demonstrated moderate to low compliance to oral antidiabetic therapy.14,24,39,43

The poor patient compliance and dose adherence associated with diabetes therapy is an outcome of a multitude of factors, two major factors being (i) forgetting the dose due to multiple doses every day, and (ii) needle phobia. While there are no published estimates on prevalence of fear of needles, it is assumed that at least 10% of the US population suffers from needle phobia.8,51 This translates into a large patient population, thus contributing significantly toward noncompliance to injectable antidiabetic therapies such as insulin and exenatide.

Considering the chronic nature of antidiabetic therapy, it has become important to develop a more patient compliant route of delivery. Oral delivery of therapeutics has long been a preferred route of administration.36 However, therapeutic delivery of protein and peptide drugs by oral route poses significant technical hurdles due to rapid degradation by gastric enzymes and acidic environment, and poor permeation across tight junction intestinal epithelium.31 Oral delivery of antidiabetic macromolecules including insulin is likely to present a more physiologically relevant means of glucose control by mimicking endogenous insulin secretion pathway via hepatic portal circulation. Being the primary site of glucose regulation homeostasis, liver is the most important target of endogenous insulin.13 Studies suggest that only 20% of subcutaneously injected insulin reaches to the liver, thus disturbing finely controlled glucose balance in the body.15 At the same time, oral delivery of antidiabetics may reduce the occurrence of hypoglycemic episodes, insulin resistance, and immunogenic reactions while significantly improving patient compliance and associated healthcare costs.36

While there have been numerous studies describing a variety of alternatives for oral delivery of antidiabetic peptides, none of these approaches are currently used in the clinic. Some of the most widely studied approaches include the use of micro/nanoparticles, oral films, absorption enhancers, protease inhibitors, bioadhesive polymers and chemical modifications to peptides to name a few. While these approaches have provided a novel outlook toward the problem, only some of them have been tested in clinical studies and the challenges associated with low efficacy and high toxicity persist.28

In this study, we describe the utility of a mucoadhesive device, inspired from transdermal drug delivery patches, in enhancing oral/intestinal absorption of insulin and exenatide. This study capitalizes on findings from our laboratory describing the utility of these devices in promoting oral delivery of peptides.20,49 Here, we exploit the potential of mucoadhesive devices in promoting intestinal absorption of therapeutic macromolecules by following engineering and formulation principles.

Materials and Methods

Materials

For synthesis of mucoadhesive devices, pharmaceutical grade Carbopol-934 was obtained from Lubrizol Advanced Materials Inc. (Cleveland, OH, USA). Sodium carboxymethylcellulose (SCMC), pectin, and ethyl cellulose (EC) were obtained from Sigma Aldrich (St. Louis, MO, USA). Bovine insulin (MW = 5800 Da; ≥ 25 Units/mg) and recombinant human insulin (MW = 5800 Da; ≥ 27.5 Units/mg) were obtained from Sigma Aldrich (St. Louis, MO, USA). Exenatide (Exendin-4; MW = 4186.6, net peptide content >90%) was purchased from Tocris Biosciences (Minneapolis, MN, USA). Transwell Caco-2 system was set up using 24 well BD-BiocoatTM HTS Caco-2 assay system (fibrillar collagen coated, 1 µm pore size) obtained from BD Biosciences (Bedford, MA, USA). ELISA kits for bovine and human insulin were purchased from Mercodia, Inc. (Winston Salem, NC, USA); and extraction-free ELISA kit for exenatide was purchased from Bachem Americas, Inc. (Torrance, CA, USA). Supplies for Caco-2 culture and in vivo studies were obtained from Fisher Scientific (Pittsburgh, PA, USA). All other chemicals used were of analytical grade and were obtained from various vendors.

Mucoadhesive Devices

The mucoadhesive polymeric devices were prepared by following previously published protocol.20,48 Briefly, a homogenous mixture of carbopol 934/pectin/SCMC in a dry weight ratio of 1:1:2 was prepared by manual grinding. 110 mg of this homogeneous mixture was poured into a 13 mm pellet press (Pike Technologies, Madison, WI, USA), and was directly compressed under a pressure of 3 tons using hydraulic press (Carver Inc., Wabash, IN, USA) for 5 min, thus producing a 400 µm thick disk. The produced disk was cut into smaller disks of 3–5 mm using disposable biopsy punches (Miltex Inc., Plainsboro, NJ, USA). A moisture and permeation barrier was placed on all but one side of the disks by coating with 5% w/v ethylcellulose solution in acetone thus resulting in approximately 50 µm thick ethylcellulose layer following evaporation of acetone at room temperature.

For preparation of insulin loaded devices, either human or bovine insulin was added to the polymeric mixture by manual grinding so as to produce a final insulin concentration in the range of 1–10% w/w (dry weight %). For loading of exenatide, 100 µL of 1 mg/mL exenatide solution in acetone was manually added to premade blank disk so as to produce 0.1% final drug loading for in vitro studies. The disk was allowed to air dry before cutting the small 5 mm devices. FITC-insulin was loaded onto mucoadhesive devices for in vitro release studies in 3–30% w/w ratio. For in vivo studies, exenatide powder was added to the polymeric mixture as described above to produce a final concentration of 3%.

Human insulin-loaded rectangular devices were synthesized by using the same manual grinding and compression techniques, but instead of cutting round disks using biopsy punches, the devices were cut into rectangles of 2 mm × 13 mm, weighing approximately 15 mg. As mentioned earlier, a moisture and permeation barrier was placed on all but one side of the disks by coating with 5% w/v ethylcellulose solution in acetone, followed by air-drying.

In vitro Drug Release Studies

In-vitro drug release studies were performed to evaluate efficacy of the devices in (i) carrying and releasing varying therapeutic payloads; and (ii) providing a sustained and prolonged release of therapeutic macromolecules. For the first study, 5 mm (≈17 mg) devices were prepared, loaded with varying amount (3–30% w/w) of FITC-Insulin. The devices were incubated with 10 mL sterile PBS at room temperature with gentle shaking (approximately 5 rocks/min). Samples were collected at predetermined time intervals for 5 h followed by immediate replacement with equivalent amount of experimental medium. Withdrawn samples were analyzed for FITC-insulin content using a Tecan SaffireTM fluorescent microplate reader at excitation wavelength of 488 nm and emission wavelength of 525 nm.

To determine efficacy in providing a sustained and prolonged drug release, a 5 mm (≈17 mg) device loaded with either bovine insulin (0.3 mg/device) or exenatide (16 µg/device), was incubated at room temperature with 10 mL sterile PBS (pH 7.4) with gentle rocking (approximately 5 rocks/min), so as to emulate mucociliary and peristaltic movement in small intestine. At predetermined time points, 100 µL samples of the dissolution media were withdrawn followed by immediate replenishment with fresh PBS, up to 5 h. Withdrawn release samples were analyzed using commercially available ELISA kits, as described in “Materials” Section, to quantify the amount of bovine insulin or exenatide released from the devices due to time-dependent swelling and degradation.

Enhancement of Insulin and Exenatide Transport Across Caco-2 Monolayers

In-vitro testing for mucoadhesive devices’ capability to increase oral macromolecule absorption via small intestine was performed by using a Caco-2 monolayer transport model. Previously, our laboratory reported the use of a rapid 3-day Caco-2 (HTB-37; American Type Culture Collection, Manassas, VA, USA) transwell system following a protocol using BD BiocoatTM HTS Caco-2 assay system (BD Biosciences, Bedford, MA, USA).18 To assess the efficacy of mucoadhesive devices, caco-2 monolayers were developed on fibrillar collagen-coated inserts (1 µm pore size). Integrity of the monolayer was tested by TEER measurements with Millicell-ERS electrical resistance measuring system (Millipore, Bedford, MA, USA) with optimum resistance reading in the range of 150–200 Ω cm2 representing a monolayer with tight junctions.

To test transport of anti-diabetic macromolecules (insulin or exenatide), a 2 mm device loaded with 0.3 mg bovine insulin; or 3 µg exenatide, was placed on confluent monolayer in 24 well plates and was incubated for 5 h with gentle shaking at 37 °C. At predetermined time points, 100 µL samples were withdrawn from the basolateral chamber to quantify the total amount of insulin/exenatide transported across the monolayer, with immediate replenishment with fresh media. Monolayer integrity was also determined at same time points by TEER measurements. The withdrawn samples were quantified for bovine insulin or exenatide content using ELISA kits mentioned in the “Materials” Section, following manufacturer’s protocol. Caco-2 monolayers exposed to plain drug solutions of equivalent concentrations served as controls for the studies.

In-vivo Efficacy of Devices in enhancing Intestinal Absorption of Insulin (Bovine and Human) and Exenatide

In-vivo efficacy of mucoadhesive devices in delivery of anti-diabetic therapeutics was tested in adult male Sprague–Dawley (SD) rats (275–300 g; Charles River Laboratories, Wilmington, MA, USA). All animal studies were performed under aseptic conditions in accordance with protocols approved by University of California Santa Barbara (UCSB) Institutional Animal Care and Use Committee (IACUC). In-vivo studies were performed following a previously established protocol of direct jejunal placement of drug-loaded mucoadhesive devices, in our lab.20 Briefly, the animals were fasted for 7 h before the experiments and were anesthetized by 1.5–3.0% isoflurane before the surgery. To directly place the devices into the jejunal area of small intestine, small intestine was exposed by a midline abdominal incision of 2.5–3.0 cm, following which a small longitudinal incision (0.2–0.3 cm) was made about 10–15 cm from the proximal end of the small intestine, so as to confirm device placement in the jejunum. Three drug-loaded devices [(3 mm diameter; 0.04 or 0.2 mg insulin/device (bovine or human), or 0.1 mg exenatide/device] were inserted through the opening into the intestinal lumen, so as to provide approximately 10 or 50 U/kg dose for bovine/human insulin, or 1 mg/kg dose for exenatide.

For determining efficacy of rectangular devices, the same surgical procedure was followed, but instead of placing three 3 mm circular devices, a single 2 mm × 13 mm device containing 0.6 mg human insulin was inserted through the opening into the intestinal lumen, so as to provide a dose of 50 U/kg of human insulin.

Following insertion, the devices were pushed approximately 5 cm into the intestinal lumen with a sterile metal plunger so as to avoid contact with incision site. Following device placement, the intestinal incision was closed with surgical tissue glue, Vetbond® (3 M Corporation, St. Paul, MN, USA). 15 min after device insertion, the adhesion site was flushed with bolus injection of 0.5 mL sterile saline, and midline muscular incisions were closed with sterile 3–0 vicryl surgical sutures followed by sealing the skin incision with Vetbond® tissue glue.

Following device placement, blood samples were collected from the tail vein at pre-determined time intervals up to 5 h in blood collection tubes without any anticoagulant (insulin) or Li-heparin coated tubes (exenatide), and placed on ice. Blood serum/plasma was separated by centrifuging the blood samples at 5000 rpm for 10 min and stored at −20 °C until analysis. At each blood collection time point, blood glucose levels were also measured using AccuChek® Aviva® monitoring system. In addition to the test groups, control groups for each macromolecule (i) 1 U/kg subcutaneous (SC) injection (ii) blank devices and (iii) 10 U/kg intestinal injection were run for bovine insulin were also tested. Similarly, 3 control groups (i) 20 µg/kg SC injection (ii) blank devices and (iii) 1 mg/kg intestinal injection were run for exenatide so as to compare efficacy of mucoadhesive devices with conventional oral/systemic delivery.

In-vivo Efficacy of Permeation Enhancer in Providing Enhanced Intestinal Absorption of Human Insulin

In some experiments, performance of the device was tested in presence of Dimethyl palmitoyl ammonio propanesulfonate (PPS), a permeation enhancer recently reported by our lab for efficacy in enhancing oral absorption of salmon calcitonin (sCT).19,47 For in vivo studies, surgical protocol was the similar to what has been described earlier, but before inserting the rectangular device into jejunum, a 5 cm stretch of jejunum was irrigated with 100 µL of 0.5% w/v aqueous solution of PPS (or with sterile saline for administration of devices alone) for 15 min by ligating the two ends of jejunal portion with hemostats. After 15 min, hemostats were removed and the rectangular devices were administered, blood glucose levels were measured, and serum samples were collected for human insulin quantification using ELISA.

Analysis of Pharmacokinetic and Pharmacological Efficacy

Pharmacokinetic profile of oral absorption of bovine and human insulin; and exenatide from mucoadhesive devices was established by quantifying serum/plasma concentrations of respective drugs at different time-points up to 5 h using ELISA kits [Bovine and human insulin from Mercodia, Inc. (Winston Salem, NC, USA); and extraction-free ELISA kit for exenatide from Bachem Americas, Inc. (Torrance, CA, USA)] following manufacturer’s protocols. To establish pharmacological availability (PA) of bovine and human insulin released from the devices, reduction in blood glucose concentration (mg/dl) was quantified using a commercially available Accuchek® Aviva® blood glucose monitoring system using the blood from tail vein bleeding, and was plotted as % reduction in blood glucose with a 100% baseline with pre-treatment blood glucose concentration (mg/dL). Pharmacokinetic parameters were calculated from concentration–time curves for human and bovine insulin, exenatide, and blood glucose reduction using Microsoft Excel (Microsoft Inc., Redmond, WA). Data were analyzed for Cmax, Tmax, t1/2, area under concentration–time plot (AUC0-5h), and relative bioavailability utilizing Microsoft Excel Extensions for Pharmacokinetic Analyzes provided by Usansky et al.45

Statistical Analyzes

Data analysis was performed by using Student’s t test and/or one-way ANOVA followed by appropriate post hoc analysis (GraphPad Prism 5.0, GraphPad Software, La Jolla, CA). Values of p < 0.05 were considered statistically significant (mean ± SE unless otherwise mentioned).

Results

The current study was designed to evaluate the efficacy of mucoadhesive devices in providing sustained delivery of anti-diabetic therapeutics across the intestine. This study builds upon previous reports from our laboratory.20,49 We used mucoadhesive devices synthesized with a blend of 3 polymers used in FDA-approved products; carbopol, pectin, and SCMC, decorated with a flexible backing layer of ethyl cellulose (EC) on all but one sides to prevent unwarranted drug leakage and also to ensure unidirectional drug release through intestinal epithelium. Most devices used in this study were 2–5 mm in diameter with ≈400 µm thickness, and carried varying amount of drugs as described in “Materials and Methods” Section.

The mucoadhesive devices possess a distinct morphology characterized by a uniform appearance with no distinguishable difference between top and bottom surfaces except the presence of invisible EC backing layer on top surface.20 These devices have been known to adhere to the intestinal luminal surface by the adhesive side with a force in the range of 8–20 mN (corresponding to 50–100 times devices’ own weight), depending on the experimental conditions. Strong mucoadhesion is necessary for the success of such a delivery system, which faces the challenge of dislodgement by several factors including its own weight, intestinal peristaltic movement, and mucus flow.40,49 In this study, the mucoadhesive devices were loaded with either insulin (bovine or human) or exenatide (a GLP-1 agonist), and were tested for drug release, in vitro Caco-2 monolayer transport and in vivo enhancement of intestinal absorption of anti-diabetic therapeutics.

In vitro Drug Release Studies

Mucoadhesive devices provided a near-complete release of loaded FITC-insulin irrespective of the amount of peptide loading in 5 h (Fig. 1a) and the release kinetics was independent of the dose which is evident from overlapping released curves (Fig. 1b). Devices loaded with bovine insulin showed similar results with approximately 90% (269.4 ± 50.5 µg out of 300 µg loaded) of bovine insulin coming out at the end of 5 h (Fig. 2a). Exenatide-loaded devices demonstrated approximately 47% (7.5 ± 0.1 µg out of 16 µg loaded) drug release in 5 h (Fig. 2b). Further analysis of release-time curve revealed a zero order drug release kinetics with r2 ≈ 0.96 (insulin) and ≈0.95 (exenatide) with release rate constants of 53.9 ± 10.1 µg/mL/h (insulin) and 1.5 ± 0.03 µg/mL/h (exenatide).
Figure 1

Effect of Amount of Drug Loading on In-vitro Release from Devices. FITC-insulin release from mucoadhesive devices following loading of varying amounts in vitro. FITC-insulin was loaded onto the devices at 3% (closed circle), 10% (closed square), or 30% (closed triangle) w/w ratio of the device weight and drug release was measured at room temperature in PBS (pH 7.4) with gentle shaking for up to 5 h. (a) Cumulative release of FITC-insulin from the devices (mg); and (b) Cumulative insulin fraction released at various drug loading (fraction of maximum). Data represent mean ± SD (n = 3).

Figure 2

Cumulative in-vitro dissolution and release studies from bovine insulin (A) and exenatide (B) loaded devices. Drug release from the devices was measured in vitro for therapeutic peptides loaded. Bovine insulin was loaded into mucoadhesive devices at 300 µg/device, and exenatide was loaded at 16 µg/device. Release of both bovine insulin and exenatide was measured in PBS (pH 7.4) at room temperature with gentle shaking for 5 h. Quantification of both peptides were performed using commercially available ELISA kits as described in “Materials and Methods” Section. (a) Cumulative release (µg) of bovine insulin; and (b) exenatide from mucoadhesive devices. Data represent mean ± SE (n = 3–6).

In vitro Drug Transport Across Caco-2 Monolayers

The ability of mucoadhesive devices to deliver insulin and exenatide was assessed in vitro using 3-day Caco-2 monolayers.18 Application of drug-loaded devices resulted in significant enhancement (approximately twofold) of peptide transport across Caco-2 monolayers following 5-h incubation (Fig. 3a for insulin and Fig. 4a for exenatide). For bovine insulin, mucoadhesive devices (0.3 mg insulin loading) provided a cumulative insulin transport of 17.7 ± 4.9 µg in 5 h as compared to 8.9 ± 1.7 µg with loading of same dose of bovine insulin as solution (Fig. 3a). Similarly, exenatide-loaded devices (3 µg exenatide loading) showed cumulative transport of 127.3 ± 15.0 ng as compared to 63.8 ± 19.5 ng with loading of plain drug (Fig. 4a). Considering the fact that the devices loaded onto the Caco-2 monolayers were only 2 mm in diameter, we expect that the enhancement of flux (delivered amount per unit area) offered by devices over solution is even higher that the twofold enhancement seen in the total delivered dose.
Figure 3

Enhancement of bovine insulin permeability across Caco-2 monolayers by mucoadhesive devices. Bovine insulin transport across Caco-2 intestinal epithelial cell monolayers was tested using short-term 3-day Caco-2 system at 300 µg loading concentration. Bovine insulin was loaded as either plain solution (300 µg/well; closed circles); or as a 2 mm device (300 µg/device; open triangles) on confluent Caco-2 monolayers and apical-to-basolateral permeability was analyzed from the samples collected from basolateral chambers for 5 h using commercially available ELISA kits. TEER values were also measured to account for monolayer integrity during the experiment. (a) Cumulative insulin transport to basolateral chamber (µg) at different time-points, (b) % reduction in TEER values of Caco-2 monolayers following exposure to bovine insulin. Data represent mean ± SE (n = 3) of three individual experiments.

Figure 4

Enhancement of exenatide permeability across Caco-2 monolayers by mucoadhesive devices. Exenatide transport across Caco-2 intestinal epithelial cell monolayers was tested using short-term 3-day Caco-2 system at 3 µg loading concentration. Exenatide was loaded as either plain solution (3 µg/well; closed circles); or as a 2 mm device (3 µg/device; open triangles) on confluent Caco-2 monolayers and apical-to-basolateral permeability was determined by analyzing the samples collected from basolateral chambers for 5 h using commercially available ELISA kits. TEER values were also measured to account for monolayer integrity during the experiment. (a) Cumulative exenatide transport to basolateral chamber (ng) at different time-points, (b) % reduction in TEER values of Caco-2 monolayers following exposure to exenatide. Data represent mean ± SE (n = 3) of three individual experiments.

Application of mucoadhesive devices did not exhibit significant and destructive changes in TEER values of Caco-2 monolayers. TEER values for device incubated monolayers remained in close conjunction of TEER values of control monolayers with drug solution application (122.5 ± 9.6% of the initial values as compared to 104.0 ± 15.0% with direct drug solution loading for insulin; and 124.0 ± 9.0% of the initial values as compared to 114.2 ± 3.0% with direct drug solution loading for exenatide) (Figs. 3b, 4b). These data suggest that the mucoadhesive devices did not compromise monolayer integrity.

In vivo Efficacy of Mucoadhesive Devices in Enhancing Intestinal Absorption of Macromolecules

The efficacy of mucoadhesive devices in promoting intestinal absorption of macromolecules was tested in adult male SD rats following intrajejunal administration of the devices. A direct intestinal injection of insulin (10 U/kg) did not produce any noticeable reduction in blood glucose levels (Fig. 5a). However, intrajejunal placement of bovine insulin loaded mucoadhesive devices led to significant decrease in blood glucose down to 75.3 ± 3.8% of initial value with 10 U/kg and 59.9 ± 1.1% with 50 U/kg loading (Fig. 5a). To confirm that the delivery is not specific to bovine insulin, patches loaded with 50 U/kg human insulin were also tested and led to similar trends (Fig. 6a). Mucoadhesive devices led to sustained hypoglycemic effect compared to rapid and reversible effects of the subcutaneous injection. Additional studies are necessary to fully understand the efficacy of mucoadhesive devices in delivering human insulin over simple administration as a solution. However, data in Fig. 5 indicate that simple intestinal administration of bovine insulin does not provide significant uptake and the same conclusion is expected for human insulin. Mucoadhesive devices also demonstrated significant pharmacological bioavailability compared to negligible bioavailability for intestinal insulin injection (Table 1).
Figure 5

In-vivo efficacy of mucoadhesive devices in enhancing pharmacokinetic absorption and therapeutic benefits of bovine insulin by oral route. Efficacy of mucoadhesive devices in enhancing in vivo absorption of bovine insulin in a rat model. Bovine insulin loaded devices or plain drug solution was placed in jejunal region of rats’ small intestine by surgical procedures followed by blood sample collection up to 5 h. Reduction in blood glucose concentration was determined in real time, while serum insulin concentration was determined by ELISA as described in “Materials and Methods” Section. Administration of plain insulin solution (10 U/kg; closed diamonds), bovine insulin-loaded mucoadhesive devices (10 U/kg; open triangles), bovine insulin-loaded mucoadhesive devices (50 U/kg; closed squares), subcutaneous insulin injection (1 U/kg; open circles), and blank devices (closed circles) were performed. (a) Pharmacological efficacy of insulin-loaded devices in reducing blood glucose concentration. Data are plotted as % reduction in blood glucose as compared to baseline glucose levels. (b) Concentration–time plot of bovine insulin-loaded devices following intra-jejunal placement depicting serum insulin concentration (ng/mL). Data represent mean ± SE (n = 4–6 animals/group).

Figure 6

In-vivo efficacy of mucoadhesive devices in enhancing pharmacokinetic absorption and therapeutic benefits of human insulin by oral route. Efficacy of mucoadhesive devices in enhancing in vivo absorption of human insulin in a rat model. Human insulin loaded devices or plain drug solution was placed in jejunal region of rats’ small intestine by surgical procedures followed by blood sample collection up to 5 h. Reduction in blood glucose concentration was determined in real time, while serum insulin concentration was determined by ELISA as described in “Materials and Methods” Section. Administration of human insulin-loaded mucoadhesive devices (50 U/kg; open triangles), bovine insulin-loaded mucoadhesive devices (50 U/kg; open circles), and subcutaneous insulin injection (1 U/kg; closed circles) were performed. (a) Pharmacological efficacy of insulin-loaded devices in reducing blood glucose concentration. Data are plotted as % reduction in blood glucose as compared to baseline glucose levels. (b) Concentration–time plot of human insulin-loaded devices following intra-jejunal placement depicting serum insulin concentration (ng/mL). Data represent mean ± SE (n = 4–6 animals/group).

Table 1

Pharmacokinetic parameters of mucoadhesive devices loaded with therapeutic macromolecules following intra-jejunal administration. Data represent mean ± standard error (n = 4–6).

Formulation

Cmax (ng/mL)

T1/2 (h)

AUC0–5 (ng/mL/h)

% relative bioavailability

% pharmacological bioavailability

Bovine insulin

 Subcutaneous Injection (1 U/kg)

2.7 ± 0.4

1.7 ± 0.4

4.8 ± 0.5

 Intestinal injection (10 U/kg)

0.22 ± 0.03

Non-quantifiable

0.6 ± 0.03

0.25 ± 0.01

0.00

 Devices (10 U/kg)

0.4 ± 0.08

1.5 ± 0.4

1.6 ± 0.5

3.3 ± 0.9

6.5 ± 1.5

 Devices (50 U/kg)

2.4 ± 0.4

3.6 ± 1.0

7.9 ± 0.7

3.3 ± 0.3

1.8 ± 0.25

Human insulin

 Subcutaneous injection (1 U/kg)

3.5 ± 1.2

0.3 ± 0.05

2.8 ± 0.7

 Round devices (50 U/kg)

1.4 ± 0.5

1.0 ± 0.3

1.7 ± 0.6

1.2 ± 0.4

2.3 ± 0.3

 Rectangular Devices (50 U/kg)

1.5 ± 0.5

1.1 ± 0.2

5.5 ± 1.6

3.9 ± 1.1

2.9 ± 0.2

 Rectangular devices (50 U/kg) + 0.5% w/v PPS

2.8 ± 1.5

1.6 ± 0.2

10.8 ± 4.6

7.7 ± 3.3

3.3 ± 0.4

Exenatide

 Subcutaneous injection (20 µg/kg)

6.0 ± 0.6

3.5 ± 0.5

19.2 ± 1.4

 Intestinal injection (1 mg/kg)

Non-quantifiable

Non-quantifiable

Non-quantifiable

Negligible

 Devices (1 mg/kg)

1.7 ± 0.4

7.5 ± 1.2

7.8 ± 1.6

0.8 ± 0.2

Quantification using ELISA indicated presence of significant amounts of insulin and exenatide in blood following intrajejunal placement of the devices. Bovine insulin-loaded devices resulted in a Cmax of 0.4 ± 0.1 ng/mL (10 U/kg) and 1.4 ± 0.3 ng/mL (50 U/kg) compared to ~ 0.1 ± 0.03 ng/mL with intestinal injection of drug solution (10 U/kg) (Fig. 5b). Similarly, human insulin loaded devices resulted in a Cmax of 1.4 ± 0.5 ng/mL (50 U/kg) (Fig. 6b) and exenatide-loaded devices showed a Cmax of 1.0 ± 0.4 ng/mL (1 mg/kg) compared to no detectable concentrations with intestinal injection (1 mg/kg) (Fig. 7).
Figure 7

In-vivo efficacy of mucoadhesive devices in enhancing pharmacokinetic absorption of exenatide following intestinal administration. Efficacy of mucoadhesive devices in enhancing in vivo absorption of exenatide in a rat model. Exenatide-loaded devices or plain drug solution was placed in jejunal region of rats’ small intestine by surgical procedures followed by blood sample collection up to 5 h. Plasma exenatide concentration was determined by ELISA as described in “Materials and Methods” Section. Administration of exenatide-loaded mucoadhesive devices (1 mg/kg; open triangles), plain drug solution (1 mg/kg; closed squares), subcutaneous exenatide injection (20 µg/kg; open circles), and blank devices (closed circles) were performed. Concentration–time plot of exenatide-loaded devices following intra-jejunal placement depicting plasma exenatide concentration (ng/mL). Data represent mean ± SE (n = 4–6 animals/group).

Effect of Device Geometry and Permeation Enhancers on Intestinal Absorption of Human Insulin

Human insulin-loaded rectangular devices demonstrated better blood glucose reduction and serum insulin concentration compared to circular devices (Figs. 8a and 8b) although the blood glucose reduction after 5 h was similar for both round and rectangular devices. The % pharmacological bioavailability for rectangular devices was higher (2.9 ± 0.2%) compared to circular devices (2.3 ± 0.3%) (Table 1). At the same time, significantly higher serum insulin concentrations were detected for rectangular devices (Fig. 8b). While both round and rectangular devices showed similar Cmax, rectangular devices demonstrated consistently constant serum insulin levels resulting in a better relative bioavailability (Table 1).
Figure 8

In-vivo efficacy of rectangular mucoadhesive devices in enhancing pharmacokinetic absorption and therapeutic benefits of human insulin by oral route. Efficacy of device geometry in enhancing in vivo absorption of human insulin in a rat model. Human insulin loaded round (3 devices) or rectangular devices (1 device) were placed in jejunal region of rats’ small intestine by surgical procedures followed by blood sample collection up to 5 h. Reduction in blood glucose concentration was determined in real time, while serum insulin concentration was determined by ELISA as described in “Materials and Methods” Section. Administration of human insulin-loaded round mucoadhesive devices (50 U/kg; closed circles), and rectangular devices (50 U/kg; open circles) were performed. (a) Pharmacological efficacy of insulin-loaded devices in reducing blood glucose concentrations. Data are plotted as % reduction in blood glucose as compared to baseline glucose levels. (b) Concentration–time plot of human insulin-loaded devices following intra-jejunal placement depicting serum insulin concentration (ng/mL). Data represent mean ± SE (n = 4–6 animals/group).

Efficacy of combining a permeation enhancer along with mucoadhesive devices was also studied by administration of PPS (0.5% w/v) and human insulin-loaded mucoadhesive devices. Inclusion of PPS during in vivo device administration resulted in higher % reduction in initial blood glucose levels (43.4 ± 0.1% as compared to 55.9 ± 4.7% for rectangular devices alone) (Fig. 9a). The % pharmacological bioavailability also increased to 3.3 ± 0.4% compared to 2.9 ± 0.2% for rectangular devices alone. Similarly, serum insulin concentrations also increased significantly with PPS administration (Fig. 9b). Cmax of 2.8 ± 1.5 ng/mL was observed for devices + PPS as compared to 1.5 ± 0.5 ng/mL for rectangular devices alone (Fig. 9b; Table 1).
Figure 9

In-vivo efficacy of PPS in enhancing mucoadhesive device induced pharmacokinetic absorption and therapeutic benefits of human insulin. Efficacy of PPS, an oral permeation enhancer, in increasing human insulin absorption via oral route was tested by simultaneous administration of PPS (0.5% w/v aqueous solution) and rectangular mucoadhesive devices (50 U/kg human insulin) in a rat model. A 5 cm jejunal segment was irrigated with PPS solution for 15 min before placing the mucoadhesive devices, and rectangular devices were placed as described earlier followed by blood sample collection up to 5 h. Reduction in blood glucose concentration was determined in real time, while serum insulin concentration was determined by ELISA as described in “Materials and Methods” Section. Administration of human insulin-loaded rectangular devices (50 U/kg; open circles), and human insulin-loaded rectangular devices + PPS (50 U/kg + 0.5% w/v; open triangles) were performed. (a) Pharmacological efficacy of human insulin-loaded devices in reducing blood glucose concentration. Data are plotted as % reduction in blood glucose as compared to baseline glucose levels. (b) Concentration–time plot of human insulin-loaded devices following intra-jejunal placement depicting serum insulin concentration (ng/mL). Data represent mean ± SE (n = 4–6 animals/group).

Pharmacokinetic Analysis of In-vivo Data

Pharmacokinetic analysis revealed that using mucoadhesive devices for oral delivery of therapeutic peptides resulted in significant enhancement of their bioavailability. Bovine insulin-loaded devices showed a % relative bioavailability of 3.3 ± 0.9% (10 U/kg) and 3.3 ± 0.3% (50 U/kg) as compared to 0.2 ± 0.01% for drug solution. Similarly, human insulin-loaded devices showed a % relative bioavailability of 1.2 ± 0.4% (50 U/kg) and exenatide-loaded devices showed 0.8 ± 0.2% (1 mg/kg) compared to negligible oral bioavailability of the drug solution, respectively (Table 1). Rectangular devices showed significantly better % bioavailability of 3.9 ± 1.1% as compared to 1.2 ± 0.4% for round human insulin-loaded devices. Incorporation of PPS further increased the % bioavailability to 7.7 ± 3.3% (Table 1). In addition, pharmacokinetic data also revealed that encapsulation of therapeutic macromolecules also resulted in substantial increases in the elimination half-lives of the peptides. The t1/2 of the peptides increased to 3.6 ± 1.0 h (bovine insulin), 1.0 ± 0.3 h (human insulin), and 7.5 ± 1.2 h (exenatide) as compared to 1.7 ± 0.4 h, 0.3 ± 0.05 h, and 3.5 ± 0.5 h respectively for drugs injected by the subcutaneous route (Table 1). While rectangular devices increased the bioavailability, no effect was seen on human insulin t1/2 as compared to round devices (1.1 ± 0.2 h as compared to 1.0 ± 0.3 h). At the same time, PPS inclusion increased the human insulin half-life to 1.6 ± 0.2 h (Table 1). This increase in the half-lives is consistent with the observation made in evaluation of blood glucose data regarding the capability of mucoadhesive devices for providing sustained drug release over a prolonged period of time.

Discussion

Here we investigated the efficacy of oral mucoadhesive devices in providing a non-invasive alternative to injections for management of diabetes. This study builds upon earlier work underlining the efficacy of mucoadhesive devices in providing significant enhancement of oral absorption of insulin48 and salmon calcitonin (sCT).20 Delivery of three anti-diabetic molecules i.e. bovine insulin, human insulin, and exenatide (exendin-4 analog) was tested. While being highly sought after route for drug delivery, oral administration has proved challenging for delivering peptides and proteins. One of the prime reasons for the same is their negligible bioavailability via oral route due to rapid degradation by gastric enzymes and acidic environment, and poor permeation across tight junction intestinal epithelium.36

Mucoadhesive devices reported here released the loaded drug in a time-dependent fashion, with nearly complete release of insulin in 5 h, irrespective of the amount of drug loading (Figs. 1a, 1b). Later experiments determined the amount of active macromolecules (bovine insulin and exenatide) released by the mucoadhesive devices over 5 h by ELISA assay with approximately 90% (insulin) and 50% (exenatide) active drug releasing from the devices (Figs. 2a, 2b). The drug release from the devices followed a nearly zero-order profile, which suggests that the drug was uniformly encapsulated into the devices, a vital prerequisite for sustained time-dependent drug release. The ideal release profile depends on the intended therapeutic goal (basal vs. pre-meal insulin). Mucoadhesive devices are better suited for basal delivery and thus a steady release over a long period is desired. The release data shown in Fig. 2 is close to this profile. Eventually, it might be feasible to design devices with different release profiles and deliver them from the same capsule so that a more complex delivery profile can be accomplished. Presence of impermeable backing layer of ethyl cellulose protects the drug from release away from target site, and thus provides protection against proteolytic and enzymatic degradation.16,37 The drug release from mucoadhesive polymer network is primarily due to swelling in the intestinal lumen. This swelling in turn allows increase in pore size and aqueous content, thus pushing the hydrophilic peptides out of the polymeric network into dissolution media.33 While these data underline the efficacy of mucoadhesive devices in providing controlled delivery of therapeutic peptides, there may be concerns regarding the bioavailability of peptides due to presence of multiple carboxyl groups in peptide drugs that are known to interact with mucoadhesive polymers and mucin, the key component of mucus for mucoadhesive interactions.44 While these interactions are significant, they do not seem to hamper the drug release from the devices. At the same time, the flexible design of the devices makes it amenable to customization for specific drugs by varying the composition and physicochemical properties.

The ability of devices in delivering insulin and exenatide across the biological membrane was tested in vitro using a 3-day Caco-2 monolayer system.18 Caco-2 monolayers represent a practical model system to study drug permeation across intestinal epithelium.23 Transport of insulin and exenatide across the monolayer is likely to occur by paracellular route i.e. by passing through the intercellular space across cellular tight junctions,5 one of the main reasons for their intestinal impermeability is inability to cross the intestinal epithelium.47 Many delivery systems have been tested to enhance the drug transport across intestinal epithelium by modulating intercellular tight junctions but their efficacy is mostly hindered by the associated toxicity.46

The drug-loaded mucoadhesive devices were able to enhance overall transport of both bovine insulin and exenatide across Caco-2 monolayers by approximately twofold (Figs. 3b, 4b), without compromising monolayer integrity as indicated by TEER measurements (Figs. 3a, 4a). Ability of the devices in enhancing peptide transport without hampering intercellular tight junctions underlines their exemplary safety. The major mechanisms of enhanced peptide transport by mucoadhesive devices is believed to be ‘Concentration Gradient Dependent Diffusion’ across the membrane, which is a well-known mechanism for drug transport across the skin by transdermal patches.35 One of the major problems with current oral peptide systems is dose dilution across the intestinal lumen resulting in loss of drug into luminal fluids.34 Mucoadhesive devices, on the other hand, offer a high drug concentration reservoir close to the intestinal epithelium, thus providing a high concentration gradient across the intestinal epithelium for transport.

Mucoadhesive devices delivered significant doses of drugs into systemic circulation in vivo as measured by % reduction in real-time blood glucose concentrations. Intestinal administration of peptide-loaded devices resulted in increased intestinal transport for all three peptides as evident by their Cmax as compared to negligible drug absorption following direct intestinal injection of equivalent doses (Table 1). At the same time, a significant reduction in blood-glucose was observed with both bovine and human insulin following administration of drug-loaded devices (approximately 42% reduction as compared to negligible reduction with intestinal injection) (Figs. 5a, 6a). Pharmacological efficacy with exenatide-loaded devices was not assessed since exenatide’s primary effect is on HbA1c (not quantifiable following acute single dose administration). Administration of drug-loaded mucoadhesive devices significantly enhanced both pharmacokinetic (≈13-fold for insulin and ≈80-fold for exenatide), and pharmacological bioavailability (as compared to intestinal injections of the drug solutions) (Table 1). Another important point to note is that encapsulation of therapeutic peptides also significantly enhanced their elimination half-lives (t1/2) by several-fold compared to subcutaneous injections (Figs. 5a, 6a; Table 1). Capability of mucoadhesive devices in providing controlled release of the anti-diabetic peptides is of paramount importance as this would solve the problems of multiple administrations, dose overshooting, postprandial hyperglycemia, and life-threatening hypoglycemia by providing consistent insulin concentrations throughout the day, and will provide for better disease management.4

Intestinal transport and absorption of therapeutics largely depends on the intestinal epithelium exposed to the delivery system, and on the amount of the drug being lost from the devices into the intestinal lumen. While the mucoadhesive devices work by creating a high concentration drug reservoir on a small area of intestinal epithelium, some drug may be lost form the edge of the devices. We hypothesized that this drug loss can be minimized by using a single device with reduced edge length. To test this hypothesis, we tested the administration of single rectangular devices carrying the same amount of drug/polymer instead of 3 round devices administered in vivo. Rectangular devices significantly enhanced the drug absorption resulting in greater reduction in blood glucose levels (Fig. 8a) and increased serum insulin concentrations (Fig. 8b).

To further enhance transport of insulin across intestinal epithelium, we tested the compatibility of rectangular devices together with PPS, a permeation enhancer reported by our lab, which is capable of enhancing macromolecule intestinal transport by both paracellular and transcellular routes.19 Use of PPS significantly reduced blood glucose levels by as much as 57% of the initial values (Fig. 9a), thus accomplishing a % bioavailability of 7.7 ± 3.3% (Fig. 9b; Table 1) which is approximately 6.5 fold higher than the round devices carrying the same amount of drug.

A common observation with oral delivery and any alternate route of drug delivery (transdermal, nasal etc.) has been the usage of elevated drug dosage for therapeutic efficacy. While oral delivery offers lower bioavailability compared to subcutaneous injections, it still offers advantages of painless delivery, ease of use and enhanced patient compliance. Diabetes being a chronic disorder, non-invasive delivery of anti-diabetic peptides remains a widely investigated area of research. Although a multitude of delivery systems and routes have been tested for efficacy in non-invasive diabetic therapy in recent years, there is strong interest following FDA approval of Mannkind Pharmaceuticals’ inhaled insulin Afrezza®.2 Oral route of delivery offers a compliant and physiologically relevant route for anti-diabetic therapy.17

Many research groups have been actively working on developing oral alternatives for delivery of both insulin and exenatide, which may include use of nano/microparticulate systems, chemical modifications, permeation enhancers, and techniques for increasing gastrointestinal retention times. In a recent study published by Zhu et al., efficacy of PLGA nanoparticles was tested in enhancing oral bioavailability of insulin. Encapsulation of insulin in PLGA nanoparticles conjugated with a variety of cell penetrating peptides enhanced its bioavailability by a modest 1.71 times.53 Efficacy of chitosan conjugated nanoparticles in enhancing oral insulin bioavailability was tested by Sheng et al. which showed ≈twofold enhancement in pharmacological bioavailability of insulin.38 Many other groups have investigated unique nanoparticle-based approaches to enhance oral insulin bioavailability.10,13,21,29,41 Another unique approach being tested for delivery of insulin and other macromolecules is the mucus permeating nanoparticle delivery systems as discussed by Xu et al. and Karamanidou et al.26,50 Permeation enhancers and cell penetrating peptides have also been used to enhance oral insulin delivery. Nielsen et al. tested the efficacy of penetratin, a cell penetrating peptide, in enhancing oral insulin absorption with significant effects on blood glucose reduction and half-life enhancement.32 In another unique approach, Chuang et al. tested the efficacy of novel CO2 bubbles as oral insulin carriers loaded with insulin and DTPA/SDS, an absorption enhancer.11 These bubble carriers, when transported to the intestine using a hard gelatin capsule, provided significant bioavailability.

Exenatide being a newer drug, a limited number of studies have investigated the efficacy of oral route for its delivery. In one of the few studies, Ahn et al. tested the efficacy of chitosan-conjugated exenatide via oral route.1 This study reported a 6.4% relative bioavailability by oral route with significant hypoglycemic effects. In another study, Zhang et al. tested the efficacy of pH-sensitive alginate and hyaluronate cross-linked microspheres in oral exenatide delivery.52 In another recent study, Chuang et al. reported efficacy of co-administration of insulin and exenatide by means of chitosan nanoparticles in diabetic rats.12 Co-administration of insulin and exenatide resulted in superior therapeutic benefits as compared to single drug administration, which may be beneficial in patients with uncontrolled glucose levels.

Mucoadhesive delivery systems have long been studied for delivery of various small molecules and therapeutic peptides by multiple non-invasive routes including transdermal, buccal and oral routes.6,30,42 More recently, mucoadhesive films and patches have been investigated for oral delivery of peptides.7 The mucoadhesive devices tested in this project provide a unique perspective toward oral delivery of therapeutic peptides. These devices enhance drug absorption across the intestine by creating a localized high drug concentration reservoir thus enabling concentration dependent intercellular diffusion (Fig. 10). While the size of the mucoadhesive devices prevents dose dilution in intestinal luminal fluids, the impermeable backing layer protects the encapsulated peptides from off-target release and proteolytic degradation. Other major advantages of the devices include the design flexibility, customization, and ability to carry high drug payload, which is of pivotal importance given the negligible oral bioavailability of anti-diabetic peptides. Mucoadhesive devices discussed in this manuscript are capable of being encapsulated in enteric-coated hard gelatin capsules and subsequent release into small intestine without getting destroyed in the stomach, as established in our recently published study.20
Figure 10

Schematic representation of mechanism of adhesion, drug release and absorption across intestinal epithelium from mucoadhesive devices.

While this study demonstrates feasibility of mucoadhesive devices in oral delivery of anti-diabetic peptides, further studies are required to test their long-term efficacy in preclinical models. Future studies should focus on testing optimal device configurations in diabetic animals. Further studies may also explore (i) the likelihood of co-administration of multiple anti-diabetic therapeutics to enhance patient compliance, and (ii) customization of the devices to enhance therapeutic peptide delivery by various techniques including chemical conjugation of therapeutics, simultaneous loading of PPS. Use of chemical permeation enhancers such as PPS is particularly interesting as the mucoadhesive devices may facilitate the use of novel permeation enhancers at lower concentrations, given their proven ability to enhance peptide absorption across the intestine.

Conclusions

This study underlines the efficacy of mucoadhesive devices fabricated with a unique blend of mucoadhesive polymers, in providing enhanced oral bioavailability of anti-diabetic therapeutic macromolecules, which has been a major hurdle in promoting dose adherence and compliance among both Type-1 and Type-2 diabetes patients. The devices tested in this study not only increased systemic drug absorption via oral route, but also demonstrated significant pharmacological effects on real-time blood glucose measurements. Based on the principles of transdermal drug delivery, these mucoadhesive devices provide localized drug reservoir close to the intestinal epithelium, which in turn promotes concentration dependent drug diffusion into intestinal microcirculation while preventing luminal dose dilution, and proteolytic/enzymatic degradation. These mucoadhesive devices not only have the potential to provide prolonged glycemic control by controlled insulin/exenatide release, but are also capable of being customized according to the drug/patient requirements thus promoting personalized medicine, patient compliance and quality of life.

Acknowledgements

This research was funded by Otis Williams Fellowship from Santa Barbara Foundation (VG) and Enlight Biosciences. SM is a scientific advisor and shareholder of Entrega Inc.

Funding information

Funder NameGrant NumberFunding Note
Enlight Biosciences
    Otis Williams Fellowship

      Copyright information

      © Biomedical Engineering Society 2016

      Authors and Affiliations

      • Vivek Gupta
        • 1
        • 2
      • Byeong-Hee Hwang
        • 1
      • Nishit Doshi
        • 1
      • Amrita Banerjee
        • 1
      • Aaron C. Anselmo
        • 1
      • Samir Mitragotri
        • 1
      1. 1.Department of Chemical EngineeringUniversity of CaliforniaSanta BarbaraUSA
      2. 2.School of PharmacyKeck Graduate InstituteClaremontUSA

      Personalised recommendations