In vivo oral insulin delivery via covalent organic frameworks

With diabetes being the 7th leading cause of death worldwide, overcoming issues limiting the oral administration of insulin is of global significance. The development of imine-linked-covalent organic framework (nCOF) nanoparticles for oral insulin delivery to overcome these delivery barriers is herein reported. A gastro-resistant nCOF was prepared from layered nanosheets with insulin loaded between the nanosheet layers. The insulin-loaded nCOF exhibited insulin protection in digestive fluids in vitro as well as glucose-responsive release, and this hyperglycemia-induced release was confirmed in vivo in diabetic rats without noticeable toxic effects. This is strong evidence that nCOF-based oral insulin delivery systems could replace traditional subcutaneous injections easing insulin therapy.


Reagents and techniques
All reagents and starting materials were purchased from Sigma-Aldrich and used without further purification. The precursor, 2,6-diformylpyridine (DFP) was synthesized according to the published procedure with no modifications. Deionized water was used from Millipore Gradient Milli-Q water purification system. Thin-layer chromatography (TLC) was performed on silica gel 60 F254 (E. Merck). The plates were inspected under the UV light. Column chromatography was performed on silica gel 60F (Merck 9385, 0.040-0.063 mm). Infrared spectra were recorded on an Agilent Technologies Cary 600 Series FTIR Spectrometer using the ATR mode. PXRD patterns of the samples were recorded by using an X-ray Panalytical Empyrean diffractometer. High resolution transmission electron microscopy (HRTEM) images were obtained using a Talos F200X Scanning/Transmission Electron Microscope (STEM) with a lattice-fringe resolution of 0.14 nm at an accelerating voltage of 200 kV equipped with CETA 16M camera. The high resolution images of periodic structures were analyzed using TIA software. N 2 adsorption-desorption isotherms were obtained at 77 K using Micrometrics ASAP 2020 surface area analyzer. The topography of the self-templated samples was analyzed by dynamic atomic force microscopy (5500 Atomic Force Microscope; Keysight Technologies Inc., Santa Rosa, CA). We acquired topography, phase and amplitude scans simultaneously. Silicon cantilevers (NanosensorsTM, Neuchatel, Switzerland) with resonant frequencies of 250-300 kHz and force constants of 100-130 Nm −1 were used. The set point value was kept at 2.5V. AFM scans were collected at 1024 points/lines with scan speed of 0.20 at fixed scan angle of 0 o . Scan artifacts were minimized by acquiring a typical scan at an angle of 90o under identical image acquisition parameters. We used GwyddionTM free soſtware (version 2.47), an SPM data visualization and analysis tool for postprocessing the AFM scans. Emission spectra in water at room temperature were recorded on a Perkin Elmer LS55 Fluorescence Spectrometer. Dynamic light scattering (DLS) measurements were performed on a Malvern Zetasizer NanoSeries to obtain the size and -potential of the nanoparticles. The XPS experiments were carried out on a Kratos Axis Ultra DLD spectrometer under a base pressure of ∼ 2×10 −10 mbar. A monochromated Al Kα X-ray source (1486.69 eV) was used to irradiate samples at room temperature. Far-UV spectra were recorded between 200 and 280 nm on a Chirascan CD spectrometer (Applied Photophysics, UK) with the lamp supplied with a flow of nitrogen. Fifty microlitres of the solution were added to a 0.1 mm path-length quartz cuvette (Hellma, UK) and the measurements were carried out at 20 °C (1 nm bandwidth resolution and 1 s acquisition time). Typically, at least two scans were recorded, and baseline and HEPES spectra were subtracted from each spectrum. Data were processed using Applied Photophysics Chirascan Viewer and Microsoft Excel. Phase contrast and fluorescence images were observed on an Olympus FV1000MPE confocal scanning microscope. Flow cytometry analyses were performed on Accuri C6 Flow Cytometer. The most favorable location of insulin molecules between COF layers was calculated with a simulated annealing process, using the Adsorption Locator module of Biovia Materials Studio. For this, the TTA-DFP COF structure was first modified by separating sets of three layers at a 25 Å distance, to allow the incorporation of insulin molecules. The insulin monomer was obtained from the 1zni structure of the protein data bank. 1 One insulin monomer was incorporated per nCOF unit cell, which corresponds to ~ 70 wt%. The Monte Carlo simulation was then run with the use of a universal forcefield after charge assignment, and the conformation with lowest adsorption energy was selected.    The solution (pH 7.4) was stirred at room temperature overnight. The solution was then cleaned with water several times by centrifugation and washed with deionized H 2 O to remove unloaded glucose molecules.

High resolution transmission electron microscopy (HRTEM).
High resolution transmission electron microscopy (HRTEM) images were obtained using a Talos F200X Scanning/Transmission Electron Microscope with a lattice-fringe resolution of 0.14 nm at an accelerating voltage of 200 kV equipped with a CETA 16M camera. The samples were prepared on holey carbon film mounted on a copper grid. A drop of diluted particle solution was spotted on the grid and dried overnight at room temperature (298 K). The obtained images of periodic structures were analyzed using TIA software. All the relevant areas were marked using bright field imaging mode at spot size 3 and the marked areas were also scanned using the STEM-HDAAF mode at spot size 9 for imaging and spot size 6 for conducting the STEM-EDAX. The STEM mode helps in providing the elemental composition as it works on the principle of mass determination.
Such measurements can be performed at low electron dose by collecting the high-angle darkfield signal using an annular detector. This mode is generally used to image the elements that have different masses, with the heavier mass element appearing brighter. The samples were scanned at spot size 9 and with screen current of 60 pA. The data was analyzed using Velox analytical software.

Elemental Mapping
The chemical mapping was carried out in STEM-EDAX mode wherein the energy-dispersive X-ray analysis (EDAX) was carried out using a super-X EDS detector. The system has superior sensitivity with resolution of ≤ 136eV@Mn-Kα for 10kcps at zero-degree sample tilt. The detector provides quick data even for low intensity EDS signals. The data is the sum of 4 detectors and the collection time for the elemental maps in fast mapping mode can be reduced to minutes from hrs. The data 15 was analyzed using Velox analytical software. The samples for the HRTEM study were prepared on holey carbon film mounted on a copper grid.

Powder X-ray diffraction (PXRD) measurements
Powder X-ray diffraction (PXRD) measurements were carried out to confirm the crystalline nature of the framework. The TTA-DFP-nCOFs were found highly crystalline in nature. In fact, we observed a strong peak at 2θ of 4.9 ° assigned to the (110) plane of the regularly ordered lattice.
TTA-DFP-nCOF shows a broad peak at ~24.80 corresponding to the reflection from the (003) plane. Before measurements, the TTA-DFP-nCOF (empty or loaded with Insulin) was activated at 358 K for 24 h to remove the solvent and trapped gas. Based on the IUPAC classification system, TTA-DFP-nCOF exhibited type-II isotherms, which are indicative of microporous materials. BET surface area was found to be 384.52 m 2 g -1 .
Figure S13. Nitrogen adsorption/desorption isotherms and pore size distribution curves (inset) at 77 K of TTA-DFP-nCOF before (yellow) and after loading with Insulin (green). The experiment was performed in triplicate.

Fourier Transform infrared (FTIR) spectroscopy
The TTA-DFP-nCOF formation as well as insulin loading was confirmed and characterized by ATR-IR spectroscopy using an Agilent Technologies Cary 600 Series FTIR spectrometer. The spectral data within the range of 4000 to 600 cm −1 were recorded, and 512 scans were averaged for each spectrum with a spectral resolution of 2 cm −1 . The spectrum of the background was recorded first and it was subtracted from the spectra of samples automatically.           At regular intervals, samples were withdrawn and the fluorescence intensity was measured.

Circular dichroism (CD) spectroscopy
To evaluate the changes of the activity and structure of insulin released from the nanoparticles, circular dichroism (CD) spectroscopy was performed as a common method to analyze the secondary structure of a protein with high reliability. In the CD spectra of the native insulin in HEPES (pH 7.4), there were two extrema at 208 and 222 nm, due to the α-helix structure and βstructure, respectively.
We evaluated the chemical stability of insulin loaded in TTA-DFP-nCOF exposed to the GI fluid simulations (pH 2.0, 24 hours). At acidic pH, insulin is not released from the nanoparticle, therefore in order to perform CD analysis we exposed TTA-DFP-nCOF/insulin to NaOH (0.1M) to release the protein from the NPs. As presented in Figure  native and GI fluid exposed insulin from TTA-DFP-nCOF were both 1.2 which reflected that there was no significant difference in the secondary structure between the native and GI fluid exposed The CD spectrum of the insulin after releasing for 12 hours from the TTA-DFP-nCOF/insulin under the hyperglycemic environment was similar to that of the native insulin with two extrema ( Figure   S30). The ratios between bands ([φ]208/[φ]222) for the native and released insulin were 1.25 and 1.24, respectively. Therefore, the secondary structure of the insulin released from the nanoparticles was similar to the original insulin. Accordingly, the released insulin maintained its structure and properties.

Glucose interaction
In order to study the release mechanism of insulin triggered by glucose, we incubated the TTA-DFP-nCOF 24 hours and 37 °C with i) insulin alone, ii) glucose alone (5 mg.mL -1 ) and iii) insulin followed by 24 hours with glucose (5 mg.mL -1 ). Samples were washed thoroughly and freezedried using lyophilization. Loading efficiency (wt%) was calculated using mass differences by comparing with the TTA-DFP-nCOF mass.

In vitro cell viability
Cell viability was assessed using CellTiter-Blue® Cell Viability assay (CTB, Promega). The assay measures the metabolic reduction of a non-fluorescent compound, resazurin, into a fluorescent product, resofurin, in living cells. As non-viable cells rapidly lose their metabolic activity, the amount of the resofurin product can be used to estimate the number of viable cells following treatment. Once produced, resofurin is released from living cells into the surrounding medium.
Thus, the fluorescence intensity of the medium is proportional to the number of viable cells present.
96-well plates were seeded with Hep-G2cells (~5,000 cells per well in 100 μL of DMEM) and incubated at 37 °C for 24 hours. The medium was removed and replaced with fresh medium (control) or various concentrations of test compounds and incubated at 37 °C for 48 hours.
Thereafter, cells were incubated with 80 μL DMEM and 20 μL of CTB per well for 6 hours at 37 °C. The fluorescence of the resofurin product (λ ex/em 560/620) was measured. Untreated wells were used as control.
The percentage of cell viability were calculated using the following formula: All assays were conducted in triplicate and the mean IC 50 ± standard deviation was determined.  Figure   S39). In RKO cells, no nanoparticles could be detected ( Figure S40).

Hemolysis assay
When the external membrane of the erythrocytes is destroyed, hemoglobin is released. [35][36][37] It is possible to estimate the amount of destroyed erythrocytes in a given test by measuring the quantity of hemoglobin in a sample by spectrophotometry. 38

Ex Vivo permeation study across mouse intestinal sac
Ex vivo absorption evaluation was carried out by permeation measurements in excised rat small intestine as described elsewhere. 39,40 Mice (25 g

In vivo animal assessments
All animals were raised in accordance with the policies of the University of Tlemcen Institutional Animal Care and Use Committee (IACUC) (accreditation number: D01N01UN130120150006).  Figure   3g-v) also due to STZ administration to induced diabetes. The kidneys of the rats treated with TTA-DFP-nCOF/insulin showed fewer alterations than the subcutaneous insulin rat kidneys, smaller Bowman's spaces and well-individualized tubules.

Biochemical determinations
Liver function test was carried out using serum biomarkers such as aspartate amino-transferase (AST) and Alanine transaminase (ALT) measured from the plasma obtained from the tail vein using SPINREACT kit. Kidney function test was performed using as urea and creatinine measured by SPINREACT kit.