The chemical composition of PEOT/PBT polymer is indicated as aPEOTbPBTc, in which ‘a’ is the molecular weight of the starting poly(ethylene glycol) PEG segments used in the polymerization process, while ‘b’ and ‘c’ refer to the weight percentage of PEOT and PBT blocks, respectively. The composition 4000PEOT30PBT70 was selected for its slow degradation rate and its limited cell-adhesive properties. The thin film microwell scaffolds were prepared from ∼50 µm thick 4000PEOT30PBT70 block copolymer films (IsoTis Integra Orhobiologics S.A., Irvine, USA). These polymer films were fabricated by solvent casting. A 10% (w/w) polymer solution was prepared in 20% (w/w) 1,1,1,3,3,3-hexafluoro-2-isopropanol (Biosolve, Valkenswaard, The Netherlands) and 80% (w/w) chloroform (Merck, Darmstadt, Germany) and casted on a silicon wafer at ambient temperature (23±3°C). The films were placed under nitrogen-flow for 12 hours to form dense films, incubated in ethanol overnight to remove solvent residue, and dried in a vacuum oven (Heraeus, Hanau, Germany) at 30°C for 3 days.

Electrospun microwell scaffolds were fabricated from 150 µm thick electrospun 4000PEOT30PBT70 block copolymer meshes. The meshes were fabricated from a 10% (w/w) polymer solution in 20% (w/w) 1,1,1,3,3,3-hexafluoro-2-isopropanol and 80% (w/w) chloroform using an electrospinning apparatus as previously described [30]. Electrospinning was performed at 12 kV constant voltage, with a 1.2 mm diameter needle, a 10 cm distance between the syringe and the collector, and a flow rate of 2 ml/h. Electrospun meshes were dried in a vacuum oven (Heraeus, Hanau, Germany) at 30°C for 3 days.

Microwell scaffolds were fabricated from the PEOT/PBT thin films and electrospun meshes by micro back molding [31], a variant of microthermoforming, which involves shaping a heated polymer film by three-dimensional stretching into a negative mold using a backing material (Figure S1) [32]–[35]. The negative mold was produced by Lightmotif BV (Enschede, The Netherlands) out of a 350 µm thick stainless steel (AISI 304) foil. Approximately 500 circular holes with a diameter of 400 µm were ablated by a titanium-sapphire laser (800 nm wavelength, circular polarized, 50 kHz pulse repetition rate, 5 µJ energy, and 200 fs pulse duration). A 25 µm laser spot was achieved by a 100 mm focal length lens. The laser bundle was scanned over the foil by a two-mirror galvanometer (Scanlab Scangine 14). After machining, the mold was cleaned in an ultrasonic bath. Molding temperature and pressure for thin film scaffolds were 85°C and 25 kN, respectively and for the electrospun microwell scaffolds they were 17°C and 3 kN, respectively. Scaffold geometry and architecture were characterized by scanning electron microscopy (SEM) on a Philips XL 30 ESEM-FEG. Scaffolds were gold sputter coated (Cressington, UK) prior to SEM analysis. The dimensions of the microwells (n = 10) were measured in three different samples per type of scaffold using ImageJ software (http://rsb.info.nih.gov/ij/).

To assess the effect of micro back molding on polymer wettability and water uptake, 4000PEOT30PBT70 block copolymer films were heat-treated (85°C) to simulate the scaffold fabrication cycle. The wettability was determined by static water contact angle measurements using the captive bubble method before and after heat-treatment. Measurements were performed using a video-based optical contact angle meter OCA 15 (DataPhysics Instruments GmbH, Filderstadt, Germany). The polymer films were mounted onto Scaffdex CellCrown inserts (Tampere, Finland) and incubated for three days in ultrapure Milli-Q water at 37°C. For measurements, the samples were placed with the flat polymer surface downward in an optical cuvette filled with ultrapure Milli-Q water. Water contact angle was determined by applying an air bubble (∼10 µL) on the film using an electronically regulated Hamilton syringe containing a curved needle. The contact angle was calculated using SCA20 software (DataPhysics Instruments GmbH, Filderstadt, Germany). The equilibrium water-uptake in ultrapure Milli-Q water was defined as the percentage weight gain of the polymer film after incubation at 37°C for 7 days. The swollen weight was measured after blotting the films with filter paper to remove surface water using a high-precision balance with an accuracy of ±0.01 mg (Sartorius BP210D, Göttingen, Germany).

Heat-treated (85°C) 4000PEOT30PBT70 block copolymer films were tested for cytotoxicity using a minimum essential medium (MEM) extract test (adapted from ISO 10993/EN 30993 standard). Small pieces of the polymer films (total surface of 36 cm²) were sterilized in 70% (v/v) ethanol overnight and dried for at least 12 hours. For extraction, the films were incubated in medium (α-MEM (Invitrogen, Carlsbad, USA) supplemented with 2 mM l-glutamine, 1 mM sodium pyruvate, 10% FBS, 100 U/ml penicillin and 10 µg/ml streptomycin) for 72 hours in a shaking water bath (37°C, 60 rpm). Natural rubber (Gelria Packing BV, Enschede, The Netherlands), polylactic acid films (PLA; Plastic Suppliers/Sidaplax, EarthFirst Packaging Film, Gentbrugge, Belgium), and polypropylene films (Walothen O 25 E, Germany) were extracted identically and used as a positive control, a reference, and a negative control for cytotoxicity, respectively. The extracts were added to a subconfluent monolayer of MC3T3-E1 cells (mouse pre-osteoblasts) and cultured for 72 hours. After 24, 48 and 72 hours the cells were evaluated under light microscopy and scored in a single-blinded manner three times for confluency, degree of floating cells, and change of cell morphology related to the negative control. After 72 hours, the cells were trypsinized and counted (Beckman Coulter, Miami, USA). Based on these results, the cytotoxic response index was scored according to ISO 10993/EN 30993 standard.

As an indicator for nutrient transport through thin film and electrospun scaffolds, glucose diffusion was assessed as previously described [36]. Briefly, the setup consisted of two double-walled compartments: a donor compartment filled with RPMI-1640 (Gibco) containing 11.1 mM d-glucose, and an acceptor compartment containing RPMI-1640 without d-glucose (both compartments were supplemented with 100 U/ml penicillin and 10 µg/ml streptomycin). The scaffolds were pre-wetted for several hours with an ethanol to medium (without d-glucose) gradient and placed between the two compartments which were maintained at 37°C with circulating water. At the indicated time intervals, 80 µl samples were obtained from both compartments and analyzed for glucose concentration using a Vitros DT60 II chemistry system (Ortho-Clinical Diagnostics, Raritan, USA). The glucose flux, the amount of glucose that passes through a certain surface in time, was calculated using the equation:(1)where C_acceptor is the glucose concentration in the acceptor compartment, V_acceptor is the volume of the acceptor compartment, and S, the surface area of the construct.

Human islets of Langerhans from pancreata of organ donors (n = 4) were obtained from the Human Islet Isolation Laboratory at the Leiden University Medical Center (Leiden, The Netherlands), which has permission from the Dutch government to isolate human islets for clinical purposes. Human islets of four different donor organs that were not eligible for clinical transplantation were used in these experiments, in accordance with Dutch law and institutional requirements.

Microwell scaffolds were sterilized in 70% (v/v) ethanol overnight and washed in PBS. Before islet seeding, scaffolds were centrifuged in islet culture medium (CMRL-1066 (Mediatech, Cellgro, Herndon, USA) containing 5.5 mM d-glucose, and supplemented with 0.02 mg/ml ciproxin, 0.05 mg/ml gentamicin, 2 mM l-glutamine, 0.25 µg/ml fungizone, 10 mM HEPES, 1.2 mg/ml nicotinamide and 10% (v/v) human serum) at 3500 rpm for 5 minutes to remove air bubbles and were mounted between two Teflon rings. Islets were handpicked and seeded drop-wise onto scaffolds. After 5 minutes of incubation, 8 ml of culture medium was added. Medium was changed every other day. As a reference, islets were also cultured free-floating in ultra-low attachment petri-dishes (Corning, New York, USA).

To assess morphology and viability, human islets were cultured in thin film and electrospun microwell scaffolds for 7 days. To assess islet morphology, samples were fixed in 4% (w/v) paraformaldehyde for 2 hours at ambient temperature (23±3°C), permeabilized with 0.1% Triton-X for 4 minutes and stained with alexa-488-phalloidin for 30 minutes in the dark to visualize actin in islet cells. Samples were imaged using a Zeiss LSM510 confocal MicroscopeAdditionally, islet morphology was assessed using SEM. As a reference for islet attachment, islets were cultured for 7 days on tissue culture plastic (NUNC ).Prior to SEM analysis, samples were fixed in 4% (w/v) paraformaldehyde for 2 hours at ambient temperature (23±3°C), dehydrated in graded ethanol series, critical-point dried, and gold sputter coated.

To assess islet function, two measures of islet response to glucose were made: insulin secretion and total insulin content were assessed at day 1 and day 7. Scaffolds containing islets or free-floating islets were transferred to ultra-low attachment 24-well plates (Corning, New York, USA). Each well contained approximately 25 islets. The islets were incubated in pre-warmed KRBH buffer (115 mM NaCl, 5 mM KCl, 24 mM NaHCO₃ and 2.2 mM CaCl₂, pH 7.4), supplemented with 20 mM HEPES, 2 mg/ml human serum albumin and 1.7 mM d-glucose for a 90 minutes at 37°C and 5% CO₂. The buffer was replaced by KRBH buffer containing either 1.7 mM or 16.7 mM glucose and the islets were incubated for 60 minutes. Experiments were performed in triplicate for each condition. The supernatant was removed and frozen for future analysis. Islets were harvested, washed in PBS and disrupted by sonication in distilled water (Braun, Melsungen, Germany). Samples from the homogenate were extracted overnight at 4°C in acidic ethanol (0.18 M HCl in 95% (v/v) ethanol). The supernatants and islet homogenate were assayed for insulin using an ELISA immunoassay (Mercodia, Uppsala, Sweden) according to the manufacturer’s instructions. For normalization, total DNA content was quantified using a Quant-iT PicoGreen dsDNA kit (Molecular Probes, Eugene, USA) according to the manufacturer’s instructions.

To identify α and β cells and assess morphology, histology was performed. Islets cultured in either microwell scaffolds or under free-floating conditions were washed in PBS and fixed in 4% (w/v) paraformaldehyde for 2 hours at ambient temperature (23±3°C). Subsequently, the samples were embedded in 2% (w/v) agarose and prepared for paraffin embedding. The samples were sectioned at 4 µm, deparaffinized in buthylene, and rehydrated through graded ethanol series to distilled water. Sections were blocked with an avidin-biotin blocking kit (Vector Laboratories, Burlingame, USA)) and normal donkey serum (1∶50 (in PBS containing 1% lamb serum), Jackson ImmunoResearch, West Grove, USA, 1 hour at room temperature). Primary and secondary antibodies were diluted in PBS containing 1% lamb serum. The following primary antibodies, dilutions, and incubation times were used: guinea pig anti-insulin (1∶200, Linco Research, St Charles, USA, 1.5 hours at room temperature) and rabbit anti-glucagon (1∶100, Vector Laboratories, Burlingame, USA, overnight at 4°C). The secondary antibodies were donkey anti-guinea pig-TRITC (1∶400, Jackson ImmunoResearch, West Grove, USA, 1 hour at room temperature) and biotinylated donkey anti-rabbit (1∶200, Jackson ImmunoResearch, West Grove, USA, 1 hour at room temperature) followed by streptavidin-Alexa Fluor488 (1∶200 Jackson ImmunoResearch, West Grove, USA, 1 hour at room temperature). Sections were mounted with mounting medium containing DAPI to stain the nuclei (VectaShield, Vector Laboratories, Burlingame, USA) prior to imaging on a confocal microscope (Carl Zeiss, Sliedrecht, the Netherlands).

Results are presented as the mean ± standard deviation. Statistical analyses were performed using an unpaired, two-tailed, Student’s t-test, or analysis of variance (ANOVA) with a Bonferroni post-test using SPSS statistics software (Chicago, USA). Statistical significance was considered at p<0.05.

We first determined the effects of our processing conditions on the 4000PEOT30PBT70 block copolymer. Water uptake and contact angle of the unprocessed polymer was 52.1±4.0% (w/w) and 39±2°, respectively. After heating the polymer to 85°C, the temperature used for scaffold fabrication, water uptake was significantly decreased to 38.3±3.4% (w/w), whereas no change in water contact angle was observed (Figure 1). We then sought to ensure the processing conditions did not affect the biocompatibility. To evaluate the cytotoxicity of heat-treated 4000PEOT30PBT70 block copolymer films, extracts of these materials were tested for their effects on growth and morphology of MC3T3-E1 cells. In contrast to exposure to extracts of natural rubber, which served as the positive control for cytotoxicity, nearly confluent monolayers were observed after exposure to extracts of polypropylene (negative control), PLA, and 4000PEOT30PBT70 films (Figure 2 A–B). No significant effect on cell morphology was observed for cells cultured in 4000PEOT30PBT70 extracts compared to PLA or polypropylene, indicating the polymer is biocompatible.

SEM analysis demonstrated homogeneously shaped microwells throughout the entire scaffold area (Figure 3 B), resembling the tapered shape of the mold, (Figure 3 A) and indicating successful demolding. The microwells were stable upon handling, due to the sufficiently high ratios of the thicknesses of the walls to the dimensions of the volumes enclosed by the walls. Per scaffold type, the upper and lower well diameters (Figure 3 E–F) of 10 microwells in three different samples were measured using ImageJ software (http://rsb.info.nih.gov/ij/). The upper well diameter of the thin film scaffolds was 377±5 µm and the lower well diameter 251±9 µm. For the electrospun microwells the upper diameter was 355±20 µm and the lower diameter was 217±36 µm. The fiber-diameter of the electrospun meshes was 1.71±0.42 µm. The height of the scaffold corresponded to the height of the mold (350 µm) (Figure 3 C–F).

As glucose is an important nutrient for islets, the diffusion of glucose through thin film and electrospun microwell scaffolds was assessed. Figure 4 A shows typical results of glucose diffusion in time through a thin film and an electrospun scaffold. No adsorption of glucose to the scaffolds was observed, since the difference between the initial concentration of glucose in the donor compartment and the concentration in the donor compartment in time equals the glucose concentration in the acceptor compartment. Due to the water uptake of this polymer composition, some glucose diffusion was observed through thin film microwell scaffolds (Figure 4 A), but the maximum glucose flux was significantly higher for the electrospun scaffolds: 4.76±0.53E10⁻⁴ gm⁻²s⁻¹ compared to 1.70±0.58E10⁻⁴ gm⁻²s⁻¹ for the thin film scaffolds (p<0.01) (figure 4 B).

Islets were seeded into the scaffolds by sedimentation. It is known that islets aggregate when they are cultured at high density. In rare cases, multiple islets were observed in one microwell. After prolonged culture, these islets start to fuse, but the final size of the fused islets remained within the dimensions of the microwell (∼350 µm), comparable to the upper size limit of human islets [37]. After 7 days of culture, the cells did not spread out and islets maintained their rounded shape (Figure 5 A–F) in both thin film and electrospun mesh microwell scaffolds. This in contrast to islets that were cultured on adherent tissue culture plastic used as an islet adhesion control (Figure 5 G–H). Some attached cells could be observed at the surface of the scaffold, but the degree of cell attachment was limited.

To assess whether the microwell scaffolds affect islets functionality, human islets were cultured for 7 days in both thin film and electrospun microwell scaffolds and subjected to a glucose-stimulated insulin secretion (GSIS) test. This test involves exposing islets to low (1.7 mM), high (16.7 mM), and then low glucose concentration and measuring insulin secretion by ELISA. After 7 days of culture, the increase in insulin release upon stimulation, compared to basal insulin release levels, did not significantly differ between islets seeded in either scaffold design (3.9±2.9 and 3.6±1.6 pmol/µgDNA/h for thin film and electrospun microwell scaffolds, respectively) and free-floating control islets (3.6±0.7 pmol/µgDNA/h, p>0.05). Following high glucose stimulation, insulin release levels returned to basal levels when incubated in low glucose buffer, indicating normal responsiveness (Figure 6 A). Total insulin content was also measured and did not differ between islets in the microwells or free-floating control islets (Figure 6 B). With no significant differences in insulin secretion between islets cultured in either the thin film or the electrospun mesh microwells, we sought to further demonstrate that donor variability was not influencing our results. We therefore compared insulin secretion and total insulin content in islets derived from three donors either in thin film microwells and free-floating islets and found no statistically significant differences (Figure 7 A–B) indicating that the microwell scaffold platform does not hamper islet functionality in terms of islet’s glucose-induced insulin secretion.

Having found that both the thin film and the electrospun microwell scaffolds supported islet insulin response equally, we performed histology of islets seeded in the thin film scaffold and compared this to free-floating islets. Islet morphology was unaffected by culture in the thin film microwell scaffolds after 7 days of culture (Figure 8). In both groups, glucagon and insulin expressing cells were typically present throughout the islets, further indicating that the microwell scaffolds did not negatively affect islet morphology nor function.

This study is the first to report on a microwell scaffold platform as a potential transplantation device for pancreatic islets. Reproducible microwell scaffolds were prepared from PEOT/PBT thin films and electrospun meshes. The morphology of human islets was preserved in the microwell scaffolds and remained stable during 7 days of culture and the islets showed an insulin release and total insulin content comparable to free-floating control islets. Glucagon and insulin immunostaining were also comparable between these groups. Together, these data indicate that the microwell scaffold platform does not hamper islet functionality and is a promising platform for use in extrahepatic islet transplantation.